Search is not available for this dataset
article
stringlengths 4.36k
149k
| summary
stringlengths 32
3.35k
| section_headings
sequencelengths 1
91
| keywords
sequencelengths 0
141
| year
stringclasses 13
values | title
stringlengths 20
281
|
|---|---|---|---|---|---|
Immunoregulatory cytokine interleukin-10 ( IL-10 ) is elevated in sera from patients with systemic lupus erythematosus ( SLE ) correlating with disease activity . The established association of IL10 with SLE and other autoimmune diseases led us to fine map causal variant ( s ) and to explore underlying mechanisms . We assessed 19 tag SNPs , covering the IL10 gene cluster including IL19 , IL20 and IL24 , for association with SLE in 15 , 533 case and control subjects from four ancestries . The previously reported IL10 variant , rs3024505 located at 1 kb downstream of IL10 , exhibited the strongest association signal and was confirmed for association with SLE in European American ( EA ) ( P = 2 . 7×10−8 , OR = 1 . 30 ) , but not in non-EA ancestries . SNP imputation conducted in EA dataset identified three additional SLE-associated SNPs tagged by rs3024505 ( rs3122605 , rs3024493 and rs3024495 located at 9 . 2 kb upstream , intron 3 and 4 of IL10 , respectively ) , and SLE-risk alleles of these SNPs were dose-dependently associated with elevated levels of IL10 mRNA in PBMCs and circulating IL-10 protein in SLE patients and controls . Using nuclear extracts of peripheral blood cells from SLE patients for electrophoretic mobility shift assays , we identified specific binding of transcription factor Elk-1 to oligodeoxynucleotides containing the risk ( G ) allele of rs3122605 , suggesting rs3122605 as the most likely causal variant regulating IL10 expression . Elk-1 is known to be activated by phosphorylation and nuclear localization to induce transcription . Of interest , phosphorylated Elk-1 ( p-Elk-1 ) detected only in nuclear extracts of SLE PBMCs appeared to increase with disease activity . Co-expression levels of p-Elk-1 and IL-10 were elevated in SLE T , B cells and monocytes , associated with increased disease activity in SLE B cells , and were best downregulated by ERK inhibitor . Taken together , our data suggest that preferential binding of activated Elk-1 to the IL10 rs3122605-G allele upregulates IL10 expression and confers increased risk for SLE in European Americans .
The gene cluster that includes interleukin 10 ( IL10 ) , IL19 , IL20 and IL24 is located on chromosome 1q31-32 , a genomic region that is linked with susceptibility to systemic lupus erythematosus ( SLE , OMIM 152700 ) [1] , [2] . Recent genome-wide association ( GWA ) and follow-up replication studies in European ancestry have identified an association between the minor allele of rs3024505 , a SNP located at 1 kb downstream of IL10 , and increased risk for SLE [3] , inflammatory bowel disease ( IBD ) including both Crohn's disease ( CD ) [4] and ulcerative colitis ( UC ) [5] , [6] , and decreased risk for type 1 diabetes [7] , [8] . In addition , SNPs in the second intron ( rs1518111 ) and the promoter region ( rs1800871 ) of IL10 have been reported to be associated with Behçet's disease ( BD ) in GWAS of Turks [9] and Japanese [10] . These findings indicate IL10 as a common susceptibility locus shared by SLE and several other autoimmune diseases . Dysregulation of IL-10 family cytokines contributes to autoimmune disease and tissue damage ( reviewed in [11] ) . IL-10 is an important immunoregulatory cytokine with a wide variety of functions in T cells , B cells , natural killer cells , dendritic cells and macrophages [12] . The observations of elevated serum IL-10 levels in SLE patients correlating with increased disease activity [13] , [14] , and promising findings of anti-IL-10 monoclonal antibody treatment in patients with SLE [15] support a pivotal role for IL-10 in the pathogenesis of SLE . Of interest , elevated IL-10 levels were also reported in first-degree relatives of SLE patients [16] , [17] , suggesting that levels of IL10 expression may be determined genetically . In this study , we fine mapped the IL10 gene cluster for genetic association with SLE in 15 , 533 case and control subjects from four diverse ancestries , identified a causal variant rs3122605 at IL10 5′ upstream using both genetic and functional assays , and explored the underlying molecular mechanism in explaining the elevated IL-10 levels in patients with SLE associated with increased disease activity .
To fine map the IL10 gene cluster , we genotyped 19 tag SNPs in 15 , 533 case and control subjects from four ancestries , including European American ( EA , 3 , 820 cases vs . 3 , 412 controls ) , African American ( AA , 1 , 670 vs . 1 , 904 ) , Asian ( AS , 1 , 252 vs . 1 , 249 ) and Amerindian/Hispanic ( HS , 1 , 445 vs . 781 ) . Each SNP was assessed for the association with SLE susceptibility under a logistic regression model adjusted for gender and global ancestry ( Figure 1A ) . In the largest EA dataset , rs3024505 located at 1 kb downstream of IL10 exhibited the strongest association with SLE ( minor allele frequency of 18 . 2% in cases vs . 14 . 8% in controls , P = 2 . 7×10−8 , OR [95%CI] = 1 . 30 [1 . 19–1 . 43] ) , which exceeded the GWAS significance level of P<5×10−8 ( Table S1 ) . To identify additional SLE-associated SNPs , we performed SNP imputation using 1000 Genomes Project data as a reference . In EA , a total of 109 well-imputed SNPs spanning 154 kb from IL10 downstream to FAIM3 were assessed for association with SLE . Of them , three imputed SNPs ( rs3122605 , rs3024493 and rs3024495 located at 9 . 2 kb upstream , intron 3 and 4 of IL10 , respectively ) , which were in tight linkage disequilibrium ( LD , r2>0 . 9 ) with the genotyped SNP rs3024505 , were strongly associated with SLE and remained significant after Bonferroni correction ( rs3122605: P = 1 . 3×10−8 , OR = 1 . 34 [1 . 21–1 . 48]; rs3024493: P = 5 . 0×10−8 , OR = 1 . 29 [1 . 18–1 . 42] ) ; rs3024495: P = 1 . 0×10−7 , OR = 1 . 29 [1 . 17–1 . 41] ) ( Figure 1B and Table S1 ) . None of the genotyped or imputed SNPs was significantly associated with SLE in three non-European datasets including AS , AA and HS after Bonferroni correction ( Table S2 ) . These data confirmed that IL10 is a risk locus for SLE in EA , and thus we subsequently focused on EA only to identify the causal variant ( s ) . IL10 promoter SNPs rs1800872 ( also named as −592T/G ) , rs1800871 ( −819G/A ) and rs1800896 ( −1082T/C ) , which were identified to be associated with elevated IL-10 production and SLE susceptibility in some , but not all of previous studies ( reviewed in [18] , [19] ) , showed nominal association with SLE in our EA dataset ( Table S1 ) . The BD-associated SNP rs1518111 showing effect on decreased IL-10 levels [9] was not associated with SLE in EA ( Table S1 ) . We performed haplotypic analysis to investigate relationships between these four previously reported SNPs and rs3024505 , rs3024495 , rs3024493 and rs3122605 . Only the haplotype H4 carrying risk alleles of rs3024505 , rs3024495 , rs3024493 and rs3122605 was strongly associated with SLE ( frequency of 16 . 8% in cases vs . 13 . 6% in controls , P = 1 . 3×10−7 ) ( Figure 1D ) . Due to strong LD among rs3122605 , rs3024493 , rs3024495 and rs3024505 in EA ancestry , their associations with SLE were highly correlated and could not be distinguished from each other using the conditional haplotype-based association test ( Table S1 ) . Conditioning on rs3122605 , rs3024493 , rs3024495 and rs3024505 , respectively , association signals ( P<0 . 05 ) of all other SNPs within the IL10 gene cluster were completely eliminated ( Table S1 ) , suggesting that these four SNPs within tight LD could capture all associations of the IL10 cluster region with SLE in EA . Of note , searching +/−200 kb of IL10 based on the 1000 Genomes Project data , we found that rs61815643 located at 10 . 3 kb upstream of IL10 was also in strong LD with rs3122605 ( r2 = 0 . 9 ) in European subjects , which suggested that this SNP might account for association signals detected within the IL10 gene cluster . However , because the imputation quality of rs61815643 did not reach the threshold of information score >0 . 9 , it was not included for association test in this study . Taken together , our data provide evidence supporting IL10 as a risk locus for SLE in EA and the underlying causal variant ( s ) might be or tagged by rs3122605 , rs3024505 , rs3024495 and rs3024493 . To explore potential functional consequences of the SLE-associated SNPs ( rs3122605 , rs3024493 , rs3024495 and rs3024505 ) , we assessed their genetic effects on influencing IL10 expression . IL10 mRNA levels in peripheral blood mononuclear cells ( PBMC ) and IL-10 protein levels in plasma from EA subjects were measured by quantitative real-time PCR and ELISA , respectively . Using rs3122605 as a surrogate of the other three SNPs , we compared IL10 expression levels among subjects carrying different genotypes of rs3122605 . In control subjects , the SLE-risk ( G ) allele of rs3122605 was dose-dependently associated with elevated IL10 expression at both mRNA ( n = 55; P = 0 . 001 , R2 = 0 . 18 in linear regression ) and protein ( n = 116; P = 3 . 3×10−7 , R2 = 0 . 21 ) levels ( Figure 2 ) . Consistently , dose-dependent association of the risk allele with elevated IL10 expression was also observed in patients with SLE at both mRNA ( n = 58; P = 6 . 4×10−6 , R2 = 0 . 31 ) and protein ( n = 132; P = 1 . 4×10−6 , R2 = 0 . 16 ) levels ( Figure 2 ) . Compared to healthy controls , higher IL10 expression was observed in patients with SLE carrying the same genotype at both mRNA ( genotype AA: P = 1 . 3×10−4 , AG: P = 4 . 3×10−4 , GG: P = 3 . 9×10−5 , cases vs . controls in t test ) and protein ( AA: P = 4 . 6×10−4 , AG: P = 4 . 9×10−6 , GG: P = 0 . 047 ) levels ( Figure 2 ) , probably due to the activated immune status of SLE patients . We hypothesized the presence of transcription factors activated in SLE patients upregulating IL10 expression and prepared nuclear extracts of peripheral blood lymphocytes from active SLE patients ( defined as SLEDAI score≥4 ) [20] , [21] to perform electrophoretic mobility shift assays ( EMSA ) for testing allelic differences in transcription factor binding conferred by the SLE-associated SNPs rs3122605 , rs3024505 , rs3024495 and rs3024493 . Because we could not exclude the possibility that rs61815643 is a SLE-risk SNP affecting IL10 expression , it was also tested by EMSA . Upon incubation with nuclear extracts , specific mobility-shift bands were only detected using the oligodeoxynucleotide probe containing the risk , but not the non-risk allele of rs3122605 ( Figure 3A ) . Of interest , no specific binding of nuclear proteins was observed with the oligodeoxynucleotide probes containing either the risk or the non-risk allele of rs3024505 , rs3024493 , rs3024495 and rs61815643 ( Figure S1 ) . In silico analysis using the program TFSEARCH indicated that the risk allele of rs3122605 might create a novel binding site of the transcription factor Elk-1 ( ETS-like transcription factor 1 ) . We validated this prediction by showing that the addition of polyclonal rabbit IgG anti-ElK-1 antibody produced a super shift band only to the probe containing the risk but not the non-risk allele of rs3122605 in EMSA ( Figure 3A ) . Taken together , these data showed that rs3122605 , the risk allele of which binds to Elk-1 in peripheral blood lymphocytes from SLE patients with active disease , is more likely to be the causal variant upregulating IL10 expression than the other four candidate SNPs . Elk-1 is activated through phosphorylation and the phosphorylated Elk-1 ( p-Elk-1 ) translocates into the nucleus to induce gene transcription [22] . To investigate the role of Elk-1 in SLE , we compared the distribution and activation of Elk-1 in PBMCs between SLE cases ( n = 4 ) and healthy controls ( n = 4 ) using Western blot . The amount of total Elk-1 was higher in nuclear ( Figure 3B ) but lower in cytoplasmic extracts ( Figure 3C ) of cases than controls . Of interest , p-Elk-1 was detected only in nuclear extracts of cases but not in controls ( Figure 3B ) and not in cytoplasmic extracts of either cases or controls ( Figure 3C ) . These data suggest that Elk-1 is aberrantly activated and accumulates in nuclei of SLE PBMCs . Furthermore , the amount of total Elk-1 and p-Elk-1 appeared to be increased in SLE patients with higher SLEDAI scores ( Figure 3B ) . Using flow cytometry , we quantified the co-expression of IL-10 and p-Elk-1 in specific cell subsets of PBMCs including CD3+ T cells , CD19+ B cells and CD14+ monocytes from healthy controls and SLE patients and showed representative data plotted in two-dimensional dot plots in Figure 4A . Compared to healthy controls ( n = 3 ) , percentages of IL-10+p-Elk-1+ cells were significantly increased in B ( n = 11 ) , T cells ( n = 12 ) and monocytes ( n = 12 ) from SLE patients ( Student's t-test: P = 0 . 013 , 0 . 012 and 0 . 012 , respectively , Figure 4B ) . Active SLE patients had significantly elevated IL-10+p-Elk-1+ double positive B cells compared to non-active SLE patients ( n = 7 vs . 4 , P = 0 . 038 ) . Similar trends for association with SLE disease activity were detected in IL-10+p-Elk-1+ T cells and monocytes , but the difference was not statistically significant ( P = 0 . 74 and 0 . 57 , respectively ) . Elk-1 is known to be activated by mitogen-activated protein kinases ( MAPK ) , including ERK ( extracellular-signal-regulated kinases ) , JNK ( c-Jun N-terminal kinases ) and p38 . We wondered which MAPK inhibitor could best down-regulate the elevated co-expression of IL-10 and p-Elk-1 in PBMCs . Because of the low expression of IL-10 in freshly isolated control PBMCs ( Figure S2 ) , we stimulated control PBMCs with IFN-α , a pivotal cytokine upregulated in most SLE patients [23] , to mimic SLE PBMCs and incubated them with the MAPK inhibitors specific to ERK ( PD 98059 ) , JNK ( SP 600125 ) or p38 ( SB 203580 ) , respectively . IFN-α-stimulation could significantly increase the percentage of IL-10+p-Elk-1+ cells in B , T cells and monocytes ( n = 5 , P = 0 . 0002 , 0 . 0018 and 0 . 0079 , respectively , Figure 4C ) , and such increase could be best down-regulated by the addition of the ERK inhibitor ( P = 2 . 8×10−5 , 0 . 0046 and 0 . 0079 in B , T cells and monocytes , respectively , Figure 4C ) . To a lesser extent , the p38 inhibitor also significantly suppressed IFN-α-induced double positive cells in all three cell subsets ( P = 0 . 0016 , 0 . 01 and 0 . 0039 in B , T cells and monocytes , respectively ) . Under our experimental conditions , the JNK inhibitor significantly inhibited the percentage of IL-10+p-Elk-1+ cells in B cells and monocytes ( P = 0 . 0048 and 0 . 009 , respectively ) , but not in T cells ( P = 0 . 10 ) .
Our data provide strong evidence for a dose-dependent association between SLE-predisposing IL10 genotypes and corresponding mRNA and protein levels of IL-10 , and identify one underlying molecular mechanism to explain previous findings of elevated IL-10 serum levels in SLE patients that positively correlated with increased disease activity . In addition to confirming the previously reported association with SLE at the IL10 3′ downstream SNP ( rs3024505 ) in European Americans , we identified a SLE-associated risk haplotype , defined by the minor alleles of four SNPs in tight LD , rs3024505 , rs3024495 , rs3024493 and rs3122605 , which could best explain the association with SLE and capture underlying causal variant ( s ) within the IL10 gene cluster in EA ancestry . The minor allele of rs3122605 , which tags the IL10 SLE-risk haplotype , exhibited a dose-dependent association with elevated IL10 expression at both mRNA levels in PBMCs and protein levels in plasma samples from SLE patients and healthy controls , suggesting that these four SLE-associated SNPs may act by influencing IL10 regulation . Further functional studies showed that only rs3122605 was experimentally validated to confer preferential allele-binding to the transcription factor Elk-1 present in SLE PBMCs , hence the most likely functional variant present on the IL10 risk haplotype . Compared to normal PBMCs , nuclear localization of activated p-Elk-1 was observed only in SLE PBMCs . Co-expression of p-Elk-1 and IL-10 was significantly increased in all SLE PBMC subsets compared to normal PBMC subsets . Of interest , SLE patients with active disease had higher double positive ( p-Elk-1+IL-10+ ) B cells than those with inactive disease . These data suggested that nuclear accumulation of activated Elk-1 in SLE peripheral lymphocytes contributes to overproduction of IL-10 in SLE patients associated with disease activity . An abnormally high production of IL-10 in patients with SLE has been consistently demonstrated in many studies ( reviewed in [18] ) , but the underlying molecular mechanism remains less well-characterized . The observation that healthy relatives of SLE patients also exhibit increased levels of IL-10 [16] , [24] , [25] suggests a possibility of genetically regulated IL-10 production . A number of genetic polymorphisms in the IL10 promoter region have been reported [26]–[30] , in particular , three SNPs shown in Figure 1D located at −1082 ( rs1800896 , C/T ) , −819 ( rs1800871 , G/A ) and −592 ( rs1800872 , G/T ) have been inconsistently associated with IL-10 production levels and risk of SLE ( reviewed in [18] , [19] ) . Identification of SNPs in the −1 . 3 to −4 kb region of the IL10 promoter associated with both IL-10 production phenotypes and SLE susceptibility [31] suggested that further evaluating the contribution of SNPs in the more distal promoter of IL10 might be warranted . Consistently , the SNP ( rs3122605 ) we identified that tags the SLE-risk haplotype in EA ancestry and confers genetic effect on IL10 expression is located at 9 . 2 kb upstream of IL10 . According to the ENCODE Project , rs3122605 , rs3024505 , rs3024493 and rs61815643 were located within DNasel hypersensitive and transcription factor binding sites in at least one cell type ( as shown in UCSC genome browser ) , suggesting that each of them may affect gene expression through interaction with regulatory elements . Given that gene regulation by genetic variants often occurs within the specific cell types most relevant to the disease phenotype [32] , we used nuclear extracts from PBMCs of active SLE patients to perform EMSA assays and found only the SLE-risk allele of rs3122605 preferentially binds to the transcription factor Elk-1 . Therefore , our data support rs3122605 as the most likely causal variant on the SLE-associated haplotype and implicate a potential importance for Elk-1-mediated upregulation of IL10 expression in SLE patients , particularly in those patients carrying the risk allele of rs3122605 . Elk-1 is a member of the Ets oncogene family of transcription factors characterized by a conserved DNA-binding domain and a C-terminal activation domain containing multiple phosphorylation sites targeted by three major MAP kinase pathways [33] . Different phosphorylation patterns of Elk-1 mediated through activation of MAPK signaling cascades by distinct external stimuli are important for Elk-1 to execute its physiologic functions [34]–[37] . We used an antibody to measure phosphorylation of Elk-1 at S383 , and detected the presence of Elk-1 pS383 in the nuclei of SLE but not normal PBMCs . Quantification of IL-10 and p-Elk-1 co-expression confirmed a higher proportion of IL-10+p-Elk-1+ cells in SLE than normal PBMC subsets . These findings imply that Elk-1 in nuclei of SLE PBMCs has been biologically activated likely due to a higher baseline immune activation status , which may enhance ability of Elk-1 to regulate IL10 transcription . In the absence of microenvironmental activation in control PBMCs , Elk-1 remains within the cytoplasm and less translocated into nuclei , limiting its regulation effects . This hypothesis may in part explain the observation of higher IL-10 expression in SLE patients than healthy controls even if they carry the same risk genotype of rs3122605 . In support of this possibility , previous studies revealed an increased expression of activation markers on peripheral lymphocytes of SLE patients , including phosphorylated ERK , JNK and p38 which are prerequisite for subsequent activation of Elk-1 [38] , [39] . In addition , our data showed a significantly increased proportion of IL-10+p-Elk-1+ cells in normal PBMC subsets ( Figure 4C ) when exposed to IFN-α which may induce a partial activation phenotype in lymphocytes mimicking that of SLE . We used healthy control , rather than SLE , PBMCs for testing effects of MAPK inhibitors on co-expression of IL-10 and p-Elk-1 , because disease activity and medications of SLE patients might confound the results and leucopenia of SLE patients could limit the amount of PBMCs available for our experiments . In all three cell subsets of PBMCs , inhibition of ERK could best suppress IFN-α induced increase in IL-10 and p-Elk-1 co-expression , highlighting the importance for phosphorylated Elk-1 via ERK signaling in regulation of IL10 transcription . The ERK-dependent activation of Elk-1 has been clearly demonstrated in neuronal cells ( reviewed in [40] ) in which phosphorylation of Elk-1 at S383/389 by ERK is tightly linked to its activation and nuclear translocation and inhibition of phosphorylation results in cytoplasmic Elk-1 retention , limiting its transcriptional properties [22] . Elk-1 , like all members of Ets-domain containing transcription factors , can bind genomic regions similarly as well as uniquely to regulate distinct classes of target genes [41] . Another member of the Ets family , Ets-1 , has a dual function in regulating IL10 gene expression acting as both a transcriptional activator with the binding partner Sp-1 in HIV-1Tat-induced IL10 transcription in THP-1 cells [42] , and as a repressor interacting with histone deacetylase 1 ( HDAC1 ) in Th1 cells [43] . Increasing evidence indicates the involvement of Ets-1 in the pathogenesis of SLE: ( 1 ) ETS1 has been identified as a risk locus for SLE in GWAS [44] , [45] and the risk allele is associated with decreased levels of ETS1 transcripts in healthy control PBMCs [45]; ( 2 ) Ets-1 is critical in maintaining B cell identity and its absence drives terminal differentiation of B cells into immunoglobulin-secreting plasma cells [46] , [47]; ( 3 ) Ets-1 functions as a cofactor for T-bet essential for Th1 effector function and differentiation [48]; ( 4 ) Ets-1 negatively regulates Th17 cell differentiation and Ets-1 deficiency results in elevated production of IL-17 and IL-17-related cytokines by Th17 cells [49] . These emerging findings support an important role of Ets family transcription factors in the development of SLE manifestations . Previous findings indicated that increased production of IL-10 by SLE PBMCs was mainly derived from B cells [16] , [50]–[53] . One explanation might be due to elevated expression of toll-like receptor 9 ( TLR9 ) on B cells of SLE patients with active disease , as the study showed that TLR9-CpG interaction could enhance the production of anti-dsDNA antibody and IL-10 [54] . B cell receptor ( BCR ) stimulation or BCR-TLR9 costimulation have been shown to activate the Erk pathway in B cells of NZB×NZW F1 mice [55] . A plausible explanation of elevated co-expression of IL-10 and p-Elk-1 in B cells from SLE patients , especially during active disease , might be attributable to activated ERK and Elk-1 involved in the BCR-dependent IL-10 production in SLE B cells . Overexpression of B-cell-derived IL-10 contributes to the pathogenesis of SLE likely dependent on its ability to promote B cell proliferation , differentiation and autoantibody production [56] , [57] . Recently identified IL-10 producing regulatory B cells ( Bregs ) may exert immunosuppressive effects to modulate murine lupus [58]–[63] . The phenotypic markers of human Breg cells have not reached consensus; they may be enriched in CD19+CD24hiCD38hi and CD24hiCD27+ peripheral blood cells [64] , [65] . Interestingly , CD19+CD24hiCD38hi B cells from SLE patients produce less IL-10 upon stimulation and are functionally impaired in suppressive capacity [64] . Thus , it seems unlikely that Bregs are the major producer of elevated IL-10 we observed in SLE patients . The minor allele of rs3024505 showed consistently higher frequency in SLE patients than controls in all four ancestries , but only reached statistical significance for association with SLE in EA . Upon considering different genetic models , the additive model yielded the best genotypic association in EA ( P = 2 . 7×10−8 , OR [95%CI] = 1 . 30 [1 . 19–1 . 43] ) . Given the lack of evidence for genetic heterogeneity across EA , AS , AA and HS ( P = 0 . 66 for Q statistic ) , the lack of significant association between rs3024505 and risk of SLE in AS , AA and HS might be due to low minor allele frequency and small sample size . Under the assumption that the minor allele of rs3024505 confers genetic risk with an odds ratio of 1 . 3 ( determined in EA ) , the power to detect a significant association ( P<0 . 05 ) for EA samples reaches 100% , whereas it is only 31% in AS , 55% in AA and 58% in HS datasets . According to the 1000 Genome Project data , the LD strength between rs3024505 and our proposed causal SNP rs3122605 is similarly strong in Asians ( r2 = 1 . 0 ) and Americans/Hispanics ( r2 = 0 . 8 ) as in Europeans ( r2 = 0 . 85 ) , but not in Africans ( r2 = 0 . 1 ) , suggesting that rs3122605 can be tagged by rs3024505 in non-Europeans except for the African-derived population . It is possible that other SLE-associated variant ( s ) specific for AS , AA or HS failed to be captured by SNPs used in this study due to different LD pattern in each ancestry . Among non-European populations , SLE GWAS conducted in AA and HS are not in the currently available literature . To our knowledge , there have been four published SLE GWA studies conducted in AS , including two Chinese [44] , [45] , one Japanese [66] and one Korean GWAS [67] , and a meta-analysis study based on Chinese GWAS [68] . In these studies , the Illumina Human 610 BeadChip is the commonly used genotyping platform , which contains nine IL10 SNPs ( including rs3024505 ) and can capture 23 of the 38 common SNPs ( MAF>1% ) within +/−5 kb of IL10 with r2>0 . 9 in Asians ( according to the 1000 Genome Project data ) . Because none of these studies have reported association signals of IL10 SNPs with SLE , current data are consistent with our findings that the IL10 locus is not a strong genetic risk factor for SLE in AS . Taken together , the association of IL10 with SLE in non-EA ancestries awaits further investigation . We have searched four publically available cis-eQTL datasets conducted in immune cells from healthy Europeans [69]–[72] but found no convincing evidence to support the presence of another SNP that can better capture the association signal of IL10 expression trait in these datasets than our data of SLE-associated rs3024505/rs3122605 . In addition , there was no convincing evidence to support the association of rs3024505 with differential expression of other genes within +/−1 Mb flanking region of IL10 . Given that the SNP rs3024505 confers increased risk for both SLE and IBD and could be tagged by rs3122605 , it is possible that patients with IBD carrying the risk allele of rs3024505 may exhibit high serum levels of IL-10 . There is evidence supporting elevated circulating IL-10 levels in both CD and UC patients that positively associated with disease activity [73]–[75] , similarly to previous reports in SLE patients . However , genetically engineered mice exhibiting low to no IL-10 signaling in the intestinal tract develop severe IBD manifestations [76]–[78] , supporting a pivotal role of IL-10 in down-regulation of inflammation . Increased levels of circulating IL-10 may be elicited by chronic inflammation in IBD , but may not be sufficiently strong to dampen intestinal inflammation [75] , raising the possibility of defective IL-10 signaling at sites of organ damage in patients with SLE . In conclusion , by characterizing genetic variations within the IL10 gene cluster region , we have identified the IL10 upstream SNP rs3122605 as the best likely causal variant responsible for association with SLE in European Americans . The SLE-associated rs3122605 G-allele preferentially binds to the activated Elk-1 conferring elevated IL10 expression . The observation that SLE patients , particular those with increased disease activity , showed enhanced activation of Elk-1 in nuclei and elevated co-expression of IL-10 and phospho-Elk-1 in peripheral lymphocytes highlights the involvement of aberrant Elk-1 signaling in development of SLE and suggests potential targeting therapy for disease amelioration .
This study was approved by the Institutional Review Boards ( IRBs ) or the ethnic committees at the institutions where subjects were recruited . All subjects were enrolled after informed consent had been obtained . The overall study was approved by the IRB of the Oklahoma Medical Research Foundation ( OMRF ) . To test the association of IL10 family genes with SLE , we used a large collection of case-control subjects from the collaborative Large Lupus Association Study 2 ( LLAS2 ) [79] , including European American ( EA: 4 , 248 cases vs . 3 , 818 controls ) , African American ( AA: 1 , 724 vs . 2 , 024 ) , Asian ( AS: 1 , 328 vs . 1 , 348 ) and Hispanic enriched for the Amerindian-European admixture ( HS: 1 , 622 cases vs . 887 controls ) . African Americans included 286 Gullahs ( 155 cases vs . 131 controls ) , who are subjects with African ancestry . Asians were composed primarily of Koreans ( 906 cases and 1 , 012 controls ) but also included Chinese , Japanese , Taiwanese and Singaporeans . Cases were defined by meeting at least four of the 1997 American College of Rheumatology ( ACR ) revised criteria for the classification of SLE [80] . To test functional consequences of SLE-associated variants , SLE patients and healthy controls of European ancestry were recruited at the University of California , Los Angeles and through the Lupus Family Registry and Repository ( LFRR , lupus . omrf . org ) for blood donations . DNA samples were processed at the Lupus Genetics Studies Unit of OMRF . SNP genotyping was performed using an Illumina custom bead array on the iSCAN instrument for 19 tag SNPs covering over 135 kb of IL10-IL24 region and 347 admixture informative markers ( AIMs ) . SNPs meeting the following criteria were included for subsequent genetic association tests: well-defined cluster scatter plots , SNP call rate >90% , minor allele frequency >1% , total proportion missing <5% , P>0 . 05 for differential missing rate between cases and controls , and Hardy-Weinberg proportion ( HWP ) test with a P>0 . 01 in controls and P>0 . 0001 in cases . Subjects with individual genotyping missing rate >10% ( due to low quality ) , shared identity by descent >0 . 4 or showing mismatch between the reported and estimated gender were removed . The global ancestry of each subject was estimated based on genotype of AIMs , using principal components analysis ( PCA ) [81] and ADMIXMAP [82]–[84] , as described in another LLAS2 study [85] , and then genetic outliers were removed . Finally , a total of unrelated 15 , 533 subjects including EA ( 3 , 820 cases vs . 3 , 412 controls ) , AA ( 1 , 670 vs . 1 , 904; composed of 92 . 5% African Americans and 7 . 5% Gullahs ) , AS ( 1 , 252 vs . 1 , 249; composed of 74 . 6% of Koreans , 16 . 1% of Chinese and subjects from Japan and Singapore ) and HS ( 1 , 445 vs . 781 ) were analyzed for 19 SNPs . Imputation was performed using IMPUTE 2 . 1 . 2 [86] , with SNP genotypes of 379 Europeans ( CEU , TSI , GBR , FIN and IBS ) , 246 Africans ( YRI , ASW and LWK ) , 286 Asians ( CHB , JPT and CHS ) and 181 Americans ( MXL , PUR and CLM ) from the 1000 Genomes Project ( version 3 of the phase 1 integrated data , March 2012 release ) as references in imputation for our EA , AA , AS and HS subjects , respectively . Imputed genotypes had to meet the threshold of information score >0 . 9 , as well as the quality control criteria as described above . After imputation , we obtained an additional 109 SNPs for EA , 45 for AA , 80 for AS and 64 for HS ( the number varied due to different LD structure ) for further analysis . Total RNA was purified with TRIzol reagent ( Life Technologies ) from PBMCs of EA individuals ( 58 SLE cases and 55 healthy controls ) and reverse-transcribed into cDNA with SuperScript II Reverse Transcriptase kit ( Life Technologies ) . Messenger RNA levels of IL10 and a housekeeping gene RPLP0 were measured by quantitative real-time PCR using TaqMan assays ( IL10 probe: Hs00961622_m1; RPLP0 probe: Hs99999902_m1 , Applied Biosystems ) . All samples were run in triplicate . Relative IL10 mRNA levels were normalized to that of RPLP0 , calculated by the 2−ΔΔCt method and Log10 transformed . Plasma IL-10 levels from 132 SLE patients and 116 healthy controls of EA ancestry were measured by ELISA ( R&D systems ) . To examine whether inhibition of MAPK pathway may affect co-expression of IL-10 and p-Elk-1 , control PBMCs ( 1×106 ) were cultured in growth medium with or without interferon alpha ( IFNα ) ( 1000 U/ml; PBL Biomedical Laboratories ) in the presence or absence of one MAPK inhibitor ( EMD Millipore ) , PD 98059 ( 20 µM; ERK/MEK inhibitor ) , SP 600125 ( 20 µM; JNK inhibitor ) or SB 203580 ( 10 µM; p38-MAPK pathway inhibitor ) , respectively . Addition of Brefeldin A ( eBiosciense ) to cells in culture blocks intracellular protein ( IL-10 ) transport processes . The patients with SLE recruited in this part of study were evaluated for disease activity by the SLE Disease Activity Index ( SLEDAI ) 2000 [87] at the time of blood draw , and SLEDAI≥4 was considered as active disease [20] , [21] . Freshly isolated or cultured PBMCs were incubated with mouse reference serum to block nonspecific binding to Fcγ receptors and then incubated with peridinin chlorophyll protein ( PerCP ) -conjugated anti-human CD3 , allophycocyanin ( APC ) -conjugated anti-human CD19 and phycoerythrin ( PE ) -conjugated anti-human CD14 ( eBiosciense ) to identify T cell , B cell and monocyte subpopulations , respectively . For intracellular staining of IL-10 and p-Elk-1 , cells were fixed with IC Fixation Buffer ( eBiosciense ) , washed with Permeabilization Buffer ( eBiosciense ) , and stained with fluorescein isothiocyanate ( FITC ) -conjugated anti-human IL-10 ( eBiosciense ) and PE- or Alexa Fluor647-conjugated anti-phospho-Elk-1 antibody ( BD Biosciences ) . Background fluorescence was assessed using appropriate isotype- and fluorochrome-matched control antibodies . Cells were collected and analyzed by FACSCalibur flow cytometer equipped with the manufacturer's software ( CellQuest; BD Biosciences ) . Bioinformatic analysis using the program TFSEARCH ( conducted on 01/18/2011 ) showed predicted binding to Elk-1 and STRE in DNA sequence containing the minor but not the major allele of rs3122605 ( Figure S3 and S4 ) . In addition , HSF , ADR1 and MZF1 were predicted to bind with sequence containing either allele of rs3122605 . Given that we were interested in identifying transcription factors that preferentially bind to the minor allele of rs3122605 , and that the STRE ( stress response element ) binding factor includes two yeast transcription factors , Msn2p and Msn4p , we prioritized to test Elk-1 using EMSA . EMSA and supershift assays were performed as previously described [88] . Nuclear extracts were prepared from peripheral blood lymphocytes of SLE patients with active disease using NE-PER Nuclear Extraction Reagent ( Thermo scientific ) and incubated with biotin-labeled oligodeoxynucleotides ( synthesized by Integrated DNA Technologies , depicted in Table S3 ) . EMSAs were performed with the LightShift Chemiluminescent EMSA kit ( Thermo scientific ) . The antibody used in the supershift reactions was polyclonal rabbit anti-human Elk-1 ( Santa Cruz Biotechnology ) . Cytoplasmic or nuclear proteins from PBMCs were prepared using NE-PER Nuclear Extraction Reagent ( Thermo scientific ) . Following SDS/PAGE , proteins were transferred onto Immobilon-P membrane ( Millipore ) . After blocking with membrane blocking solution ( Invitrogen , Life Technologies ) , the membrane was successively incubated with the anti-Elk-1 ( Santa Cruz Biotechnology ) or anti-phospho-Elk-1 ( Cell Signaling Technology ) primary antibody and the horseradish peroxidase ( HRP ) -conjugated secondary antibody ( Santa Cruz Biotechnology ) . Blots were developed using the ECLPlus Western Blotting Detection System ( GE Healthcare ) , visualized with ChemiDoc XRS imager and analyzed by Quantity One software ( BIO-RAD ) . β-actin or Lamin B was used as internal control . Allelic association tests in each ancestral group and conditional haplotype-based association tests in EA ancestry were performed by PLINK v1 . 07 software [89] under a logistic regression model adjusted for gender and the first three principal components estimated using AIMs . The Bonferroni corrected P-value threshold was adjusted to P<3 . 9×10−4 ( = 0 . 05/128 SNPs in EA ) . Pairwised LD values between SNPs and haplotypic association with SLE were evaluated using Haploview 4 . 2 [90] . The linear regression test was used to evaluate the association of SNP genotypes with IL-10 mRNA or protein levels . The Student's t-test was used to compare the mean values between two groups . A P value<0 . 05 was considered to be statistically significant . | Systemic lupus erythematosus ( SLE ) , a debilitating autoimmune disease characterized by the production of pathogenic autoantibodies , has a strong genetic basis . Variants of the IL10 gene , which encodes cytokine interleukin-10 ( IL-10 ) with known function of promoting B cell hyperactivity and autoantibody production , are associated with SLE and other autoimmune diseases , and serum IL-10 levels are elevated in SLE patients correlating with increased disease activity . In this study , to discover SLE-predisposing causal variant ( s ) , we assessed variants within the genomic region containing IL10 and its gene family member IL19 , IL20 and IL24 for association with SLE in case and control subjects from diverse ancestries . We identified SLE-associated SNP rs3122605 located at 9 . 2 kb upstream of IL10 as the most likely causal variant in subjects of European ancestry . The SLE-risk allele of rs3122605 was dose-dependently associated with elevated IL10 expression at both mRNA and protein levels in peripheral blood samples from SLE patients and controls , which could be explained , at least in part , by its preferential binding to Elk-1 , a transcription factor activated in B cells during active disease of SLE patients . Elk-1-mediated IL-10 overexpression could be downregulated by inhibiting activation of mitogen-activated protein kinases , suggesting a potential therapeutic target for SLE . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2013 | Preferential Binding to Elk-1 by SLE-Associated IL10 Risk Allele Upregulates IL10 Expression |
Tungiasis or jigger infestation is a parasitic disease caused by the female sand flea Tunga penetrans . Secondary infection of the lesions caused by this flea is common in endemic communities . This study sought to shed light on the bacterial pathogens causing secondary infections in tungiasis lesions and their susceptibility profiles to commonly prescribed antibiotics . Participants were recruited with the help of Community Health Workers . Swabs were taken from lesions which showed signs of secondary infection . Identification of suspected bacteria colonies was done by colony morphology , Gram staining , and biochemical tests . The Kirby Bauer disc diffusion test was used to determine the drug susceptibility profiles . Out of 37 participants , from whom swabs were collected , specimen were positive in 29 and 8 had no growth . From these , 10 different strains of bacteria were isolated . Two were Gram positive bacteria and they were , Staphylococcus epidermidis ( 38 . 3% ) and Staphylococcus aureus ( 21 . 3% ) . Eight were Gram negative namely Enterobacter cloacae ( 8 . 5% ) , Proteus species ( 8 . 5% ) , Klebsiellla species ( 6 . 4% ) , Aeromonas sobria ( 4 . 3% ) , Citrobacter species ( 4 . 3% ) , Proteus mirabillis ( 4 . 3% ) , Enterobacter amnigenus ( 2 . 1% ) and Klebsiella pneumoniae ( 2 . 1% ) . The methicillin resistant S . aureus ( MRSA ) isolated were also resistant to clindamycin , kanamycin , erythromycin , nalidixic acid , trimethorprim sulfamethoxazole and tetracycline . All the Gram negative and Gram positive bacteria isolates were sensitive to gentamicin and norfloxacin drugs . Results from this study confirms the presence of resistant bacteria in tungiasis lesions hence highlighting the significance of secondary infection of the lesions in endemic communties . This therefore suggests that antimicrobial susceptibility testing may be considered to guide in identification of appropriate antibiotics and treatment therapy among tungiasis patients .
Tungiasis is a parasitic disease caused by the sand flea Tunga penetrans [1] . The fleas can infest any part of the body . However majority of the cases occur on the feet [2] . Children and the elderly bear the brunt of the infection in endemic areas [3] , [4] . During the transmission period , a study that followed up individuals entering an endemic area was able to demonstrate that by the third week all the participants were infested by T . penetrans [5] . The ectoparasites can cause more than 50 lesions in a single individual in some cases [1] . This leads to severe inflammation and ulceration which is associated with intense pain . Walking , working or going to school becomes a problem . In endemic communities it’s not uncommon to find children who have dropped out of school due to the pain and stigma brough about by this condition . In some cases there is loss of toe nails and deformation of digits [6] . Secondary infection of the lesions caused by Tunga species occurs in endemic areas . A bacteriological investigation of the lesions from human tungiasis in Brazil reported isolation of various pathogenic bacteria [7] . Untreated Tungiasis is a risk factor in acquiring blood stream bacterial infections ( bacteremia ) due to broken skin . Once the jigger flea penetrates the skin , it maintains an opening ( 250 to 500μm ) in the epidermis through which it defecates , breathes and lays eggs , consequently connecting the outer surface of the skin and the blood stream as it feeds [7] , [8] ) . The exposed skin tissue is a perfect environment for bacteria to thrive . It provides warmth , moisture and nutrients , factors that are essential for microbial growth [9] . Sepsis in Tungiasis patients has been elucidated hence illustrating the medical significance of systemic infections caused by secondary bacteria infection in these patients [10] , [11] . Systemic conditions like pneumonia , meningitis , osteomyelitis , endocarditis , septicemia and Toxic shock syndrome ( TSS ) can be fatal [12] . Antibiotic resistance compounds the problem , as treatment options are greatly diminished due to bacteria developing mechanisms that neutralize available antibiotics [13] . The selection of appropriate antibiotics for treatment of severe tungiasis is critical for proper management of this parasitosis . This study therefore sought to shed light on the bacterial pathogens causing secondary infections in tungiasis lesions and their susceptibility profiles to commonly used antibiotics .
The study area was Vihiga County in Western Kenya . It is one of the most densely populated rural areas in Kenya . The population density as of 2009 , was 1 , 045 persons per square kilometre . This figure is projected to rise to 1231 persons per square kilometer in 2017 . The poverty levels are consequently very high due to the population pressure on land and other resources . The GDP per capita income was reported as US $ 1 , 103 in 2013 . Majority of the inhabitants own small uneconomical pieces of land as a result of increased subdivision occassioned by the high population and cultural practice of land inheritance . The area has two rainy seasons . Long rainy season in April , May and June and the short rains in September , October and November . The study took place during the dry and hot season from January to March , 2016 . Tungiasis peaks during the dry season . Participants were recruited with the help of Community Health Workers . Swabs were taken from lesions which clinically showed signs of secondary infection like swelling , erythema and pus . The flea was extracted with a sterile needle after disinfection of the surrounding skin with 70% alcohol for 1 minute [7] . Swabs were then collected from the surgical lesions by use of sterile cotton swabs . The cotton swabs were moistened in sterile physiological saline and gently moved in and out of the remaining cavity . One swab was used for each lesion and labeled accordingly . The swab was then transferred to a sterile transportation tube , briefly stored in an ice box and transported to the laboratory . Upon arrival at the laboratory , the swabs were cultured separately and directly onto Mannitol salt and MacConkey agar ( both from Oxoid Ltd , Basingstoke , United Kingdom ) . They were also inoculated on Brain heart infusion agar ( Oxoid Ltd , Basingstoke , United Kingdom ) supplemented with the commercially available 5% sheep blood ( TCS Biosciences , Botolph Claydon , Buckingham , United Kingdom ) and incubated aerobically at 35°C for 24 hours . Identification of suspected bacteria colonies was done by colony morphology , Gram staining , catalase , coagulase tests and biochemical tests as described previously [14] . Anaerobic bacteria were not isolated due to limited laboratory facilities at the time . The bacterial isolates were subjected to antimicrobial sensitivity testing by disk diffusion method as described by [15] . Briefly , test organisms were suspended in sterile normal saline to conform to 0 . 5 McFarland turbidity standard . With the aid of sterile cotton swab the suspended organisms were spread on Mueller-Hinton ( Oxoid Ltd , Basingstoke , United Kingdom ) Agar plate and the antibiotic disks dispensed . The plates were incubated at 37°C for 16-18h . Inhibition zone diameters were determined and recorded in Excel sheets and interpreted according to the Clinical and Laboratory Standards Institute ( CLSI ) guidelines [16] . The following panel of antibiotics ( all from Oxoid ) and their concentrations were used . For Gram negatives cefuroxime sodium 30 μg , amoxycillin\clavulanic acid 2:1 30 μg , chloramphenicol 30 μg , tetracycline 30μg , co-trimoxazole sxt 25 μg , nalidixic acid 30 μg , ampicillin 10 μg , ceftazidime 30 μg , cefotaxime 30 μg , ciprofloxacin 5 μg , norfloxacin 10μg , gentamycin 10μg , were tested . For Gram positives meropenem 10 μg , gentamycin 10μg , kanamycin 30μg , clindamycin 2μg , norfloxacin 10μg , ofloxacin 5μg , oxacillin 5μg , erythromycin 10μg , nalidixic acid 30 μg , trimethoprim-sulfamethoxazole 25μg , chloramphenicol 30 μg , tetracycline 30μg were tested . The study got approval from the KEMRI Scientific and Ethics Review Unit ( SERU ) . Approval number KEMRI/SERU/CTMDR/015/3116 . Informed written consent was obtained from all participants including children , where the guardian provided an informed consent on their behalf . All data analyzed was coded and identity of participants kept confidential . All participants were treated for tungiasis according to the National Policy Guidelines on Prevention and Control of Jigger Infestations in Kenya [17] . This was by ( removing the embedded flea with a sterile needle and disinfection of the skin lesion ) or bathing the affected area in 0 . 05% potassium permanganate for 10 minutes . A nurse who was part of the team also vaccinated the participants against tetanus . Severe cases were referred to their health center by Community Health Extension Workers from the study area , who have a functioning referral system in place .
Out of the 29 patients , 20 ( 69% ) had a single strain of bacteria whereas nine patients ( 31% ) , had more than one strain of bacteria . Most of the single strain infection were Gram positive , S . epidermidis and S . aureus; 60% and 20% respectively . The rest were Proteus species ( 10% ) , Klebsiella species ( 5% ) and E . cloacae ( 5% ) . In the polymicrobial infection , majority 66 . 7% ( 6/9 ) were S . aureus and S . epidermidis ( 2/9 ) ( 22 . 2% ) and their combinations as shown in Table 1 . Up to 17 . 2% had two different types of bacteria , three and four polymicrobial infections had 3 . 4% each and two patients ( 6 . 9% ) had five different bacteria . Once identified , bacterial isolates were further tested for drug sensitivity to commonly prescribed drugs using the Kirby Bauer disk diffusion method . Eleven drugs were used to test susceptibility of the Gram negative isolates . All the isolates were sensitive to ciprofloxacin , cefotaxime , norfloxacin , gentamicin , nalidixic acid , chloramphenicol ( Table 2 ) . Ampicillin had the highest resistance of 52 . 6% . All the Gram negative bacterial isolates were resistant to atleast one or more drugs except E . amnigenus . Enterobacter cloacae showed resistance to ampicillin ( 75 . 0% ) , amoxicillin clavulanic acid ( 25 . 0% ) tetracycline ( 50 . 0% ) , ceftazidime ( 25 . 0% ) and cefuroxime ( 25 . 0% ) . Citrobacter species showed resistance to ampicillin ( 100 . 0% ) , amoxicillin clavulanic acid ( 100 . 0% ) and cefuroxime ( 50 . 0% ) . The two proteus species had resistance to tetracycline; 50 . 0% and 75 . 0% respectively ( Table 2 ) . All the Gram positive were sensitive to norfloxacin , ofloxacin , meropenem and gentamicin drugs . However they were resistant to clindamycin , kanamycin , oxacillin , erythromycin , nalidixic acid , trimethorprim sulfamethoxazole , chloramphenicol and tetracycline ( Table 3 ) . nalidixic acid had the highest bacterial resistance ( 31 . 0% ) followed by clindamycin ( 20 . 7% ) . Three patients ( 10 . 3% ) had S . aureus isolates that were methicillin resistant ( MRSA ) . Both Gram negative and Gram positive bacteria isolates were sensitive to Gentamicin and Norfloxacin drugs .
The findings from this study confirm the presence of resistant bacteria in tungiasis lesions hence highlighting the significance of secondary infection of the lesions in endemic communties . This therefore implies that the treatment regimen for tungiasis especially in severe cases should be expanded to include antibiotics . Antimicrobial susceptibility testing may be considered to guide in identification of appropriate antibiotics . Norfloxacin and gentamicin have shown to be very effective against both Gram negative and Gram positive bacteria . In severe tungiasis where sepsis is observed , a broad spectrum drug may be considered at the onset to avoid delay in starting treatment as results from cultures are awaited . | Secondary bacterial infection of tungiaisis lesions is a threat to jigger infested patients . Once the flea penetrates the skin , it leaves an opening on the skin through which it lays eggs , defecates and breathes throughout it’s life cycle on the host . At the same time , it feeds on the host’s blood hence a direct connection of the environment to the blood stream is established . Chances of bacteria getting into the blood stream is therefore greatly enhanced . Once bacteria gets into the blood , it can lead to life threatening conditions like septicemia , meningitis , pneumonia and toxic shock syndrome . Affected communities are oblivious to this danger and they neither seek nor get treated for this parasitosis . The health officials and scientific community also continue to ignore this disease hence it’s extremely neglected . Consequently , not much is understood concerning the disease dynamics , distribution and pathogenesis . This study therefore is a deliberate effort to bring attention to one aspect of the disease dynamics of this menace . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] | [
"bacteriology",
"antimicrobials",
"gram",
"positive",
"bacteria",
"gram",
"negative",
"bacteria",
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"drugs",
"microbiology",
"animals",
"staphylococcus",
"aureus",
"methicillin-resistant",
"staphylococcus",
"aureus",
"antibiotic",
"resistance",
"signs",
"and",
"symptoms",
"antibiotics",
"fleas",
"pharmacology",
"bacteria",
"bacterial",
"pathogens",
"antimicrobial",
"resistance",
"staphylococcus",
"medical",
"microbiology",
"lesions",
"microbial",
"pathogens",
"insects",
"arthropoda",
"eukaryota",
"diagnostic",
"medicine",
"microbial",
"control",
"biology",
"and",
"life",
"sciences",
"organisms",
"staphylococcus",
"epidermidis"
] | 2017 | Secondary bacterial infections and antibiotic resistance among tungiasis patients in Western, Kenya |
Diarrheal diseases are among the most frequent causes of morbidity and mortality in children worldwide , especially in resource-poor areas . This case-control study assessed the associations between gastrointestinal infections and diarrhea in children from rural Ghana . Stool samples were collected from 548 children with diarrhea and from 686 without gastrointestinal symptoms visiting a hospital from 2007–2008 . Samples were analyzed by microscopy and molecular methods . The organisms most frequently detected in symptomatic cases were Giardia lamblia , Shigella spp . / enteroinvasive Escherichia coli ( EIEC ) , and Campylobacter jejuni . Infections with rotavirus ( adjusted odds ratio [aOR] = 8 . 4; 95% confidence interval [CI]: 4 . 3–16 . 6 ) , C . parvum/hominis ( aOR = 2 . 7; 95% CI: 1 . 4–5 . 2 ) and norovirus ( aOR = 2 . 0; 95%CI: 1 . 3–3 . 0 ) showed the strongest association with diarrhea . The highest attributable fractions ( AF ) for diarrhea were estimated for rotavirus ( AF = 14 . 3%; 95% CI: 10 . 9–17 . 5% ) , Shigella spp . /EIEC ( AF = 10 . 5%; 95% CI: 3 . 5–17 . 1% ) , and norovirus ( AF = 8 . 2%; 95% CI 3 . 2–12 . 9% ) . Co-infections occurred frequently and most infections presented themselves independently of other infections . However , infections with E . dispar , C . jejuni , and norovirus were observed more often in the presence of G . lamblia . Diarrheal diseases in children from a rural area in sub-Saharan Africa are mainly due to infections with rotavirus , Shigella spp . /EIEC , and norovirus . These associations are strongly age-dependent , which should be considered when diagnosing causes of diarrhea . The presented results are informative for both clinicians treating gastrointestinal infections as well as public health experts designing control programs against diarrheal diseases .
Diarrheal diseases are the second leading cause of childhood mortality worldwide . In 2010 , diarrhea was responsible for 0 . 8 million deaths of children below the age of five years , accounting for 10 . 5% of all deaths within that age group [1] . Mortality and morbidity patterns differ across geographical regions , with 78% of all pediatric diarrhea-associated deaths occurring in the African and South-East Asian World Health Organization ( WHO ) Regions [2] . The etiology of diarrhea is often not completely understood , especially in developing countries , including those in sub-Saharan Africa . Knowledge of the distribution and impact of infectious agents in diarrheal diseases is crucial in guiding empirical medical treatment and in designing prevention programs . However , many studies on the epidemiology of gastrointestinal infections are restricted to only patients with diarrhea , ignoring the possibility that infections may progress asymptomatically or even influence one another . This may be of particular importance in areas in which certain infectious agents are endemic , which would results in a high probability of ongoing infections after the development of partial immunity and/or tolerance . Thus , an adequate control group is essential to determine the pathogenicity of infectious agents , their fractions attributable to gastrointestinal symptoms ( GIS ) , and the age-dependent association of infectious agents with GIS [3 , 4] . This hospital-based case-control study in a rural area of Ghana was designed to analyze gastrointestinal infections in children with and without diarrhea . The aims of this study were ( i ) to identify the causative pathogens linked to diarrhea , ( ii ) to describe their pathogenicity and contribution to the burden of diarrhea , and ( iii ) to analyze the frequency and interactive effects of co-infections .
The Committee on Human Research , Publications and Ethics , School of Medical Science , Kwame Nkrumah University of Science and Technology , Kumasi , Ghana , approved the study design and the informed consent procedure . All participants were informed of the study’s purpose and procedures . Written informed consent was obtained from the parents or the legal guardian on behalf of the study children prior to study enrolment . Non-participation had no effect on the medical treatment provided . Stool samples were refrigerated ( 4°C ) immediately after collection and transported within a day to the laboratories at the Kumasi Centre for Collaborative Research in Tropical Medicine ( KCCR ) . Sample transport took about 1 . 5 hours and a cool-box was used to maintain the cold chain . Upon arrival stool samples were aliquoted ( 3 x 0 . 2 mg ) and frozen at -20°C while the reaming sample was processed for further analyses . Microscopy of blood and stool samples took place at the KCCR laboratory and DNA extractions , polymerase chain reaction ( PCR ) assays and ELISA-tests were conducted at the Bernhard Nocht Institute for Tropical Medicine ( BNITM ) in Germany . Therefore , frozen stool samples were transported on dry ice to the BNITM maintaining a temperature of -20°C ( 3 ) . All samples were handled , stored and transported according to good laboratory practice . Sodium acetate-acetic acid-formalin ( SAF ) solution was added to the sample to preserve parasites . Following formalin-ether concentration of protozoan cysts and trophozoites , as well as helminth eggs , fractions of the concentrated material were stained with iodine or by the modified acid-fast method , the latter to improve detection of Coccidia species , such as Cryptosporidium , Cyclospora , and Cystoisospora [8 , 9] . The samples were then viewed under a light microscope . For molecular detection , DNA was extracted from the frozen stool samples using QIAamp DNA Stool Kits ( Qiagen , Hilden , Germany ) . Specific sequences were amplified by PCR to identify the following organisms: Campylobacter jejuni [10] , Cryptosporidium parvum/hominis [11] , Cyclospora cayetanensis [12 , 13] , Entamoeba dispar [14 , 15] , Entamoeba histolytica [14 , 15] , Giardia lamblia [11] , norovirus [16] , Salmonella enterica [10] , Shigella spp . /enteroinvasive Escherichia coli ( EIEC ) [10] , and Yersinia enterocolitica [10] . Diagnostic PCR assays give the Ct value as a semi-quantitative result and a cut-off of 35 cycles was applied to determine a positive test result . As Shigella spp . and EIEC both possess the ipaH gene , these organisms cannot be distinguished by PCR [10] . Further , the PCRs applied did not allow differentiation between typhoidal and non-typhoidal Salmonella or between norovirus genotypes . Rotavirus was identified by an enzyme-linked immunosorbent assay ( ELISA; Ridascreen ) . Blood samples were obtained from all participants by finger prick . Thick and thin smears were prepared from these blood samples , which were stained and examined by microscopy . Malaria was defined as any asexual parasitaemia with body temperature >38°C . All PCR reactions , as well as malaria and stool microscopy , were regularly evaluated by external and internal quality assessments . The study personal was trained to adhere to the standard operational procedures for each laboratory method . A sample size of about 500 children per group was estimated to identify isolates in 5% of the controls and 10% of the cases , considering an alpha-level of 5% and a power of 80% . Categorical variables are reported as frequencies and percentages , whereas continuous variables are reported as means ± standard deviations ( SDs ) or as medians with interquartile ranges ( IQRs ) . Missing values were excluded from analyses , thus the denominators for some comparisons differ . Because microscopy has decreased sensitivity and specificity in diagnosing diarrheal samples [8 , 17] , all statistical comparisons were based on PCR- or ELISA-based diagnoses . The associations between diarrhea and gastrointestinal infections were determined by calculating odds ratios ( OR ) and 95% confidence intervals ( CI ) . Subjects were stratified to show effects within categories of a third variable to assess and account for confounding or effect modification . Mantel-Haenszel adjusted ORs ( aOR ) were calculated from the stratified analyses . The attributable fractions ( AF ) on the diarrhea burden , defined as the proportion of diarrhea attributable to a certain pathogen , were calculated as described [18] , from logistic regression estimates , including dummy variables , for age categories . Heterogeneity in the occurrence of co-infections was assessed by comparing the probability of organism A in the presence of organism B over the probability of A in the absence of B . These associations were determined by calculating the risk ratio ( RR ) , using the formula RR = P ( A|B = 1 ) /P ( A|B = 0 ) , in which a value of about one indicates independence and a value different from one indicating dependence . RRs were calculated for the total study group and for cases and controls separately . Organisms diagnosed by PCR or ELISA and detected in more than 5% of stool samples in the respective study groups were included in this calculation to ensure a sufficient number of co-infections . Age-adjusted RRs ( aRR ) were calculated to account for the age-dependence of infections . Age was categorized into the groups 0–<1 , 1–<2 , 2–<5 and 5–15 years to analyze age specific infection dynamics . All data analyses were performed with STATA 12 ( StataCorp LP , College Station , USA ) .
Potentially pathogenic organisms as well as facultative and non-pathogenic parasites were detected 1 , 843 times in 915 ( 79 . 5% ) stool samples . The most frequent infections were with G . lamblia ( n = 470; 38 . 1% ) , Shigella spp . /EIEC ( n = 336; 27 . 2% ) , C . jejuni ( n = 242; 19 . 6% ) , Blastocystis hominis ( n = 144; 14 . 3% ) , and norovirus ( n = 139; 11 . 3% ) . Sex-dependent differences were not observed . All parasites and protozoa , apart from C . parvum/hominis , tended to be less frequently observed in diarrheal compared to non-diarrheal samples . Cyclospora cayetanensis , E . histolytica , and Yersinia spp . were not detected in any of the stool samples tested ( Table 2 ) . Fig 1 shows the age-stratified proportions and the median ages of case and control children infected with particular organisms . In children with diarrhea , rotavirus , norovirus , and C . parvum/hominis were most frequently observed in younger infants , with infected individuals having median ages of 12 months ( IQR: 8–23 months ) , 13 months ( IQR 8–22 months ) , and 14 months ( IQR: 10–20 months ) , respectively . By contrast , control children with these organisms were older , with median ages of 30 months ( IQR: 16–45 months ) , 40 months ( IQR: 20–82 months ) , and 23 months ( IQR: 18–28 months ) , respectively . For all other infections the median age in children with diarrhea ranged from 19 months ( IQR: 11–36 months ) to 68 months ( IQR: 35–123 months ) . Infants below 6 months accounted for 95 ( 8 . 1% ) hospital visits , with 16 stool samples ( 16 . 9% ) diagnosed with rotavirus and 13 ( 13 . 7% ) with norovirus . All virus-infected children , except one diagnosed with norovirus , had diarrhea . Within this age group other organisms were not detected in more than six stool samples ( 6 . 3% ) . Crude analyses showed that the strongest positive associations with diarrhea were for infections with rotavirus ( OR = 11 . 9; 95% CI: 6 . 2–24 . 9 ) , C . parvum/hominis ( OR = 4 . 3; 95% CI: 2 . 3–8 . 6 ) , and norovirus ( OR = 2 . 6; 95% CI: 1 . 8–3 . 9 ) . Inverse associations were found for G . lamblia and E . dispar suggesting that more infections were diagnosed in controls than in cases . However , this effect was attenuated after age stratification ( Table 3 ) . Age stratification revealed varying associations of most infections with diarrhea . For example , associations with S . enterica or Shigella spp . /EIEC increased with age , and associations with norovirus were highest in the youngest age group ( OR = 4 . 7; 95% CI: 1 . 4–24 . 6 ) and lower in older children . Associations between rotavirus and diarrhea were strong amongst all age groups , although the frequency of infections decreased with age . Stratified estimates for C . parvum/hominis were lower than the crude OR , indicating that age confounded effect estimates . No associations between sex ( being female ) and diarrhea ( OR = 0 . 9; 95% CI: 0 . 7–1 . 1 ) or rainy season and diarrhea ( OR = 1 . 1; 95% CI: 0 . 9–1 . 4 ) were observed . Also , in logistic regression models these factors did not alter the association between infections and diarrheal symptoms . Infections manifesting at younger ages tended to be more strongly associated with diarrhea . For example , the highest ORs were observed for rotavirus , C . parvum/hominis and norovirus , infectious agents most frequently diagnosed in the younger case age groups , with median ages of 12 months ( IQR: 8–23 months ) , 14 months ( IQR: 10–20 months ) , and 12 months ( IQR: 8–23 months ) , respectively . By contrast , E . dispar and G . lamblia were not positively associated with diarrhea and were more frequently diagnosed in older cases , with median ages around 38 months ( IQR: 35–57 months ) and 28 months ( IQR: 16–46 months ) , respectively . The highest AFs ( proportion of diarrheal symptoms attributable to a certain organism ) were observed for rotavirus ( AF = 14 . 3%; 95% CI: 10 . 9–17 . 5% ) , Shigella spp . /EIEC ( AF = 10 . 5%; 95% CI: 3 . 5–17 . 1% ) , and norovirus ( AF = 8 . 2%; 95% CI: 3 . 2–12 . 9% ) , whereas all other infections had AFs of about 5% and lower ( Fig 2 ) . Probabilities for the occurrence of distinct pairs of infectious agents were calculated for organisms detected by PCR or ELISA in more than 5% of the stool samples ( Table 4 ) . Considering the total study group , most of the infections occurred independently of other organisms . However , E . dispar , C . jejuni , and norovirus were observed more often in the presence of G . lamblia , showing rate ratios of 1 . 6 ( 95% CI: 1 . 3–1 . 9 ) , 1 . 3 ( 95% CI: 1 . 2–1 . 6 ) , and 1 . 3 ( 95% CI: 1 . 1–1 . 6 ) , respectively . These , estimates were comparable between cases and controls .
The most important cause of diarrheal disease was rotavirus , with both the highest AF and the largest risk for diarrhea across all age groups . Frequency of rotavirus infections decreased with age , but its association with diarrhea was nearly constant throughout all age groups . Similarly , a multi-center study performed in seven resource-poor countries found that rotavirus was the leading cause of diarrhea in infants , with age-dependent AFs between 16% and 28% [3] . Furthermore , 96% of children in a Mexican birth cohort were infected with rotavirus at least once by the age of 2 years . Rotavirus infections conferred protection against re-infection , resulting in less frequent and less severe manifestations in older children [19] . The proportion of norovirus infections among cases and controls was , with 16 . 6% and 6 . 8% , respectively , comparable to figures from other high-mortality developing countries , where 14% ( CI: 11–16 ) and 7% ( CI: -2–16 ) are reported , respectively [20] . In children with diarrhea the frequency of norovirus infections was similar to that of rotavirus . However , compared to rotavirus the amount of norovirus infections was higher in asymptomatic ( control ) individuals , most probably due to increasing pathogen tolerance and limited sterilizing immunity . A transmission model showed that , in highly endemic settings , protection against severe norovirus gastroenteritis could be acquired early in life , resulting in frequent asymptomatic re-infections [21] . Likewise , our study found a significant association between norovirus infection and diarrhea in infants , whereas older children were more likely to be asymptomatic carriers . In industrialized countries , cryptosporidiosis is primarily an opportunistic infection in HIV/AIDS patients and a major cause of water-born outbreaks reported from several countries [22] . In West Africa , however , cryptosporidiosis is the cause of diarrhea in 4 . 9% to 14 . 7% of immunocompetent children , depending on age and geographical location [3 , 23] . Similarly , our results showed that C . parvum/hominis infections were strongly associated with diarrhea throughout all age groups and was present in more than 10% of symptomatic children below the age of 2 years . Asymptomatic cryptosporidium carriers were not observed in this age group and rarely seen in older children . A review of cryptosporidiosis in sub-Saharan Africa reported the same age distribution , with a peak amongst children aged 6–12 months . Apparently , infection can occur throughout childhood , but symptoms become less severe with age [24] . Shigella spp . /EIEC was , after G . lamblia , the second most frequent pathogen identified in children with diarrhea , increasingly occurring in older children . However , comparing our findings with other studies is challenging because our test was PCR based . Traditionally , diagnosing shigellosis relies on culturing techniques , which selectively isolates the pathogen , followed by a biochemical identification of one of the four Shigella species [25] . Introducing PCRs techniques has their benefits but the gene sequence used for the diagnosis is also carried by enteroinvasive Escherichia coli ( EIEC ) [10] . Technically , our study cannot differentiate between Shigellosis and EIEC , which , in addition to the greater sensitivity of the PCR based test , might also be mirrored by the higher numbers of infections presented . Thus , despite its moderate association with diarrhea , the high prevalence of this pathogen group led to the second highest AF observed in our study . Generally , the used PCR methods have a higher test sensitivity compared to conventional culture methods . This improves the ability to diagnose organisms in a stool sample . However , a drawback seems to be an increase in asymptomatic detections overall [26] . Since this affects diagnostics in both cases and controls the calculated ORs should not be affected . However , the estimated disease prevalence , and consequently the estimated AF , might be overestimated . Further studies using quantitative approaches [27] are needed to establish and improve diagnostic analyses for gastrointestinal diseases in low- and middle-income countries . The burden of G . lamblia infections was quite high . Interestingly , the frequency of G . lamblia infections was lower in children with diarrhea than in asymptomatic carriers . A systematic review of the impact of G . lamblia on diarrhea highlighted that , although most studies show no or inverse effects , some studies report positive associations in children aged around 1 year , presumably as a response to initial G . lamblia infections [28] . However , the statistical power of the current study did not allow to disentangle such age-effects . In both cases as well as controls high numbers of multiple infections were observed . Most co-infections were identified as statistically expected , although G . lamblia was more often found together with E . dispar , C . jejuni , and norovirus . G . lamblia has been reported to induce apoptosis of epithelial cells leading to increased epithelial permeability [29] . Further , G . lamblia was found to secrete proteins capable of impairing the innate immune response [30] . Alternatively , co-infections may be due to shared transmission routes . However , since most gastrointestinal organisms are transmitted via the fecal-oral route , it is unlikely that this alone explains the association among co-infections . Interestingly , a recent pooled case-control study from Ecuador also identified mechanistic interactions for diarrhea symptoms between rotavirus and G . lamblia as well as between rotavirus and Escherichia coli [31] . In vitro models have indicated that rotavirus may foster the adhesion , invasion , and multiplication of bacteria in enteric cells , mechanisms that may explain these synergistic effects [32–34] . Generally , the role of co-infections in diarrheal diseases deserves more attention in order to identify the associations between infections as well as interactions with GIS . Furthermore , strategies to identify causative pathogens in the presence of multiple infections are needed since diagnosing a particular isolate may not rule out other potential infectious causes in diarrheal disease . The study presented here has several limitations , therefore the results should be interpreted with caution . Cases as well as controls were selected at a hospital OPD , thus the control group does not consist of healthy individuals . These are children seeking help for other health conditions that may increase the risk for gastro-intestinal infections . We have little background information on the total eligible study group , i . e . , children that visited the OPD during the study period , from which we selected cases and controls . Thus , we cannot judge how well characteristics of cases and controls match . Table 1 highlights differences between cases and controls . For example , falciparum malaria is more frequently observed in controls , because controls need an alternative reason to attend the hospital , which is malaria in some attendees . Controls are also more likely to have a full vaccination schedule , which can be explained by age differences as well as due to possible differences in socio-economic status . In the analyses these factors cannot be controlled for , however , we believe that they do not act as a confounder . Even though some factors are associated with gastrointestinal infections , they are not associated with diarrheal symptoms , which would be needed to qualify as a confounder . Controls had to be diarrhea free at the point of study enrollment , yet GIS before enrollment were not assessed . Thus , controls could be carriers of pathogens if infections occurred before study enrollment . For instance , norovirus can be found in stool for up to 60 days after infection [35] . In case some controls are pathogen carriers due to recent infections study results would underestimate the actual diarrheal association . Two preconditions need to be fulfilled to generalize AFs from case-control data: ( i ) the case-control selection must be representative of the source population and ( ii ) the OR must be a robust estimator of the RR . Cases in this study were recruited from children in the OPD , making this a select group of patients seeking professional care , thereby representing individuals with moderate to severe diarrheal disease . Considering our case-control sampling approach , the OR would approximate the RR only if the rare disease assumption is fulfilled . This , however , was applicable to all infections studied , especially not for Shigella spp . /EIEC , C . jejuni , and norovirus . In these cases , the OR is likely to overestimate the true RR , resulting in a higher AF . Several possible infectious causes of diarrhea were not detected by these methods , including adenovirus [36] enteropathogenic Escherichia coli [37] and enterotoxigenic Escherichia coli [37 , 38] . The AFs express the proportion of diarrheal disease that would be reduced if an organism could be removed . This measure is highly relevant to public health concerns since it demonstrates the potential effects of disease prevention and control as well as empirical disease treatment measures . In particular , it highlights the potential roles of vaccinations against rotavirus and norovirus in sub-Saharan Africa , as well as of water purification , sanitation , and hygiene measures; effective options that can reduce the burden of diarrheal diseases [39] . | Gastrointestinal infections are frequent in many low-income countries . However , their role in diarrheal diseases is still under discussion . Many epidemiological studies focus on individuals with diarrheal symptoms only , ignoring the fact that infections may progress asymptomatically as well . In order to identify infectious agents associated with diarrhea it is imperative to consider cases without symptoms as a control group . We conducted a case-control study , including 548 children with diarrhea and 651 children without gastrointestinal symptoms in order to untangle the role of gastrointestinal infections in diarrheal disease . As shown in other studies infections with rotavirus , Shigella spp . /EIEC and norovirus are responsible for the main diarrhea burden . Co-infections are frequently observed in our study group and some organisms occur more frequently in the presence of a second one . Especially Giardia lamblia , which is not associated with diarrhea , is more often observed along with Campylobacter jejuni and norovirus , which are responsible for a high number of diarrheal episodes . This may be of particular interest since G . lamblia is , with a frequency of 40% within the study group , the most prevalent organism observed . Furthermore , the high number of co-infections challenged the identification of causative pathogens since diagnosing a particular isolate may not rule out the effect of another potentially infectious agent in diarrheal disease . We observed a strong effect of age on the course of an infection , which may guide clinicians when diagnosing causes of diarrhea . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | Gastrointestinal Infections and Diarrheal Disease in Ghanaian Infants and Children: An Outpatient Case-Control Study |
Histones package DNA and regulate epigenetic states . For the latter , probably the most important histone is H3 . Mammals have three near-identical H3 isoforms: canonical H3 . 1 and H3 . 2 , and the replication-independent variant H3 . 3 . This variant can accumulate in slowly dividing somatic cells , replacing canonical H3 . Some replication-independent histones , through their ability to incorporate outside S-phase , are functionally important in the very slowly dividing mammalian germ line . Much remains to be learned of H3 . 3 functions in germ cell development . Histone H3 . 3 presents a unique genetic paradigm in that two conventional intron-containing genes encode the identical protein . Here , we present a comprehensive analysis of the developmental effects of null mutations in each of these genes . H3f3a mutants were viable to adulthood . Females were fertile , while males were subfertile with dysmorphic spermatozoa . H3f3b mutants were growth-deficient , dying at birth . H3f3b heterozygotes were also growth-deficient , with males being sterile because of arrest of round spermatids . This sterility was not accompanied by abnormalities in sex chromosome inactivation in meiosis I . Conditional ablation of H3f3b at the beginning of folliculogenesis resulted in zygote cleavage failure , establishing H3f3b as a maternal-effect gene , and revealing a requirement for H3 . 3 in the first mitosis . Simultaneous ablation of H3f3a and H3f3b in folliculogenesis resulted in early primary oocyte death , demonstrating a crucial role for H3 . 3 in oogenesis . These findings reveal a heavy reliance on H3 . 3 for growth , gametogenesis , and fertilization , identifying developmental processes that are particularly susceptible to H3 . 3 deficiency . They also reveal partial redundancy in function of H3f3a and H3f3b , with the latter gene being generally the most important .
The core histones H2A , H2B , H3 and H4 , package DNA into nucleosomes and modify higher-order chromatin structure . H3 in mammals is represented by three near-identical 135 amino acid isoforms: canonical H3 . 1 and H3 . 2 , and the variant H3 . 3 . Chromatin incorporation is replication-coupled for the canonical forms , and is replication-independent for H3 . 3 [1–4] . H3 . 3 plays fundamental roles in regulating genome function and stability . It can accumulate in terminally differentiated or non-dividing cells , and can become the predominant H3 isoform [3 , 5] . Through the chaperone HIRA ( histone cell cycle regulation defective homolog A ( S . cerevisiae ) ) , H3 . 3 is deposited at active genes and regulatory regions , forming more accessible nucleosomes [3 , 4 , 6–8] , and underlies the formation of epigenetic bivalent domains at the promoter regions of poised genes in embryonic stem ( ES ) cells [8 , 9] . Through the chaperones ATRX ( alpha thalassemia/mental retardation syndrome X-linked ) and DAXX ( Fas death domain-associated protein ) , H3 . 3 is deposited at repetitive regions such as telomeres and centromeres , serving genome stabilization functions [10–12] . H3 . 3 is required for heterochromatin formation in mammalian preimplantation-stage embryos [13 , 14] , and has been reported to preferentially localize to the heterochromatic XY-body in meiosis-I , where it could serve functions related to meiotic sex chromosome inactivation ( MSCI ) [15] . Finally , specific amino acid substitutions in H3 . 3 drive cancer [16 , 17] , underscoring the fundamental importance of H3 . 3 in regulating epigenetic states . Canonical histones are encoded by numerous clustered intronless genes , while histone variants are typically encoded by one intron-containing gene . H3 . 3 presents a unique genetic paradigm in that the identical protein is encoded by two intron-containing unlinked genes , H3f3a , and H3f3b ( histone 3 , family 3A and -B ) . We will refer to the encoded proteins as H3 . 3A and H3 . 3B , respectively . The two genes have divergent regulatory and intervening regions , suggesting some qualitatively different functional roles . The presence of two H3 . 3-encoding genes , and their genomic arrangement , is highly conserved in mammals and birds [18] . Here , we have investigated the developmental effects of null mutations of H3f3a and H3f3b in the mouse ( Mus musculus ) , each mutated by the same experimental strategy , and in the identical mouse strain—129S1/SvImJ ( 129S1 ) . H3f3a mutants were viable to adulthood . By contrast , H3f3b mutants were growth deficient , and inviable at birth . Failures in gametogenesis and fertilization were seen , with the defects severer on ablation of H3f3b . Our results demonstrate important functions of H3 . 3 in growth , viability , and fertility , and reveal stages of gametogenesis and fertilization that are adversely affected by H3 . 3 deficiency in the mouse .
H3f3a+/- heterozygous young were produced by mating ES cell chimaeras to strain 129S1 Cre-deleter females [19] ( see Materials and Methods ) . H3f3a+/- females and males were overtly normal and fertile . Timed matings were obtained from H3f3a+/- ♀ × H3f3a+/- ♂ intercrosses so that newborns could be monitored . In 14 litters in which 65 pups were born , only four dead pups were seen: two of these dead pups were H3f3a-/- . All remaining young survived to adulthood , except two H3f3a-/- runts that were culled . From all timed pregnancies , and other intercrosses , 29 , 84 and 46 animals surviving to adulthood were H3f3a-/- mutant , H3f3a+/- and H3f3a+/+ ( WT ) , respectively—not significantly different from a 1:2:1 ratio ( P>0 . 05 , chi-squared test ) . H3f3a-/- females were of normal size . By contrast , H3f3a-/- males were slightly smaller than H3f3a+/- and WT males at 3 wk and 6 wk ( Fig . 1A ) . H3f3a-/- mutants of both sexes were overtly normal in behaviour and appearance . H3f3b+/- young were bred by mating ES cell chimaeras to strain 129S1 Cre-deleter females ( see Materials and Methods ) . These young often appeared smaller than WT sibs at weaning , but developed into robust adults . H3f3b+/- females were fertile on mating with WT 129S1 males . From these matings , altricial and weaned H3f3b-/+ young were significantly smaller than WT littermates , revealing haploinsufficiency for H3f3b ( Fig . 1B ) . H3f3b-/- mutants could not be bred in intercrosses because H3f3b-/+ males were sterile , with small testes ( Fig . 1C ) . To produce H3f3b-/- zygotes , we again made use of the 129S1 Cre-deleter line . The cre cds in this line is X-linked ( Xcre ) . Floxed alleles are always mutated on sperm entry into Cre-containing oocytes when XXcre females are used [19] . Two sequential matings were performed . First , H3f3b-/+ , XXcre females were bred in H3f3bc/c , XX ♀ × WT , XcreY ♂ matings—these females were converted to the H3f3b-/+ genotype early in development because of the presence of Xcre . Second , H3f3b-/+ , XXcre ♀ × H3f3bc/c , XY ♂ matings were performed to obtain timed pregnancies . The paternal H3f3bc allele was converted to H3f3b- in the zygote in these matings . Therefore , a 1:1 ratio of H3f3b-/- to H3f3b+/- zygotes was expected . Results of fetal recovery are shown ( Table 1 ) . Overall , 37 H3f3b-/- and 69 H3f3b+/- fetuses were obtained , different from a 1:1 ratio . Thirty-six deciduae were counted in these pregnancies , these being implantation sites in which development did not proceed beyond the early implantation stage . It is possible that most of these deciduae represented failures in H3f3b-/- development , accounting for the fetal deficit . This failure was not associated with sex chromosome constitution or to the presence of Xcre ( Table 1 ) . Twenty four percent ( 9/37 ) of H3f3b-/- fetuses at 13½-18½ dpc were dead or severely abnormal ( Table 1 ) . Live 18½ dpc H3f3b-/- fetuses , two days before birth , had no gross abnormalities ( Fig . 1D ) . However , no live H3f3b-/- pups were seen in newborn litters . Thirty two H3f3b+/- animals were born , 29 of which survived to weaning . A single H3f3b-/- dead pup was found on the day of birth . These results strongly suggest that all H3f3b-/- pups died at birth , then were consumed by the mothers . The immediate death of pups at birth is usually associated with respiratory failure [20] . Examining 18½ dpc H3f3b-/- fetuses histologically revealed no gross pathology in any tissues , including the lungs ( S1 Fig . ) . To confirm that H3f3b-/- fetuses could not survive parturition , fetuses were delivered by Caesarean at 19½ dpc . From 5 pregnancies , of 11 H3f3b+/- fetuses delivered , all initiated normal respiration within a few minutes , remaining oxygenated or ‘pink’ , and mobile . By contrast , all 6 H3f3b-/- fetuses delivered failed to initiate respiration after initially gulping . Within a few minutes they were immobile , and ‘blue’ ( Fig . 1E ) . They were also smaller than H3f3b+/- littermates ( Fig . 1F ) . H3f3a-/- mutants contain residual H3 . 3B , while H3f3b-/- mutants contain residual H3 . 3A . These amounts of residual H3 . 3 were determined in immunoblots in the brain , kidney , liver and lung of 18½ dpc fetuses , and in the trunk of 13½ dpc fetuses ( Fig . 2 ) . In both mutants , H3 . 3 was detected in all organs , consistent with H3 . 3A and H3 . 3B being essentially ubiquitous in the developing fetus . Given that H3f3b-/- mutants had the severer phenotype , we expected to see relatively less residual H3 . 3 in these animals . However , the only clear difference between the two mutants was in the brain , where H3f3b-/- animals had relatively higher residual H3 . 3A ( Fig . 2 ) . Three H3f3a-/- males were fertility-tested alongside three WT littermates within the first three months after reaching adulthood . Each male was caged for one week with two nulliparous 8–10 wk outbred Swiss females , this being repeated over several weeks . For H3f3a-/- males , 7 pregnancies from 34 female exposures resulted , less than the frequency for WT males—16 pregnancies from 28 exposures ( P<0 . 001 , Fisher’s exact test ) . Also , litters sired by H3f3a-/- males were small: WT , 10 . 63 ± 2 . 55 ( 16 ) ; H3f3a-/- , 6 . 14 ± 4 . 85 ( 7 ) [mean ± s . d . ( n ) ] ( P<0 . 01 , t-test ) . These three littermate pairs , and two additional pairs , were then fertility-tested at 6–10 months of age . Over this period , each male was caged with two Swiss females . All WT males consistently sired large litters , while three H3f3a-/- males sired no litters , and two H3f3a-/- males sired the occasional small litter . At the end of this period , the ability of H3f3a-/- males to copulate was tested . Vaginal plugs , at ½ days post coitum ( dpc ) , were readily obtained when H3f3a-/- males were mated with nulliparous Swiss females . Oviducts were flushed at 1½ dpc . H3f3a-/- males fertilized substantially fewer ova than WT littermates , indicating that the subfertility was caused by a defect in spermatogenesis ( Table 2 ) . The reproductive tracts were overtly normal , except the vasa deferentia appeared to be devoid of sperm . Representative histological sections of the testes and caudae epididymides are shown ( Fig . 3 A–E ) . A substantial production of sperm was seen in H3f3a-/- males , although some sloughed germ cells were present in the caudae , indicating developmental failure of some spermatids ( Fig . 3D , E ) . Sperm isolated from the caudae showed reduced motility , failing to efficiently swim out in media . The large majority of spermatozoa possessed tail defects , with some possessing head defects ( Figs . 3F , 4 ) . These data are consistent with a requirement for H3 . 3A in spermiogenesis . H3f3b heterozygous males derived from chimaeras mated to Cre-deleter females , or from the H3f3b+/- ♀ × WT ♂ backcross , were sterile . Their testes were less than one third of normal weight , indicative of failed spermatogenesis ( Fig . 1C ) . Histological analysis revealed a reduction of round spermatids and very few elongating spermatids in the lumen of seminiferous tubules ( Fig . 3G , H ) . The caudae epididymides were devoid of sperm , and often contained sloughed germ cells ( Fig . 3I ) . H3 . 3 has been reported to preferentially localize to the XY-body , where it may play a role in MSCI [15] . Therefore , we thought there may a level of MSCI failure in H3f3b-/+ males . MSCI acts as a checkpoint . Its failure is associated with autosomal asynapsis , leads to pachytene apoptosis with no progression of spermatocytes beyond meiosis-I [21 , 22] . The presence of pachytene spermatocytes with XY-bodies , and round spermatids in H3f3b-/+ sterile males ( Fig . 3H ) , indicates that there is no serious failure of MSCI . Yet , we decided to investigate autosomal synapsis in H3f3b-/+ males more directly . First , we examined the localization of γH2A . X ( H2A histone family , member X ) in mid- to late-pachytene spermatocytes of H3f3b-/+ males , when autosomes should be fully synapsed . These stages were marked by positive staining for H1FNT ( H1 histone family , member N , testis-specific ) [23] . All H3f3b-/+ H1FNT-positive cells examined ( >50 ) displayed discrete localization of γH2A . X to the presumptive XY-body , with no labelling evident outside this region , indicating that autosomal synapsis and MSCI proceeded normally ( Fig . 5A–D ) . Complete autosomal synapsis in H3f3b-/+ males was confirmed in double-staining for SYCP3 and SYCP1 ( synaptonemal complex proteins 3 and 1 ) . SYCP3 marks chromosomal axial elements that have assembled in meiosis-I , and is present before and during synapsis , while SYCP1 marks synapsed chromatin . All spreads with 20 discrete SYCP3-positive axial elements examined ( >30 ) were equivalently positive for SYCP1 , demonstrating complete synapsis . The unsynapsed sex chromosomes remained SYCP1-negative ( Fig . 5E–H ) . Finally , no increase in apoptosis was seen in spermatocytes or round spermatids by TUNEL staining ( S2 Fig . ) . Overall , these data strongly suggest that meiosis-I is essentially normal in H3f3b-/+ spermatocytes . Thus , the failure in spermatogenesis is localized to round spermatids , which appear to arrest through mechanisms that do not involve apoptosis . These data suggest a crucial role for H3 . 3B in early spermiogenesis , one that is relatively more important than for H3 . 3A . In H3f3a mutants , spermiogenesis was also abnormal , but did lead to the formation of sperm . RNA in situ hybridization analysis has shown that the expression of H3f3b is restricted to primary spermatocytes , while the expression of H3f3a is ubiquitous across spermatogenic cell types [24] . It therefore appears that the requirement for H3 . 3B in spermiogenesis is fulfilled by translation in meiosis-I . The difference in severity of phenotype was not reflected in the relative amount of RNA , as we measured more H3f3a than H3f3b RNA in pachytene spermatocytes ( Fig . 6A ) . Also , the amounts of H3 . 3 detected in pachytene spermatocytes and round spermatids of H3f3a-/- and H3f3b+/- males were not discernibly different to WT ( Fig . 6B ) . H3f3a-/- females were of normal fertility . Of eight 6 wk females caged with WT 129S1 males , seven gave birth within 10 wk . Litter size was 4 . 86 ± 1 . 86 ( 7 ) [mean ± s . d . ( n ) ] , not significantly different to that obtained with H3f3a+/- females in intercrosses; 5 . 54 ± 2 . 54 ( 14 ) ( P>0 . 05 , t-test ) . H3f3a-/- females raised their young normally . Three H3f3a-/- second-generation female young , bred in H3f3a-/- ♀ × H3f3a-/- ♂ matings , were also fertile: when mated to a WT outbred Swiss male , each had three sequential litters no more that one month apart , and with no fewer than five pups in any litter . All pups were raised normally . These results show that H3 . 3A is dispensable in the female germ line , and for fertilization and development . To determine the requirement for H3 . 3B in folliculogenesis , we used the Zp3 ( zona pellucida protein 3 ) promoter-cre cds transgenic mouse line , Tg ( Zp3-cre ) , which excises floxed sequences at primordial follicle activation [25 , 26] . We bred a Tg ( Zp3-cre ) congenic line in strain 129S1 to perform all the experiments described . H3f3bc/c , 0/0 ♀ × H3f3bc/c , 0/Tg ( Zp3-cre ) ♂ matings produced control H3f3bc/c , 0/0 , and experimental H3f3bc/c , 0/Tg ( Zp3-cre ) females . These females were mated with WT males to obtain timed pregnancies . Less H3f3bc/c , 0/Tg ( Zp3-cre ) or experimental females became pregnant , and in these , fetal recovery was very low—only 6% in successful matings , relative to control females ( Table 3 ) . A few females were allowed to give birth , and pups were viable to adulthood . No deciduae were observed in any of the pregnant or non-pregnant plugged experimental females , indicating that there was a failure in ovulation , fertilization , or cleavage . Ova were scored at ½ dpc . The number obtained in control and experimental female matings was equivalent . However , fewer ova in experimental females were fertilized ( Table 3 ) . Zygotes , with visible pronuclei , were cultured overnight to 1½ dpc: only 10% of experimental pronuclear zygotes could cleave . Cleavage failure was also observed in vivo , when oviducts of experimental females were flushed at 1½ dpc ( Table 3 ) . Experimental zygotes that failed to cleave are shown ( Fig . 7A ) . To gain insight into the cause of cleavage failure in H3f3b-/- zygotes , we studied aspects of chromatin structure using immunofluorescence . A hallmark of late prophase and mitosis is the serine phosphorylation of histone H3 ( H3Sph ) . The deposition dynamics of this mark are highly conserved , suggestive of important roles in mitosis . Indeed , it has been recently shown that H3S10ph is required for proper chromosome condensation in budding yeast ( Saccharomyces cerevisiae ) [27] . H3f3b-/- zygotes did not accumulate this mark , even after overnight culture , indicating that they failed to progress to late prophase ( Fig . 8 ) . Staining for H3K9me3 , being specific for the maternal pronucleus [28 , 29] , revealed that the paternal pronucleus was often similar in size or smaller than the maternal , instead of being consistently larger , as in controls ( Fig . 8 ) . We also typically observed small and dispersed nucleolar precursor bodies ( NPBs ) in the maternal pronuclei of H3f3b-/- zygotes—evident under DAPI staining . Control H3f3bc/c zygotes typically contained one large NPB in each pronucleus ( Fig . 8 ) . Similarly , the central regions of intense DAPI staining in the paternal pronuclei of H3f3b-/- zygotes could be related to a reduction in the development of NPBs: normally , intense DAPI staining is concentrated at the periphery of NPBs , corresponding to constitutive heterochromatin [29 , 30] . DNA replication was examined in H3f3b-/- zygotes using Edu incorporation , concomitant with H3K9me3 immunofluorescence so that the maternal and paternal pronuclei could be definitively discerned . Three classes of H3f3b-/- zygote were seen in respect to relative pronuclear size and DNA synthesis: ( 1 ) paternal larger than the maternal , robust replication in both—3 of 21 zygotes , ( 2 ) paternal similar to the maternal , reduced replication in at least the paternal—8 of 21 zygotes , and ( 3 ) paternal smaller than the maternal , replication in only the maternal—10 of 21 zygotes ( Fig . 8 ) . Thus , all mutant zygotes appeared to have smaller NPBs , while most showed reductions in DNA replication . The developmental potential of H3f3a and H3f3b mutant oocytes as described above , indicates that H3f3b is either the predominant source , or the only source , of H3 . 3 during folliculogenesis . To distinguish between these possibilities , we first examined the relative amounts of H3f3a and H3f3b RNA in ovulated unfertilized oocytes by quantitative real-time PCR . Both RNAs are present , with the amount of H3f3b RNA being double that of H3f3a ( Fig . 6A ) . Next , we assessed the developmental potential of double-conditional experimental H3f3ac/c , H3f3bc/c , 0/Tg ( Zp3-cre ) oocytes , in which H3 . 3A and H3 . 3B should be eliminated soon after primary oocyte activation . Presumptive double-mutant primary oocytes were inviable . Experimental females at 4 wk had very small ovaries , with no apparent oocyte growth . A similar phenotype was seen in 10 wk experimental females ( Fig . 7B ) . Primordial follicles , of ~11 μm in diameter , and very early-stage primary oocytes , were of normal appearance . By contrast , nearly all primary oocytes developing a second layer of follicle cells were degenerating , reaching a maximum of ~50 μm in diameter ( Fig . 7C ) . Very few larger primordial oocytes were seen , except for a rare early-stage antral follicle . This result indicates that H3f3a and H3f3b are both sources of H3 . 3 during folliculogenesis , and clearly shows that H3 . 3 is absolutely required . Despite the inviability of double-mutant primary oocytes , they had undergone a significant development of the zona pellucida , indicating that the genes encoding the zona glycoproteins were being actively transcribed up to the stage of oocyte death ( Fig . 7C ) .
We have assessed the contribution to viability and fertility of the two unlinked genes encoding the histone variant H3 . 3 in mice . The findings provide implications for the relative importance of each gene . H3f3a and H3f3b mutants could develop to late gestation , demonstrating no absolute requirement for either gene up to this stage . H3f3a is strongly ubiquitously expressed during development [31 , 32] , and the present study indicates this is also true for H3f3b . Both genes have CpG-island promoters , consistent with housekeeper expression . Given the wide expression profile of each gene , a level of functional redundancy is likely . The co-ablation of H3f3a and H3f3b RNA in zygotes by morpholinos resulted in morula-stage arrest , revealing functional importance of H3 . 3 at this early stage [33] . It remains of interest to determine the developmental potential of double-mutant animals in our experimental system . Double-mutants may survive the preimplantation period because of the presence of maternal H3f3a and H3f3b RNAs . Also , mouse ES cells in which H3 . 3 is depleted are viable , and are able to differentiate into all three germ layers in teratomas [9] . The presence of two H3 . 3-encoding genes appears to be important , given its conservation in at least mammals and birds [18] . This gene doubling may be necessary to satisfy the high demand for this protein in DNA packaging and other downstream functions , akin to the multiplicity of canonical H3 . 1- and H3 . 2-encoding genes . Sequences outside the cds of H3f3a and H3f3b are nevertheless highly divergent . From the present results , H3f3b is generally functionally more important than H3f3a , given H3f3b mutants were more severely affected than H3f3a mutants . This was seen in fetal and postnatal growth , perinatal survivability , spermiogenesis , and zygote development . However , this conclusion assumes that there is no compensatory regulation between the two genes . This cannot be examined at the protein level in WT animals , as H3 . 3A and H3 . 3B cannot be distinguished . Surprisingly , H3f3a and H3f3b mutants contained similar amounts of residual H3 . 3 in a range of tissues examined , except brain , where the residual H3 . 3A was higher . These findings indicate that differences in expression of the two genes at the sub-organ tissue-specific level are the most important determinants of the relative functional importance of H3 . 3A and H3 . 3B . Differences in tissue-specificity of expression is suggested in the recent findings that paediatric brain tumors , and giant cell tumors of bone , contain driver amino acid substitution mutations in H3F3A , while chondroblastomas contain driver mutations in H3F3B [16 , 17] . A detailed investigation of the tissue-specific expression of each of these genes in development would be of interest . Approximately one half of H3f3b mutants appeared to die at the implantation or early postimplantation stage . The other half developed to fetal stages , with the majority reaching birth . One possible explanation for this developmental bipotentiality is chromosomal instability . A higher level of aneuploidy was reported in H3f3b mutant primary embryo fibroblasts [34] , therefore there could exist a potential for aneuploidy at the preimplantation stage . Most aneuploid blastocysts die at implantation , or shortly thereafter [35] . The failure of H3f3b mutants to initiate respiration at birth was not accompanied by any lung histopathology , but there could be lung surfactant or neural deficits that lead to immediate respiratory failure [20] . The defects induced by H3 . 3B deficiency in the germ line were striking . The constitutive loss of one H3f3b allele led to the arrest of round spermatids , and sterility . By contrast , the loss of both H3f3a alleles resulted in dysmorphic spermatozoa , and subfertility . This difference in severity occurred despite our lack of evidence that H3f3b is expressed at a higher level than H3f3a , or is the major source of H3 . 3 , in pachytene spermatocytes and round spermatids . This raises the possibility that there could be qualitative differences in the distribution of H3 . 3A and H3 . 3B in spermatogenic cells . Differential expression of the two genes during the prolonged meiotic cycle could lead to regional and functional differences in the incorporation of each protein into chromatin , depending on what sites are vacant for incorporation , and when . The phenotypes described here in H3f3a and H3f3b mutants differ substantially from phenotypes described in other studies . A hypomorphic gene-trap mutation in H3f3a led to very high mortality by weaning [36] . In explaining our milder phenotype , genetic background is an unlikely explanation , as both H3f3a mutations were propagated in a ‘Steel’ substrain of strain 129 [37] . Environmental influences are a possible explanation: the survivability of H3f3a mutants to weaning was increased when litter size was reduced [36] . Finally , the H3f3a retroviral gene-trap mutation was reported to produce up to ~20% normal levels of full length H3f3a RNA [36] . Therefore , retroviral sequences could have resulted in a deregulation of residual H3 . 3A that was detrimental to viability . In contrast to the situation for H3f3a , our H3f3b mutants were considerably more affected than those previously described [34 , 38] . In this earlier study , a normal frequency of H3f3b mutants was obtained at midgestation , although most were substantially growth retarded . Despite this prominent defect , many mutants survived parturition and developed to adulthood . No growth deficit was seen in postnatal H3f3b heterozygous or mutant animals [34] . By marked contrast , our H3f3b mutants could not survive beyond birth , and a significant number died even before this stage . Further , we saw marked reductions in growth in perinatal and postnatal heterozygous and mutant animals . The greater phenotypic severity associated with our H3f3b mutation was also seen in the male germ line . We saw a near-uniform arrest of round spermatids in heterozygous males . By contrast , previously reported heterozygous males were fertile [34] . Indeed , the phenotypic severity in our heterozygous animals exceeded even that seen in recently described H3f3b mutants , where appreciable progression through spermiogenesis was observed . These mutants were sterile due to the production of sperm with motility and head defects [38] . Other differences were apparent . We saw no increase in apoptosis by TUNEL staining in seminiferous tubules of heterozygotes , while an increase was seen in mutants [38] . Reduced protamine loading was reported in mutants , and this could be a major contributor to the defective head formation in mature sperm [38] . This phenotype was similar to that seen in our H3f3a mutant males , therefore there could be a similar mechanistic basis . However , in our H3f3b heterozygous males , other explanations are required , as the stage of arrest precedes the phase of histone displacement . In round spermatids , there is a large increase in transcriptional activity [39 , 40] , and defects in this transcriptional activation could lead to their arrest . The simplest explanation for the differences in effect of the two H3f3b mutations is genetic background . Ours was kept in strain 129S1 , while the other was kept in strain C57BL/6 [38] . A caveat in interpreting the present , and previous , results is that the H3f3b mutations were constitutive . H3 . 3B deficiency in Sertoli cells , which support spermatogenesis , could exacerbate the germ cell defects observed . H3 . 3B deficiency in folliculogenesis led to some fertilization failure , and zygotic arrest , demonstrating that H3f3b is a maternal effect gene . Loss of maternal Hira , a H3 . 3 chaperone , leads to zygote failure in flies ( Drosophila melanogaster ) . After sperm entry , H3 . 3 fails to incorporate into paternal chromatin , leading to failed decondensation of sperm chromatin [41] . A similar effect has recently been described in mouse , where depletion of HIRA in folliculogenesis using the Tg ( Zp3 . cre ) transgene resulted in zygote arrest . The loss of HIRA led to a failure of paternal pronucleus formation . Also , the loss of DNA replication and ribosomal RNA transcription in paternal and maternal chromatin was observed , and either one of these deficiencies can lead to zygote arrest [42] . The phenotype in our H3 . 3B-deficient mutant oocytes was similar to , but less severe than , the HIRA-deficient phenotype , and could be explained by the presence of residual H3 . 3A . Our phenotype was variable , ranging from reduced , to overtly normal , paternal pronuclear expansion . Only a small number of mutant zygotes showed robust DNA replication in both pronuclei . If these represent the few zygotes able to cleave , then incomplete DNA replication could be the predominant cause of failed entry into first mitosis . Another possible explanation is a deficiency of H3Sph , given the importance of H3 . 3 in DNA packaging in the zygote , and the probable requirement for H3Sph in chromosome condensation at the end of prophase [27] . This could be investigated by determining the developmental potential of H3 . 3B-deficient zygotes supplemented with mutant H3 . 3 lacking relevant serines . H3 . 3 incorporation has also been recently shown to be required for pronuclear pore formation [43] and it is possible that a defect in this assembly could explain the mitosis failure . An important requirement for H3 . 3 during the early phase of oocyte growth was seen in the developmental failure of H3f3a , H3f3b double-mutant primary oocytes . Given that HIRA-deficient oogonia can develop and be ovulated , this result clearly demonstrates a requirement for other H3 . 3 chaperones , possibly DAXX and ATRX , in regulating folliculogenesis . Conditional mutagenesis of these chaperones during folliculogenesis could be informative . The death of H3 . 3-deficient primary oocytes could be related to the lack of incorporation by a chaperone aside from HIRA , or to the combined lack of incorporation by two or more chaperones . This study provides insight into the relative importance of the two genes encoding H3 . 3 for development and fertility in the mouse , and provides a basis for further investigations into the various biological roles of this basic building block of chromatin .
The mouse lines carrying Cre/loxP conditional mutant alleles of H3f3a and H3f3b , termed H3f3ac and H3f3bc , have been described [32] . The Cre-deleter line used to derive mice carrying constitutive mutant alleles , termed H3f3a- and H3f3b- , was 129S1/Sv-Hprttm1 ( cre ) Mnn/J ( Jackson Laboratory , stock no . 004302 ) . Testes and ovaries from adult mice were fixed in Bouin’s fixative . Sections were embedded in paraffin and stained by Periodic-Acid Schiff ( PAS ) . Fetuses at 18½ dpc were decapitated , transferred to 20 mL of Bouin’s fixative , rocked slowly for 3 days , then washed in four changes of 70% ( v/v ) ethanol over 2 days . Sections were stained with haematoxylin and eosin . Spreads of mature sperm were made by squeezing the cauda epididymis with the back of curved forceps for release into 0 . 3 mL of medium M2 [44] . Most of the droplet was transferred to a tube with a wide-bore yellow tip , then an equal volume of 10% phosphate buffered formalin added , and gently mixed . After fixation ( 15 min , RT ) , a small drop of the sperm suspension was smeared onto a slide , air-dried , and stained with haematoxylin and eosin . Spermatocyte spreads from 18–22 days post partum mice and immunofluorescence ( IF ) were performed as described [45] , with the following modifications: both testes of an 18–22 days post partum ( dpp ) male were transferred to 2 mL of Dulbecco’s phosphate buffered saline ( without calcium and magnesium , with 5 . 6 mM glucose and 0 . 4 mg/mL bovine serum albumin ) , in a petri dish , the tunicae removed , and cut into small pieces with scissors . The pieces were transferred to 3 mL of 0 . 25% trypsin-EDTA ( Gibco , 25200 ) in a polystyrene tube ( Falcon , 352054 ) , incubated ( 37°C , 20 min ) , and triturated vigorously with a transfer pipette ( 30 sec ) . The crude suspension was passed through a cell strainer ( Falcon , 352350 ) into 1 mL of fetal bovine serum in a 15 mL centrifuge tube , and mixed . The spermatocytes were pelleted ( 250 × g , 4 min ) , the pellet resuspended in 3 mL of medium M2 , and the tube placed on ice . Protease inhibitors ( Sigma Aldrich , P8340 ) were added to all of the solutions . To prepare the spreads , 50 μL of the sperm suspension was combined with 0 . 1 mL of medium M2 , and pelleted ( 500 × g , 2 min ) . The supernatant was removed , and the pellet resuspended in 40 μL of alkaline sucrose solution [45] . Two drops of this suspension were then dispensed from a yellow tip into ~0 . 1 mL of the 1% ( w/v ) paraformaldehyde fixative solution , this being previously dispensed within a rectangle of ~1 . 5 cm2 drawn on a SuperFrost slide with a PAP pen . Spermatocytes were allowed to settle onto the slides in a humidified chamber ( 2 h , RT ) . Protease inhibitors were added to all solutions , except the fixative solution . After removing from the chamber , slides were rinsed in water , then in 0 . 004% ( v/v ) Photo-Flo 200 ( Kodak , 14634510 ) in water , and air-dried . Slides were stored at -20°C until hybridization with Abs . Imaging was performed with a Imager . M1 microscope and AxioCam MRm camera ( Carl Zeiss ) . Primary Abs used , dilution factor: anti-γH2A . X , rabbit mcAb , 400× ( Cell Signaling , 9718 ) ; anti-H1FNT , guinea pig pcAb , 250× ( kindly provided by Mary Ann Handel ) ; anti-SYCP3 , mouse mcAb , 100× ( Abcam , 97672 ) ; anti-SYCP1 , rabbit pcAb , 200× ( Abcam , 15090 ) . Secondary Abs used , dilution factor: anti-rabbit IgG , goat , 1000× , AlexaFluor 488; anti-guinea pig IgG , goat , 1000× , AlexaFluor 594; anti-mouse IgG , goat , 1000× , AlexaFluor 488; anti-rabbit IgG , goat , 1000× , AlexaFluor 594 ( Life Technologies , A-11008 , A-11076 , A-11001 and A-11012 , respectively ) . Pachytene spermatocytes ( larger 4c cells ) and round spermatids ( smaller 1c cells ) were purified by flow cytometry as described [46]: testis suspensions were obtained as described above , except that DNAseI ( Sigma Aldrich , D4527 ) was added to the trypsin digest at 5 U/mL , and the cell pellet resuspended in 1 . 5 mL of medium M2 containing 1 U/mL of DNAseI . The final suspension was incubated with Hoechst 33342 ( Life Technologies , H3570 ) for gating on larger pachytene spermatocytes ( 4c ) and smaller round spermatids ( 1c ) , and propidium iodide ( Life Technologies , P3566 ) for gating on live cells ( S3 Fig . ) . Cells were sorted into 0 . 3 mL of medium M2 in 1 . 5 mL centrifuge tubes using an Influx cell sorter ( Becton Dickinson ) , then pelleted ( 2 , 500 rpm , 3 min ) , the supernatant removed , and stored at -80°C . RNA was isolated using TRI Reagent ( Life Technologies , AM9738 ) according to the instructions . Pachytene spermatocytes ( 200 , 000 ) were purified from WT prepubertal 129S1 males by flow cytometry as described above . Oocytes ( 200 ) were obtained from WT superovulated prepubertal 129S1 females . The zona pellucida was removed using acid tyrode’s solution [47] before lysis in TRI Reagent . Two random-primed reverse transcription ( RT ) reactions were performed with SuperScript III Reverse Transcriptase ( Life Technologies , 18080 ) . For each RT , triplicate PCR reactions ( 20 μL , 10 ng cDNA ) were performed using GoTaq qPCR Master Mix ( Promega , A6001 ) to yield six Ct values . Equipment used was the 7500 Real-Time PCR System ( Applied Biosystems ) . These values were converted to the relative amount of RNA by multiplying the fold increase in product per cycle—equivalent to amplification efficiency , to the power of ∆Ct . Amplification efficiencies for each transcript were pre-determined in standard curves . Each value was normalized to the mean obtained for the Rsp7 ( ribosomal protein S7 ) housekeeper RNA . Primers , 5′-3′: H3f3a , GCAG CTAT TGGT GCTT TGC ( JRM-1227 ) , ATGT TTCC CCTC ATAG TGGA CTC ( JRM-1228 ) , amplicon 173 bp; H3f3b , GGAG ATCG CCCA GGAT TTC ( JRM-1231 ) , ATGC CATA AAAA CCGC TTCA AC ( JRM-1232 ) , amplicon 215 bp; Rps7 , TGGT CTTC ATTG CTCA GAGG AGG ( JRM-609 ) , TGCC ATCC AGTT TCAC ACGG ( JRM-610 ) , amplicon 177 bp . Immunoblotting was performed as described [48] . Tissues were homogenized by trituration with a 19 G needle and syringe in RIPA buffer . Primary Abs used , dilution factor: anti-H3 . 3 , rabbit pcAb , 1000× ( Millipore , 09–838 ) ; anti-H3 ( panH3 ) , rabbit pcAb , 2000× ( Abcam , 1791 ) ; anti-TUBB ( β-tubulin ) , rabbit mcAb , 1000× ( Cell Signaling , 2128 ) ; anti-GAPD ( glyceraldehyde-3-phosphate dehydrogenase ) , rabbit mcAb , 1000× ( Cell Signaling , 5174 ) . Secondary Abs used , dilution factor: anti-rabbit IgG , goat , horseradish peroxidase conjugate , 15000× ( Millipore , 12–348 ) . The light cycle in the mouse colony was 0630–1830 h . Zygotes were collected at ~1200 h on the day of obtaining a vaginal plug , when usually they were at the mid-pronuclear stage . They were cultured for ~2 h for pronuclei to become more prominent before fixation . For H3S10ph immunofluorescence , WT zygotes were not fixed until at least two in the batch had undergone pronuclear envelope breakdown , indicating entry into mitosis . Immunofluorescence was performed as described [30] , except goat serum was used in blocking , and fixation and permeabilization solutions contained 0 . 2% ( w/v ) and 0 . 1% ( w/v ) BSA , respectively , to prevent zygote sticking . Fixation , permeabilization , washing and secondary Ab incubation steps were performed in volumes of 0 . 1–0 . 15 mL in 12-well staining dishes ( ProSciTech , H421–12 ) , while blocking and primary Ab incubations were carried out in 10 μL drops under paraffin oil . In mounting , zygotes were transferred to a 1 μL drop of Fluorescence Mounting Medium ( Dako , S3023 ) on a SuperFrost slide , coverslipped , and sealed with nail polish . 5-ethynyl-2′-deoxyuridine ( Edu ) incorporation and detection was carried out using a Click-iT Edu Imaging Kit ( Life Technologies , C10339 ) : freshly isolated zygotes with pronuclei were cultured in 20 μM Edu in medium KSOM-GAA [49] ( 2 h ) prior to fixation . H3K9me3 primary and secondary Ab incubations were carried out immediately before Edu detection . Primary Abs , dilution factor: anti-H3K9me3 , rabbit pcAb , 1000× ( Abcam , 8898 ) ; anti-H3S10ph , rabbit mcAb , 1000× ( Cell Signaling , 3377 ) . Secondary Ab , dilution factor: anti-rabbit IgG , goat , 1500× , AlexaFluor 488 ( Life Technologies , A-11008 ) . The third wash after secondary Ab staining ( 10 min , RT ) contained DAPI at 1 μg/mL , then zygotes were rinsed in DAPI-free wash solution ( 1 min ) before mounting . Tail tips were digested in 0 . 1 mL of 50 mM Tris-HCl pH 8 . 0 , 0 . 5% ( v/v ) Triton X-100 , 20 mM EDTA , 400 μg/mL proteinase K , ( 55°C , overnight ) . An aliquot was diluted 20× in water , then heat inactivated ( 70°C , 10 min ) . This template solution was diluted 10× in the final PCR reaction , performed using MyTaq Red DNA polymerase ( Bioline , 21108 ) . Primers: H3f3a , CTGG TTTT GGCT GTTT TATC GCTC GG ( WT primer , JRM-1272 ) , GCTA TTGC TTTA TTTG TAAC CATT ATAA GCTG C ( mutant/conditional , JRM-1461 ) , and AGGG CGCA CTCT TGCG AGC ( common , JRM-1276 ) ; triplex reaction , amplicons WT—362 , and mutant/conditional—205 bp . H3f3b , CTCA CCGC TACA GGTA GGC ( WT primer , JRM-1465 ) , GCTA TTGC TTTA TTTG TAAC CATT ATAA GCTG C ( mutant/conditional , JRM-1461 ) , TCTC CCTC ACCA ATCT CTGG ( common , JRM-1466 ) ; triplex reaction , amplicons WT—211 , mutant/conditional—357 . Xcre , specific for the cre cds , TGCT GTTT CACT GGTT ATGC GGCG ( JRM-391 ) , TGCC TTCT CTAC ACCT GCGG TGCT ( JRM-392 ) ; amplicon 304 bp . Y chromosome , nos . 211 and 212 [50] . This work was approved by the Animal Ethics Committee of the Murdoch Children’s Research Institute , approval no . A692 . | Histones package DNA and regulate chromosome activity . Histone H3 is particularly important in this regard . The H3 . 3 isoform is unique , among H3 histones , in being able to incorporate into chromosomes independent of DNA replication . Thus , in slowly- or non-dividing somatic cells , H3 . 3 histone can take on the bulk of H3 functions . The developing germ line is very slowly dividing , and undergoes large-scale changes in chromosome activity . H3 . 3 is therefore likely to be very important in regulating these processes . Here , we have studied the effects of null mutations in each of the two genes encoding H3 . 3 . We demonstrate that H3 . 3 is very important in germ cell development , regulating oogenesis , spermatogenesis , and fertilization . Also , we reveal H3 . 3 to be important in somatic growth . Each of the two genes is required to varying extents in regulating these processes . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | Contribution of the Two Genes Encoding Histone Variant H3.3 to Viability and Fertility in Mice |
We propose a top-down approach to the symptoms of schizophrenia based on a statistical dynamical framework . We show that a reduced depth in the basins of attraction of cortical attractor states destabilizes the activity at the network level due to the constant statistical fluctuations caused by the stochastic spiking of neurons . In integrate-and-fire network simulations , a decrease in the NMDA receptor conductances , which reduces the depth of the attractor basins , decreases the stability of short-term memory states and increases distractibility . The cognitive symptoms of schizophrenia such as distractibility , working memory deficits , or poor attention could be caused by this instability of attractor states in prefrontal cortical networks . Lower firing rates are also produced , and in the orbitofrontal and anterior cingulate cortex could account for the negative symptoms , including a reduction of emotions . Decreasing the GABA as well as the NMDA conductances produces not only switches between the attractor states , but also jumps from spontaneous activity into one of the attractors . We relate this to the positive symptoms of schizophrenia , including delusions , paranoia , and hallucinations , which may arise because the basins of attraction are shallow and there is instability in temporal lobe semantic memory networks , leading thoughts to move too freely round the attractor energy landscape .
Schizophrenia is a major mental illness , which has a great impact on patients and their environment . One of the difficulties in proposing models for schizophrenia is the complexity and heterogeneity of the illness . We propose that part of the reason for the inconsistent symptoms may be a reduced signal-to-noise ratio and increased statistical fluctuations in different cortical brain networks . The novelty of the approach described here is that instead of basing our hypothesis purely on biological mechanisms , we develop a top-down approach based on the different types of symptoms and relate them to instabilities in attractor neural networks [1] . The main assumption of our hypothesis is that attractor dynamics are important in cognitive processes [2] . Our hypothesis is based on the concept of attractor dynamics in a network of interconnected neurons that in their associatively modified synaptic connections store a set of patterns , which could be memories , perceptual representations , or thoughts [3–5] . The attractor states are important in cognitive processes such as short-term memory , attention , and action selection [6] . The network may be in a state of spontaneous activity , or one set of neurons may have a high firing rate , each set representing a different memory state , normally recalled in response to a retrieval stimulus . Each of the states is an attractor in the sense that retrieval stimuli cause the network to fall into the closest attractor state , and thus to recall a complete memory in response to a partial or incorrect cue . Each attractor state can produce stable and continuing or persistent firing of the relevant neurons . The concept of an energy landscape [4] is that each pattern has a basin of attraction , and each is stable if the basins are far apart and also if each basin is deep , which is caused , for example , by high firing rates and strong synaptic connections between the neurons representing each pattern , which together make the attractor state resistant to distraction by a different stimulus . The spontaneous firing state , before a retrieval cue is applied , should also be stable . Noise in the network caused by statistical fluctuations in the stochastic spiking of different neurons can contribute to making the network transition from one state to another; we take this into account by performing integrate-and-fire simulations with spiking activity , and relate this to the concept of an altered signal-to-noise ratio in schizophrenia [7–9] . Schizophrenia is characterized by three main types of symptom: cognitive dysfunction , negative symptoms , and positive symptoms [10–12] . We consider how the basic characteristics of these three categories might be produced in a neurodynamical system , as follows . Dysfunction of working memory , the core of the cognitive symptoms , may be related to instabilities of persistent attractor states [13 , 14] , which we show can be produced by reduced firing rates in attractor networks in brain regions such as the prefrontal cortex . The negative symptoms such as flattening of affect or reduction of emotions may be caused by a consistent reduction in firing rates of neurons in regions associated with emotion , such as the orbitofrontal cortex [1] . These hypotheses are supported by the frequently observed hypofrontality , a reduced activity in frontal brain regions in patients with schizophrenia during cognitive tasks [15–17] . The positive symptoms are characterized by phenomenologically overactive perceptions or thoughts such as hallucinations or delusions , which are reflected , for example , by higher activity in the temporal lobes [17 , 18] . We relate this category of symptoms to a spontaneous appearance of activity in attractor networks in the brain and more generally to instability of both the spontaneous and persistent attractor states . We do not try to account for every detail of schizophrenic symptoms , which are diverse and vary among patients , but instead show how these three main categories of symptoms could be related to changes in the stability of dynamical attractor systems in the brain , and how the changes in the stability might be produced by changes at the level of the currents being passed through receptor-activated ion channels at synapses . There are specific symptoms such as aberrant eye movements that cannot be accounted for by this general scheme . In particular , we were interested in how these symptoms are related . Negative and cognitive symptoms typically precede the first psychotic episode [19 , 20] . Positive symptoms can be treated in most cases with neuroleptics , whereas negative and cognitive symptoms persist , at least for typical neuroleptics . Can a mapping onto a dynamical system help to understand these relations ? After proposing a dynamical systems hypothesis for the different symptoms of schizophrenia , we study a standard neural network model [21] of cortical dynamics specifically in relation to our hypothesis . We were especially interested in how excitation and inhibition implemented by NMDA and GABA synapses affect the network dynamics . Alterations in the efficacies of the NMDA and GABA channels have been identified in the pathology of schizophrenia [22 , 23] , and transmitters such as dopamine influence the currents in these receptor-activated channels [24] . Do NMDA and GABA currents have antagonistic effects , or do they have a special role in the network dynamics ? How could this be related to our hypothesis of schizophrenia ? Building upon the current body of neural network research , we specifically conduct neural network simulations to substantiate our dynamical systems hypothesis of schizophrenia . While focussing on NMDA and GABA synapses in the Results , in the Discussion we consider how altered transmission at D1 and D2 receptors by modulating NMDA and GABA conductances could not only influence working memory , which has been investigated previously [21 , 25–27] , but could in particular influence the different symptoms of schizophrenia .
We relate the three types of symptoms of schizophrenia to the dynamical systems attractor framework described at the end of the Introduction as follows [1] . The cognitive symptoms of schizophrenia include distractibility , poor attention , and the dysexecutive syndrome [10–12 , 28] . The core of the cognitive symptoms is a working memory deficit in which there is a difficulty in maintaining items in short-term memory [29 , 30] , which could directly or indirectly account for a wide range of the cognitive symptoms . We propose that these symptoms may be related to instabilities of persistent states in attractor neural networks , consistent with the body of theoretical research on network models of working memory [13] . The neurons are firing at a lower rate , leading to shallower basins of attraction of the persistent states , and thus a difficulty in maintaining a stable short-term memory , normally the source of the bias in biased competition models of attention [5 , 6] . The shallower basins of attraction would thus result in working memory deficits , poor attention , distractibility , and problems with executive function and action selection [6 , 31] . The negative symptoms refer to the flattening of affect and a reduction in emotion . Behavioural indicators are blunted affect , emotional and passive withdrawal , poor rapport , lack of spontaneity , motor retardation , and disturbance of volition [10–12] . We propose that these symptoms are related to decreases in firing rates in the orbitofrontal cortex and/or anterior cingulate cortex [1] , where neuronal firing rates and activations in functional MRI investigations are correlated with reward value and pleasure [1 , 32] . Consistent with this , imaging studies have identified a relationship between negative symptoms and prefrontal hypometabolism ( i . e . , a reduced activation of frontal areas [33 , 34] ) . The positive symptoms of schizophrenia include bizarre ( psychotic ) trains of thought , hallucinations , and ( paranoid ) delusions [10–12] . We propose that these symptoms are related to shallow basins of attraction of both the spontaneous and persistent states in the temporal lobe semantic memory networks and to the statistical fluctuations caused by the probabilistic spiking of the neurons . This could result in activations arising spontaneously and thoughts moving too freely around the energy landscape , loosely from thought to weakly associated thought , leading to bizarre thoughts and associations that may eventually over time be associated together in semantic memory to lead to false beliefs and delusions . Consistent with this , neuroimaging studies suggest higher activation , especially in areas of the temporal lobe [17 , 18 , 35] . To further investigate our hypothesis , we use an attractor network , as this is likely to be implemented in many parts of the cerebral cortex by the recurrent collateral connections between pyramidal cells , and has short-term memory properties with basins of attraction that allow systematic investigation of stability and distractibility . The particular neural network implementation we adopt includes channels activated by AMPA , NMDA , and GABAA receptors , and allows not only the spiking activity to be simulated , but also a consistent mean-field approach to be used [21] . First , we introduce an analytical approach to the concepts of how changes in transmitters could affect the depth of the basins of attraction in networks in ways that may be related to the symptoms of schizophrenia . The depth of the basins of attraction can be assessed by calculating the flow toward the attractors using the mean-field reduction ( see Methods and [36] ) . The flow is the force that draws the dynamical system to one of the attractors . Figure 1 shows the flow between the spontaneous and persistent state in a network featuring one selective pool . The curve for ( Normal ) shows the flow for the reference baseline condition . The crossing of the curve and the 0-flow axis represent the fixed points , which are either stable ( negative derivative ) or unstable ( positive derivative ) . We use this line as a reference to assess the relative depth of the energy landscape . The system flows either into a spontaneous firing state ( of approximately 2 Hz ) , or a high-firing attractor state . A reduction of NMDA ( −NMDA ) shows a stronger flow than the unchanged condition at low firing rates toward the spontaneous attractor ( at about 2 Hz ) . The absolute values of the function are higher compared with the normal condition until the first unstable fixed point ( at around 6–7 Hz ) . The basin of attraction toward the persistent attractor at high firing rates yields the reverse picture . Here , the −NMDA curve is clearly below the unchanged condition , and the flow toward the attractor is smaller . Overall , the basin of attraction is deeper for the spontaneous state and shallower for the persistent state compared with the unchanged condition . This pattern fits with the cognitive symptoms of schizophrenia as proposed in our hypothesis . Also , note that the firing rate of the persistent fixed point is reduced in the −NMDA condition ( crossing with the flow = 0-axis ) , which is consistent with the hypothesis for the negative symptoms . A reduction of the GABA conductance ( −GABA ) yields the opposite pattern to that in the reduced NMDA condition . Here , the basin of attraction of the persistent state is deeper . This is not a condition that we suggest is related to the symptoms of schizophrenia . However , in the condition in which both the NMDA and the GABA conductances are reduced ( −NMDA , −GABA ) , the persistent-state basin of attraction is shallower , and the spontaneous-state basin is a little shallower . This condition corresponds to the proposed landscape for the positive symptoms as considered above . In particular , in the −NMDA , −GABA condition , the system would be less stable in the persistent state , tending to move to another attractor easily , and less stable in the spontaneous state , also tending to move too readily into an attractor from spontaneous activity . Overall , the flow analysis suggests that both the cognitive and negative symptoms could be related to a decrease in the NMDA conductances . This is consistent with the fact that these two symptoms usually appear together . The flow analysis suggests that the positive symptoms are related to a reduction in both NMDA and GABA . Thus , the transition from the cognitive and negative symptoms to the positive , psychotic symptoms might be caused by an additional decrease in the GABA conductance . It is notable that excitation and inhibition do not cancel each other out as assumed by many models , but have distinct influences on the network dynamics . The flow analysis provides insight into how the depth of the basins of attraction and the firing rates are influenced by changes in the conductivities of the channels activated via NMDA and GABA receptors . However , the overall stability of the different attractors is affected not only by the depth of the basins of attraction , but also by the breadth and distance apart of the basins and by the statistical noise generated by the randomness of the spiking of the different neurons . These statistical fluctuations play a role in the way in which the system moves from one state to another , for these statistical fluctuations can cause hills in the energy landscape to be crossed stochastically . Since the mean-field analyses do not take these properties into account , we investigate the system further using large-scale integrate-and-fire network simulations and measuring the statistics of the network behavior . To clarify the concept of stability , we show examples of trials of spontaneous and persistent simulations in which the statistical fluctuations have different effects on the temporal dynamics . Figure 2 shows the possibilities , as follows . In the spontaneous-state simulations , no cue is applied , and we are interested in whether the network remains stably in the spontaneous firing state , or whether it is unstable and , on some trials due to statistical fluctuations , enters one of the attractors , thus falsely retrieving a memory . Figure 2A shows an example of a trial on which the network correctly stays in the low spontaneous firing rate regime , and another trial ( labelled spontaneous unstable ) in which statistical spiking-related fluctuations in the network cause it to enter a high-activity state , moving into one of the attractors even without a stimulus . In the persistent-state simulations , a strong excitatory input is given to the S1 neuronal population between 0 and 500 ms ( see Analysis section ) . Two such trials are shown in Figure 2B . In one , the S1 neurons ( correctly ) keep firing at approximately 30 Hz after the retrieval cue is removed at 500 ms . However , due to statistical fluctuations in the network related to the spiking activity , on the trial labelled persistent unstable , the high firing rate in the attractor for S1 was not stable , and the firing decreased back toward the spontaneous level , in the example shown starting after 1 . 5 s . This trial illustrates the failure to maintain a stable short-term memory state , even when no distractor is applied . In Figure 2 , the transitions to the incorrect activity states are caused by statistical fluctuations in the spiking activity of the integrate-and-fire neurons and the depth of the basins of attraction , which has been reduced in the simulations shown by reducing both the NMDA and the GABA currents , as indicated in the caption . We hypothesize that such instabilities are related to the symptoms of schizophrenia . We note that there are two sources of noise in the spiking networks that cause the statistical fluctuations: the randomly arriving external Poisson spike trains , and the statistical fluctuations caused by the spiking of the neurons in the finite sized network . The magnitude of these fluctuations increases as the number of neurons in the network becomes smaller [37] . For our investigations , we selected w+ = 2 . 1 , which with the default values of the NMDA and GABA conductances yielded stable dynamics; that is , a stable spontaneous state if no retrieval cue was applied , and a stable state of persistent firing after a retrieval cue had been applied and removed . To investigate the effects of changes ( modulations ) in the NMDA and GABA conductances , we chose for demonstration purposes a reduction of 4 . 5% and 9% , respectively , as these could cause instabilities , as illustrated in Figure 2 . However , the exact values are not crucial to observe the effects described . The magnitudes of these reductions are smaller than those that can be produced experimentally [24 , 26] . A strength of our approach is that we show that even quite small reductions in the synaptic currents can alter the global behaviour of the network , e . g . , the stability of its attractors . We assessed how the stability of both the spontaneous and persistent states changes when NMDA and GABA efficacies are modulated . Specifically , we ran multiple-trial integrate-and-fire network simulations and counted how often the system maintained the spontaneous or persistent state , assessed by the firing rate in the last second of the simulation ( 2–3 s ) of each 3-s trial . Figure 3 shows the stability of the spontaneous and persistent attractors relative to the unmodulated reference state ( Normal ) . A negative percentage means that the system was less stable than in the unmodulated state . A reduction of the NMDA conductance ( −NMDA ) reduces the stability of the persistent state drastically , while slightly increasing the stability of the spontaneous state ( see Figure 3 ) . We hypothesized that this type of change might be related to the cognitive symptoms , since it shows a reduced stability of the working memory properties . A reduction of GABA shows the opposite pattern: a slight reduction in the stability of the spontaneous state , and an increased stability of the persistent ( i . e . , attractor ) state ( see Figure 3 ) . When both NMDA and GABA are reduced , one might think that these two counterbalancing effects ( excitatory and inhibitory ) would either cancel each other out or yield a tradeoff between the stability of the spontaneous and persistent states . However , this is not the case . The stability of both the spontaneous state and the persistent state is reduced ( see Figure 3 ) . We relate this pattern to the positive symptoms of schizophrenia , in which both the spontaneous and attractor states are shallow , and the system merely jumps by the influence of statistical fluctuations between the different ( spontaneous and attractor ) states . To investigate more directly the wandering between spontaneous and several different persistent attractor states , we simulated the condition with decreased NMDA and GABA conductances over a long time period in which no cue stimulus input was given . Figure 4 shows the firing rates of the two selective pools S1 and S2 . The high activity switches between the two attractors due to the influence of fluctuations , which corresponds to spontaneous wandering in a shallow energy landscape , corresponding , for example , to sudden jumps between unrelated cognitive processes . These results are consistent with the flow analysis and demonstrate that the changes in the attractor landscape influence the behavior at the stochastic level . As distractibility is directly related to the symptoms of schizophrenia , we ran simulations specifically to assess this property using persistent and distractor simulations ( see Analysis section ) . A distractor strength of 0 Hz corresponds to the persistent condition described in the preceding section ( Stability ) . Figure 5 shows the stability and distractibility for reductions of NMDA and GABA currents . The reference state is labelled “Normal . ” In this state , pool S1 continued to maintain its attractor firing without any distractor ( distractor strength = 0 Hz ) throughout the delay period on almost 90% of the trials . In both conditions that reduce the NMDA current ( labelled −NMDA ) , the network was less and less able to maintain the S1 attractor firing as the distractor stimulus strength was increased through the range of 0–80 Hz . The stability of the persistent state was reduced , and the distractibility was also increased , as shown by the fact that increasing distractor currents applied to S2 could move the attractor away from S1 . The implication , therefore , is that a reduction of the NMDA currents could cause the cognitive symptoms of schizophrenia by making short-term memory networks less stable and more distractible , thereby reducing the ability to maintain attention . Reducing only the GABA currents ( −GABA ) reduces the distractibility for low distractor strengths and coincides with the reference ( Normal ) condition at high values of the distractor strengths . We further investigated the signal-to-noise ratio in relation to the changes in synaptic conductances . The signal-to-noise ratio denotes the level of a signal relative to the level of background noise . In an attractor network , a high signal-to-noise ratio indicates that the network will maintain the attractor stably , as it will be unlikely to be disrupted by spiking-related statistical fluctuations that are the source of the noise in the network . Figure 6 shows the signal-to-noise ratio of a measure related to the functional MRI blood oxygenation level–dependent signal . ( This measure described in the caption to Figure 6 and below was used because the experimental data with which we wish to compare the simulation results use functional MRI measures [7–9] . The index we used of the activity of the network was the total synaptic current of selective pool 1 averaged over the whole simulation time of 3 s to take the temporal filtering properties of the blood oxygenation level–dependent signal into account , given the typical time course which lasts for several seconds of the haemodynamic response function [38] . Further , we subtracted the averages of the spontaneous trial simulations that represent the baseline activity from the persistent trial simulation values . The signal-to-noise ratio was calculated from the mean of this index across trials divided by the standard deviation of the index , both measured using 1 , 000 simulation trials . If the network sometimes had high activity , and sometimes low , then the signal-to-noise measure gave a low value . If the network reliably stayed in the high persistent firing states , then the signal-to-noise ratio measure was high . ) As shown in Figure 6 , we found that in all the cases in which the NMDA or the GABA conductance , or both , are reduced , the signal-to-noise ratio , computed by the mean divided by the standard deviation , is also reduced . This relates to recent experimental observations which show a decreased signal-to-noise ratio in schizophrenic patients [7–9] . Here , we directly relate a decrease in the signal-to-noise ratio to changes ( in this case , decreases ) in receptor-activated synaptic channel conductances . Given these results , it would be of interest in future studies to model the exact paradigm used by Winterer et al . [8] .
We have proposed a hypothesis that relates the cognitive , negative , and positive symptoms of schizophrenia [10–12] to the depth of basins of attraction and to the stability properties of attractor networks caused by statistical fluctuations of spiking neurons . This assumes that some cognitive processes can be understood as dynamical attractor systems , which is an established hypothesis in areas such as working memory , but has also been used in many other areas [2 , 5] . Our approach applies this concept to mental illnesses [39] . Due to the diversity of schizophrenic symptoms , our general hypothesis is meant to serve as a heuristic of how the different kinds of symptoms might arise and are related . We investigated the hypothesis empirically in a computational attractor framework to capture an important aspect of cortical functionality . Figure 7 summarizes our hypothesis and its relation to the investigations of an attractor neural network . The middle column in Figure 7 shows the overview for the cognitive and negative symptoms . The core of the cognitive symptoms is a failure of working memory and attentional mechanisms . Working memory activity is related to the ongoing ( i . e . , persistent ) firing of neurons during the delay period of cognitive tasks [29 , 30] . This could be implemented by associatively modifiable synapses between the recurrent collateral synapses of cortical pyramidal cells [13 , 14 , 40 , 41] . We propose that the cognitive symptoms of schizophrenia could arise because the basins of attraction of the persistent states in the prefrontal cortex become too shallow . This leads in combination with the statistical fluctuations due to randomness of the spiking activity to either a fallout of an active working memory state or to a shift to a different attractor state , leading to a failure to maintain attention and thereby impairing executive function . The hypofrontality in schizophrenia , that is , less activation in frontal brain regions during working memory tasks [15 , 42] , is in line with our hypothesis , since the firing rates of the persistent state are lower in the reduced NMDA condition ( Figure 1 ) , and the system spends on average less time in the persistent state , since it is less stable than in the normal condition ( Figure 3 ) . In addition , a reduced signal-to-noise ratio as shown in our simulations ( Figure 6 ) has also been identified in imaging studies [7–9] . Our simulations suggest that a reduction in NMDA conductance at the synaptic level ( see Figure 7 ) can account for this phenomenon . This is in line with previous work on the stability of working memory networks [14 , 27 , 43] . A reduction of the NMDA conductance also results in a reduction of the firing rates of the neurons in the persistent state ( see Figure 1 and [21] ) . We relate this , following [1] , to the negative symptoms , which include flattening of affect , a reduction in emotion , emotional and social withdrawal , poor rapport , passive withdrawal , lack of spontaneity , motor retardation , apathy , and disturbance of motivation . These symptoms are related to decreases in activity in the orbitofrontal cortex and/or anterior cingulate cortex [33 , 34] , both of which are implicated in emotion [1 , 32 , 44] . The emotional states represented in the orbitofrontal cortex and anterior cingulate cortex include states elicited both by rewards and punishers . Our hypothesis is that both would be reduced by the mechanism described . Correspondingly , motivation would be reduced in the same way , in that motivation is a state in which we work to obtain goals ( rewards ) or avoid punishers [1] . Both the negative and cognitive symptoms thus could be caused by a reduction of the NMDA conductance in attractor networks . The proposed mechanism links the cognitive and negative symptoms of schizophrenia in an attractor framework and is consistent with a close relation between the cognitive and negative symptoms: blockade of NMDA receptors by dissociative anesthetics such as ketamine produces in healthy subjects schizophrenic symptoms , including both negative and cognitive impairments [45 , 46]; agents that enhance NMDA receptor function reduce the negative symptoms and improve the cognitive abilities of schizophrenic patients [47]; and the cognitive and negative symptoms occur early in the illness and precede the first episode of positive symptoms [12 , 19 , 20] . Consistent with this hypothesized role of a reduction in NMDA conductances being involved in schizophrenia , postmortem studies of schizophrenia have identified abnormalities in glutamate receptor density in regions such as the prefrontal cortex , thalamus , and the temporal lobe [22 , 47] , brain areas that are active during the performance of cognitive tasks . The dopamine D1 receptor has been shown to modulate the performance of working memory tasks [30 , 48–50] . An increase in D1 receptor activation has been shown to increase the NMDA current [24 , 26] , and modeling studies have shown that this increase is related to the stability of working memory states [21 , 25 , 27] . Imaging data also support the importance of the D1 receptor in schizophrenia [51 , 52] . We therefore suggest that an increased activation of D1 receptors might alleviate both the cognitive and the negative symptoms of schizophrenia [53 , 54] by increasing NMDA receptor-mediated synaptic currents ( Figure 7 ) . Atypical neuroleptics might use this mechanism by not only blocking D2 receptors , but also by increasing the presynaptic release of dopamine , which in turn would increase the activation of the extrasynaptic D1 receptors [48 , 55] . Taken together , we suggest that the cognitive and negative symptoms could be caused by the same synaptic mechanism , namely a reduction in the NMDA conductance , which reduces the stability and increases the distractibility of the persistent attractors , and reduces the activity ( firing rates ) of neurons ( Figure 7 , middle column ) . The reduced depth of the basins of attraction can be understood in the following way . Hopfield [4] showed that the recall state in an attractor network can be thought of as the local minimum in an energy landscape , where the energy would be defined as where yi and yj are the firing rates of the ith and jth neurons in the network , which are connected by synaptic weight wij . In general , neuronal systems do not admit an energy function . Nevertheless , we can assume an effective energy function: in fact , the flow picture shown in Figure 1 resulting from the mean-field reduction associated with the spiking network analyzed here can be viewed as an indirect description of an underlying effective energy function . From this equation , it follows that the depth of a basin of attraction is deeper if the firing rates are higher and if the synaptic strengths that couple the neurons that are part of the same attractor are strong . ( The negative sign results in a low energy , and thus a stable state , if the firing rates of the neurons in the same attractor and their synaptic coupling weights are high . ) If we reduce the NMDA receptor–activated channel conductances , then the depth of the basins of attraction will be reduced both because the firing rates are reduced by reducing excitatory inputs to the neurons , and because the synaptic coupling weights are effectively reduced because the synapses can pass only reduced currents . The positive symptoms ( Figure 7 , right column ) of schizophrenia include delusions , hallucinations , thought disorder , and bizarre behavior . Examples of delusions are beliefs that others are trying to harm the person , impressions that others control the person's thoughts , and delusions of grandeur . Hallucinations are perceptual experiences that are not shared by others and are frequently auditory , but can affect any sensory modality . These symptoms may be related to activity in the temporal lobes [11 , 12 , 56] . The attractor framework approach taken in this paper hypothesizes that the basins of attraction of both spontaneous and persistent states are shallow ( Figure 7 ) . Due to the shallowness of the spontaneous state , the system can jump spontaneously up to a high activity state , causing hallucinations to arise and leading to bizarre thoughts and associations . This might be the cause for the higher activations in temporal lobe areas that are identified in imaging experiments [17 , 18] . We relate the positive symptoms to not only a reduction in NMDA conductance , but also to a reduction in GABA conductance . This is consistent with the fact that the positive symptoms usually follow the cognitive and negative symptoms and represent a qualitative worsening of the illness [12] . Alterations in GABA receptors have been identified in schizophrenia [23 , 57] . D2 receptor antagonism remains a main target for antipsychotics [58 , 59] . Dopamine receptor D2 antagonists mainly alleviate the positive symptoms of schizophrenia , whereas the cognitive and negative symptoms persist , especially for the typical neuroleptics [12] . Together with the simulations , our hypothesis suggests that an increase in the GABA current in the state corresponding to the positive symptoms ( −NMDA , −GABA ) might have the same effect as D2 antagonists . The therapeutic effect of D2 antagonists might thus be caused by an increase in GABA currents . Indeed , it has been found that D2 receptors decrease the efficacy of the GABA system [60 , 61] . ( For example , the application of D2 antagonists prevented a decrease in evoked inhibitory postsynaptic current amplitude produced by dopamine [60] . ) Thus , D2 antagonists would , in a hypersensitive D2 receptor state [62 , 63] , increase GABA inhibition in the network , and we suggest could increase the stability of attractor networks involved in the positive symptoms of schizophrenia , and thus ameliorate the positive symptoms . Since the concentration of dopamine in the cortex depends on cortical–subcortical interactions [64] , the causes of the described changes could also result from subcortical deficits . A detailed analysis of these feedback loops would require specific modeling . Earlier accounts of the relation of dopamine and schizophrenia in the cortex [24 , 60] have suggested two distinct states of dopamine modulation . One is a D2 receptor–dominated state in which there is weak gating and in which information can easily affect network activity . The other is a D1 receptor–dominated state in which network activity is stable and maintained . We have proposed a more detailed account for stability and discussed this separately for the spontaneous and persistent attractor states . This allows us to account for the dichotomy between the cognitive/negative and positive symptoms . We emphasize that in biophysically realistic network simulations , excitation and inhibition are not merely antagonistic but implement different functions in the network dynamics . Thereby , our modeling approach provides a missing link between the symptoms of schizophrenia and network models of working memory and dopamine [21 , 25 , 27] . This approach is not meant to provide a detailed discussion of specific symptoms of schizophrenia . Further research would be needed to relate specific symptoms to the overall scheme presented here in this paper . We concentrated in our study on basic effects in local cortical dynamics based on biophysically realistic spiking networks . One can extend the model to study the interaction between large-scale networks involving more than one location in the cortex and/or subcortical regions . This increases the dimensionality of the system and adds to the issues of stability [65 , 66] . Processes such as bubbling transitions and chaotic attractors also add to the notion of stability [67 , 68] . For high-dimensional systems , Kaneko uses the concept of a Milnor attractor [69] , which offers a way to formally describe the stability of an attractor . This approach measures the stability of an attractor against perturbations by introducing the concept of “return probability . ” By sampling over random perturbation and orbit positions , the return probability is defined as the fraction of trials in which the system returns back to the original point . This measures the strength of the attractor . In our work , we wanted to stay close to paradigms of working memory and address mechanisms involved in cognitive processes , attention , and distractibility . Consequently , our quantitative measures of stability are intrinsically related to these paradigms and therefore use concepts such as the escape probability instead of the return probability . Indeed , we introduced these measures in order to relate the quantitative concept of stability more directly with the cognitive symptoms . Nevertheless , in detailed investigations of specific symptoms of schizophrenia , the concepts mentioned above should be kept in mind , as they might play a role in neural dynamics . Cohen and collaborators focused in a series of computational and experimental studies on the cognitive symptoms of schizophrenia . Their connectionist models try to account for specific experimental data on context-processing deficits using working memory , gating , and attentional selection as mechanisms [70–72] . The context-processing hypothesis is compatible with the hypothesis of working memory deficits in schizophrenia , as context processing relies on stable working memory: a deficit in working memory would consequently lead to a deficit in context processing . Our model is conceptually at another level , featuring biophysically plausible single-neuron properties and specific synaptic currents . It is not intended to account for specific experimental data , but provides insights at a more generic and also biological level . In this sense , the two models complement each other . We discussed a possible cause for the proposed alterations of the attractor landscape related to schizophrenia , namely changes in NMDA and GABA conductance , as these are directly related to schizophrenia [22 , 23] . We did not investigate changes in AMPA conductance . In this particular model , the contribution of the AMPA current is relatively small [21] . A more detailed investigation could also include AMPA conductance , especially because it is known to be influenced by NMDA synaptic plasticity [73] . Indeed , if reduced NMDA currents led in turn by synaptic plasticity to reduced AMPA currents , this would amplify the effects we describe . The proposed alterations in the attractor landscape could have a variety of causes at the neurobiological level: abnormalities in glutamate and GABA receptors and signaling , modulations in synaptic plasticity , aberrant dopamine signaling , reduced neuropil , genetic mechanisms , and brain volume reduction [12 , 30 , 74 , 75] . Besides cortical mechanisms , cortical–subcortical dynamics could also cause the proposed alterations in the cortical attractor landscape , for example , via neuromodulatory influences such as dopamine or serotonin or cortical–subcortical feedback loops [64 , 76] . Our general hypothesis regarding the attractor landscape is meant to describe the aberrant dynamics in cortical regions that could be caused by several pathways . Future work could analyze further how changes of different factors such as regional differences , subcortical–cortical networks , or even more detailed neural and synaptic models might influence the stability of the type of neurodynamical system described here . We envision that our hypothesis could serve as a useful guideline for further theoretical work .
Our aim is to investigate stability and distractibility in a biophysically realistic attractor framework so that the properties of receptors , synaptic currents , and the statistical effects related to the probabilistic spiking of the neurons can be part of the model . We use a minimal architecture , a single-attractor or autoassociation network [3–5 , 41 , 77] . We chose a recurrent ( attractor ) integrate-and-fire network model which includes synaptic channels for AMPA , NMDA , and GABAA receptors [21] . These synaptic receptors are important , since there is evidence that alterations in synaptic currents are related to the symptoms of schizophrenia , and many of the drugs used to treat schizophrenia act on these receptor-mediated currents , either directly or indirectly [22 , 23] . Both excitatory and inhibitory neurons are represented by a leaky integrate-and-fire model [78] . The basic state variable of a single-model neuron is the membrane potential . It decays in time when the neurons receive no synaptic input down to a resting potential . When synaptic input causes the membrane potential to reach a threshold , a spike is emitted and the neuron is set to the reset potential at which it is kept for the refractory period . The emitted action potential is propagated to the other neurons in the network . The excitatory neurons transmit their action potentials via the glutamatergic receptors AMPA and NMDA , which are both modeled by their effect in producing exponentially decaying currents in the postsynaptic neuron . The rise time of the AMPA current is neglected because it is typically very short . The NMDA channel is modeled with an alpha function , including both a rise and a decay term . In addition , the synaptic function of the NMDA current includes a voltage dependence controlled by the extracellular magnesium concentration [79] . The inhibitory postsynaptic potential is mediated by a GABAA receptor model and is described by a decay term . The single-attractor network contains 400 excitatory and 100 inhibitory neurons , which is consistent with the observed proportions of pyramidal cells and interneurons in the cerebral cortex [80 , 81] . The connection strengths are adjusted using mean-field analysis [21] so that the excitatory and inhibitory neurons exhibit a spontaneous activity of 3 Hz and 9 Hz , respectively [82 , 83] . The recurrent excitation mediated by the AMPA and NMDA receptors is dominated by the NMDA current to avoid instabilities during the delay periods [84] . Our cortical network model features a minimal architecture to investigate stability and distractibility , and consists of two selective pools , S1 and S2 ( Figure 8 ) . We use just two selective pools to eliminate possible disturbing factors . The nonselective pool NS models the spiking of cortical neurons and serves to generate an approximately Poisson spiking dynamic in the model [21] , which is what is observed in the cortex . The inhibitory pool IH contains the 100 inhibitory neurons . The connection weights between the neurons of each pool or population are called the intrapool connection strengths w+ . The increased strength of the intrapool connections is counterbalanced by the other excitatory connections ( w− ) to keep the average input constant . The network receives Poisson input spikes via AMPA receptors that are envisioned to originate from 800 external neurons at an average spontaneous firing rate of 3 Hz from each external neuron , consistent with the spontaneous activity observed in the cerebral cortex [41 , 83] . A detailed mathematical description is provided in Protocol S1 . Our analysis is targeted to investigate the stability and distractibility with respect to NMDA and GABA receptor modulations . We use two different techniques: multiple trial spiking simulations and mean-field simulations . Spiking trial simulations integrate the complete neural and synaptic dynamics over time , including statistical components of the network model . Therefore , the spiking simulations are needed to assess the stability and distractibility of the dynamical system , for this depends in part on the statistical fluctuations that occur in a network of spiking neurons [85] . This is done by simulating a network configuration for several trials , each run with different random seeds , and running a statistical analysis on the data . We simulate three different conditions: the spontaneous , persistent , and distractor conditions ( see Figure 9 ) . In spontaneous simulations , we run spiking simulations for 3 s without any extra external input . The aim of this condition is to test whether the network is stable in maintaining a low average firing rate in the absence of any inputs , or whether it falls into one of its attractor states without any external input . In persistent simulations , an external cue of 120 Hz above the background firing rate of 2 , 400 Hz is applied to each neuron in pool S1 during the first 500 ms to induce a high-activity state , and then the system is run for another 2 . 5 s . The 2 , 400 Hz is distributed across the 800 synapses of each S1 neuron for the external inputs , with the spontaneous Poisson spike trains received by each synapse , thus having a mean rate of 3 Hz . The aim of this condition is to investigate whether the network , once in an attractor short-term memory state , can maintain its activity stably , or whether it falls out of its attractor , which might correspond to an inability to maintain attention . The distractor simulations start off like the persistent simulations with a 500-ms input to pool S1 to start the S1 short-term memory attractor states , but between 1 s and 1 . 5 s we apply a distracting input to pool S2 with varying strengths . The aim of this condition is to measure how distractible the network is . The degree of distractibility is measured parametrically by the strength of the input to S2 required to remove the high-activity state of the S1 population . These simulation protocols serve to assess the generic properties of the dynamical attractor system rather than to model specific experimental data obtained in particular paradigms . We used a mean-field approach ( described in Protocol S1 ) to calculate the stationary attractor states of the network for the delay period [21] . These attractor states are independent of any simulation protocol of the spiking simulations and represent the behavior of the network by mean firing rates to which the system would converge in the absence of statistical fluctuations caused by the spiking of the neurons and by external changes . Therefore , the mean-field technique is suitable for tasks in which temporal dynamics and fluctuations are negligible . It also allows a first assessment of the attractor landscape and the depths of the basins of attraction , which then need to be investigated in detail with stochastical spiking simulations . Part of the utility of the mean-field approach is that it allows the parameter region for the synaptic strengths to be investigated to determine which synaptic strengths will on average produce stable activity in the network , for example , of persistent activity in a delay period after the removal of a stimulus . For the spontaneous state , the initial conditions for numeric simulations of the mean-field method were set to 3 Hz for all excitatory pools and 9 Hz for the inhibitory pool . These values correspond to the approximate values of the spontaneous attractors when the network is not driven by stimulus-specific inputs . For the persistent state , a selective pool was set to a higher initial value ( 30 Hz ) to account for the excitation of these neurons during the preceding cue period . In addition , we used the mean-field technique to assess the flow between the attractor states . The flow is the force that drives the system toward the attractor given a parameter value in phase space , i . e . , the firing rates of the pools . Since we were interested in the depth of a single attractor , we used a setup with just one selective pool . This was done by fixing the value of the firing rate of the selective pool and letting the other values converge to their fixed point . Afterward , we computed the flow with this configuration [36] . | One of the hallmarks of schizophrenia is the complexity and heterogeneity of the illness . We propose that part of the reason for the inconsistent symptoms may be a reduced signal-to-noise ratio and increased statistical fluctuations in different cortical brain networks . The novelty of the approach described here is that instead of basing our hypothesis purely on biological mechanisms , we develop a top-down approach based on the different types of symptoms and relate them to instabilities in attractor neural networks . Schizophrenia is characterized by cognitive , negative , and positive symptoms . We propose which characteristic effects in a dynamical system could cause these symptoms , and investigate our hypothesis in a computational model . We implement an integrate-and-fire network model and focus on the alterations of synaptic channels activated via NMDA and GABA receptors . We found that a decrease in the NMDA receptor conductance could contribute to both the cognitive and negative symptoms by reducing the neuronal firing rates and the stability of the attractor states . A reduction of both NMDA and GABA conductance causes an instability of the attractor states related to the positive symptoms . Overall , we provide a framework to discuss schizophrenia in a dynamical system framework . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"spiking",
"dynamics",
"schizophrenia",
"attractor",
"networks",
"neuroscience",
"homo",
"(human)",
"neuronal",
"networks"
] | 2007 | A Dynamical Systems Hypothesis of Schizophrenia |
In trypanosomatids , gene expression is regulated mainly by post-transcriptional mechanisms , which affect mRNA processing , translation and degradation . Currently , our understanding of factors that regulate either mRNA stability or translation is rather limited . We know that often , the regulators are proteins that bind to the 3′-untranslated region; they presumably interact with ribonucleases and translation factors . However , very few such proteins have been characterized in any detail . Here we describe a genome-wide screen to find proteins implicated in post-transcriptional regulation in Trypanosoma brucei . We made a library of random genomic fragments in a plasmid that was designed for expression of proteins fused to an RNA-binding domain , the lambda-N peptide . This was transfected into cells expressing mRNAs encoding a positive or negative selectable marker , and bearing the “boxB” lambda-N recognition element in the 3′-untranslated region . The screen identified about 300 proteins that could be implicated in post-transcriptional mRNA regulation . These included known regulators , degradative enzymes and translation factors , many canonical RNA-binding proteins , and proteins that act via multi-protein complexes . However there were also nearly 150 potential regulators with no previously annotated function , or functions unrelated to mRNA metabolism . Almost 50 novel regulators were shown to bind RNA using a targeted proteome array . The screen also provided fine structure mapping of the hit candidates' functional domains . Our findings not only confirm the key role that RNA-binding proteins play in the regulation of gene expression in trypanosomatids , but also suggest new roles for previously uncharacterized proteins .
Kinetoplastid protists are exposed to environmental challenges in the host and vector , necessitating extensive changes in gene expression . In kinetoplastids , most protein-coding genes are transcribed by RNA polymerase II in unidirectional clusters . Individual mRNAs are excised by 5′ trans splicing , with addition of a 39mer spliced leader RNA , and by 3′ polyadenylation [1] . At the level of individual pol II-transcribed open reading frames ( ORFs ) , there is no evidence for control of expression via regulated transcription initiation . Trypanosomes and related parasites are therefore mainly dependent on post-transcriptional mechanisms for the control of gene expression [2] , and thus for both survival and pathogenesis . Control of gene expression at the post-transcriptional level is essential in all organisms , and RNA-binding proteins ( RBPs ) play critical roles at all stages: RNA processing , transport , stability and translation . The interactions of RBPs with cytosolic mRNAs - often , but not always , with the 3′-untranslated region ( 3′-UTR ) - can affect transcript stability , translational efficiency , or both [3] . In this way , RBPs can modify and regulate every step of RNA metabolism and function . Recent studies in different organisms have revealed that the repertoire of RBPs is far greater than anticipated [4]–[7] . Although members of all classical RBP domain families were abundantly represented in these analyses , several novel and unexpected candidates with no RNA-related ontology or domain homology were seen to interact with RNA . In addition to novel proteins containing low complexity amino acid repeat motifs , metabolic enzymes , and enzymes with potential RNA- and protein-modifying activities were found . However , for most of these novel RBPs , the functional consequences of the mRNA binding remain unknown . In trypanosomatids , RBPs are key factors in gene expression regulation . Over a hundred RBPs have been predicted to exist in trypanosomes , based on their containing canonical RNA-binding domains: RNA recognition motif ( RRM ) , CCCH zinc fingers , and pumilio or PUF domains . Some of these are already known to be involved in the regulation of differentiation , development , the cell cycle and rRNA processing ( recently reviewed in [8] ) . Most of the characterized RBPs control mRNA abundance by stabilizing the target transcript ( for example TbZPF3 [9] , TbZC3H11 [10] , TbZC3H20 [11] , TbPUF9 [12] ) or by increasing translation ( ALBA domain proteins , [13] ) . In mammalian cells , several proteins ( such as BRF1 , TTP , AUF1 and TIA-1 ) have been shown actively to destabilize mRNAs or to act as translation silencers ( [14] . In trypanosomatids , although indirect results suggest that some proteins cause degradation of target transcripts this has never been shown directly ( examples are UBP1 [15] , PUF6 [16] and PUF5 [17] . ) There is evidence that one protein , RBP10 , can suppress translation when attached to an mRNA [18] . With the single exception of ZC3H11 [19] , the mechanisms by which trypanosome proteins determine mRNA fate are not known . The functions of RBPs in vivo can be studied by “tethering” them to a reporter mRNA . In this technique , the protein under study is attached to the UTR of an mRNA reporter through an artificial RNA-protein interaction [20] . Hence , the functional activity of a protein can be studied independent of its intrinsic ability to bind to RNA . This assay has been applied successfully for numerous proteins in diverse organisms [21] and has proven useful in the analyses of the precise function of essential genes [22] , mapping of protein function [10] , dissecting functional maps of protein complexes [23] and visualizing tagged mRNAs [24] . The bacteriophage lambda-N protein is often used in the tethered function assay [25] . N-protein regulates bacterial transcriptional anti-termination by binding to a 15-nucleotide RNA hairpin , called boxB , within nascent transcripts [26] . The N-peptide is used in tethering assays because its small size ( only 22 amino acids ) limits the extent of likely effects on the function of the fused protein , and because of its high affinity interaction with boxB RNA . In this paper , we use the “tethering” approach in a genome-wide screen for post-transcriptional regulators in Trypanosoma brucei . We have been able to take advantage of the excellent tools for advanced genetics and very high efficiency transfection of trypanosomes [27] . Also , since introns are almost completely absent and coding sequences are closely spaced , all open reading frames could be represented in a library made from genomic DNA . We were able to map functional domains on well-characterized proteins and found evidence for functions of many RBPs . RNA-binding properties were also analysed using a targeted protein array . The results greatly increase our knowledge of the possible regulatory repertoire in trypanosomes and provide functional annotation for many previously uncharacterized genes .
To screen for proteins that increase gene expression , we used a reporter cell line constitutively expressing the blasticidin resistance ( BLA ) protein . The BLA mRNA contains five copies of the boxB hairpin RNA sequence between the resistance cassette and the actin ( ACT ) 3′-UTR ( BLA-B-ACT ) . As a secondary reporter , a cassette encoding GFP ( GFP-B-ACT ) was introduced upstream of the blasticidin reporter ( Figure 1A ) . Cell lines containing the BLA reporter mRNA lacking the boxB element were used as controls . The tandemly arranged reporters were integrated into the trypanosome alpha-beta tubulin repeat , and should be transcribed constitutively by RNA polymerase II reading through from upstream . We first introduced a plasmid that inducibly expressed trypanosome poly ( A ) -binding protein 1 ( PABP1 ) with the lambda-N peptide at the N-terminus and a myc tag at the C-terminus ( N-PABP1-myc ) . Tethering of N-PABP1 to a trypanosome mRNA is known to increase mRNA stability and reporter expression [28] . As we had hoped , co-expression of N-PABP1-myc enabled the cells to grow at 250 µg/ml blasticidin - 50 times the normal concentration - whilst cells expressing only the BLA-B-ACT reporter were rapidly killed ( Figure 1B ) . At 25 µg/ml blasticidin ( 5-fold normal ) , cells containing the BLA reporter with the boxB element grew slightly faster than the control ( Figure 1B ) ; this was independent of tetracycline addition , probably because a small amount of N-PABP1-myc was expressed in the absence of tetracycline ( Figure 1D ) . As expected , tethering of PABP1-myc also increased GFP expression ( Figure 1D ) . To find out whether we could also detect the activity of proteins that decreased expression , we tethered the deadenylase CAF1 , which is known to result in reporter mRNA degradation [29] . Indeed , expression of the N-CAF1-myc decreased cell survival under low ( 25 µg/ml ) blasticidin pressure ( Figure 1C ) . Inhibition of GFP protein expression in both cell lines confirmed specificity ( Figure 1D ) . We next designed a screen that would positively select for proteins that decreased expression upon tethering . In bloodstream trypanosomes , inducible expression of cytosolic phosphoglycerate kinase B ( PGKB ) is lethal [30] . We therefore created a cell line with a tetracycline inducible PGKB open reading frame followed by 5 boxB copies at the 3′-UTR ( Figure 1A ) . Upon tetracycline induction , the parasites died ( not shown ) . As expected , co-expression of N-CAF1-myc ( “destabilizing” , Figure 1E ) expression conferred a selective advantage while N-GFP-myc ( “neutral” , Figure 1F ) and N-PABP-myc ( “stabilizing” , Figure 1G ) did not rescue the parasites . We concluded that our two complementary selections could be used to discover proteins affecting mRNA-fate . To screen for proteins involved in mRNA-fate regulation , a plasmid library was constructed using randomly sheared trypanosome genomic DNA . The plasmid backbone was designed such that transcription of each genomic DNA fragment could be induced with tetracycline , and the resulting proteins would contain the lambda-N peptide at the N-terminus . 3×106 independent clones were obtained with an average insert size ( as judged by plasmid-specific PCR ) of about 1 . 2 kbp . Based on the gene density within the coding regions of trypanosome chromosomes we expected approximately one in twelve plasmids to encode an authentic trypanosome protein fragment fused in frame to the N-peptide . This would indicate that our plasmid pool represents a ∼10-fold coverage of the open reading frames . The inducible library was transfected into bloodstream cells expressing both the BLA-B-ACT and GFP-B-ACT reporters , using site-specific endonuclease-enhanced transfection [31] . In the first experiment ( BLA-A ) , 1 . 8 million clones were obtained and in a second ( BLA-B ) , 3 . 3 million clones ( Supplementary Figure S1 ) . To determine the sequence coverage of the fusions in the trypanosome populations , the plasmid inserts were amplified by PCR , using primers located within the lambda-N peptide and immediately 3′ to the cloning site . Control experiments with the resistance mRNA lacking the boxB elements were performed once with 3 . 1 million clones ( experiment BLA-C ) . We induced N-peptide fusion protein expression for 24 h and then grew cells for four days under different conditions . Clones that were unable to grow after tetracycline induction must express proteins that have intrinsic dominant-negative effects , unrelated to blasticidin resistance ( Figure 2 , black cloned inserts and trypanosomes ) . In addition , after 24 h induction we grew the cells for four days in various concentrations of blasticidin . To detect fusion proteins that decrease mRNA stability or translation ( Figure 2 , blue ) we grew the cells in 1x ( experiment BLA-A ) or 2x ( experiment BLA-B ) levels of blasticidin . No significant effect on overall population growth was seen ( Supplementary Figure 1A , C ) . To detect fusion proteins that increase mRNA stability or translation ( Figure 2 , red ) we grew the cells in 6x ( experiment BLA-A ) or 10x or 20x ( experiment BLA-B ) levels of blasticidin . In this case , growth was impaired but then the cells started to recover , indicating rescue by lambda-N fusions that could enhance BLA expression ( Supplementary Figure 1A , C ) . In parallel , we run a similar experiment with the cells carrying the BLA reporter without boxB elements ( Supplementary Figure 1D , experiment BLA-C ) . The cloned genomic DNA fragments were in each case recovered by plasmid-specific PCR . Examination of small aliquots from the reactions revealed similar-looking smears of many products for all populations except the high blasticidin treated cells ( Supplementary Figure 1B , E ) . All PCR mixtures were subjected to high-throughput sequencing . To screen for protein fragments that impaired expression , the plasmid library was transfected into bloodstream-form cells expressing the inducible PGKB-B-ACT reporter ( Supplementary Figure 2A ) . Three independent experiments were done , resulting in 0 . 65 million ( Library PGKB-A ) , 1 . 2 million ( Library PGKB-B ) and 4 . 8 ( Library PGKB-C ) million independent clones . Expression of both PGKB and the lambda-N fusions was induced with tetracycline and cells were grown in the presence of inducer for 5 days . In the libraries , the induced PGKB only slightly affected growth . Presumably , some of the parasites had lost inducible expression of the lethal protein during the transfection procedures . This is a common problem in trypanosomes that inducibly express toxic RNAs or proteins . Nevertheless , expression of lambda-N fusions did confer a modest growth advantage ( Supplementary Figure 2B ) . Again , the selected plasmid inserts were amplified ( Supplementary Figure 2C ) and sequenced . Sequencing reads from the blasticidin expression experiment were mapped to the T . brucei reference after trimming of the non-trypanosome boundary sequences . Only reads that were in frame with the lambda-N peptide were considered . Results are tabulated in Supplementary Tables S1 , S2 , S3 , with Supplementary Table S1 showing the reads for all mapped in-frame locations at the nucleotide level . Before selection , reads were fairly uniformly distributed with no strand or coding-region preference . However , under inducing conditions , in-frame coding-region reads on particular genes were strongly enriched . Figure 3A shows an example of aligned reads from the second blasticidin selection experiment , in the ∼12 kb region that surrounds the gene encoding the zinc finger protein ZC3H11 . After selection , the accumulation of reads over ZC3H11 was striking . Overall there were about 50-fold more in-frame reads from the ZC3H11 open reading frame after selection than in the unselected population . We preciously characterized the activity of ZC3H11 using the tethering assay , with a chloramphenicol acetyltransferase ( CAT ) reporter mRNA . We found that the entire region downstream of residue 185 was required for activity [19] . This portion contains , at residues 196–199 , a four-residue motif ( HNPY ) that is required for interaction with the post-transcriptional regulator MKT1 , and beyond that an additional region that interacts with the PABP-interacting protein PBP1 [19] . Consequently , any lambda-N fusion that commenced beyond residue 185 should not be selected , and fusions that commenced earlier , but did not include the C-terminus , should also be excluded . Figure 3B , which shows the individual reads , illustrates this: without selection reads were scattered on forward and reverse strands , but after selection , only fusions that commenced upstream of and including nt 477 ( residue 159 ) showed enrichment . However , detailed analysis showed that a fusion that should have been selected , commencing at residue 167 , was not enriched . This particular fusion protein may have non-functional folding pattern , or it may be truncated at the C-terminus . Interestingly , overall , many fusions that were selected by blasticidin were not detected in the population prior to selection . To further check the results we examined additional genes whose activities had previously been studied in the tethering assay . Both PABP1 and PABP2 yielded multiple clones that increased blasticidin resistance . Averaging the entire open reading frames , PABP1 gave 80-fold more reads after 6-fold or 10-fold blasticidin selection than were present without blasticidin , and PABP2 gave 25-fold read-count enrichment . Pab1p-binding protein , PBP1 - which is also active in the tethering assay [19] - gave over 150-fold enrichment . In each case , it is possible to deduce approximately , from the positions of selected fusions , the portion of the protein that is required for activity in the tethering assay . We now analyzed the entire dataset from the blasticidin selection experiment ( Supplementary Table S1 ) . To do this , we used a list of 6933 non-redundant protein-coding sequences predicted in the assembled T . brucei genome ( [32] with modifications ) ( Supplementary Table S2 ) . We obtained at least one in-frame read for more than 6700 ( >96% ) of these . The mapped sequence reads represented over 340 thousand independent locations , equivalent to >10 locations and >50 reads for each CDS ( Supplementary Table S2 ) . For data analysis we included reads that started from position −36 relative to the ATG , to maximize retention of N-terminal sequences . We also excluded all reads that would include less than 6 amino acids of the protein C-terminus . To obtain an overview of the data , and a list of candidate activators , we selected proteins from which at least two different fragments showed read-count enrichment of at least 3-fold , and there was also at least 3-fold read-count enrichment when the reads for the whole ORF were considered ( see Methods for details ) . We will henceforth discuss these proteins as activators , although strictly speaking we have only shown activation activity for at least two fragments . If only one or two clones from an ORF contain the entire region required for activity , averaging over the entire ORF will eliminate real effects . Our selection was therefore biased against long proteins , and against any proteins for which regions near the N-terminus are required for activity . A few proteins that were not covered at all in experiment BLA-A were also excluded . For these reasons it is essential to refer to Supplementary Table S1 in individual cases . Applying the stringent criteria listed above , there were 197 putative up-regulating proteins . We classified them functionally using TritrypDB annotations supplemented by other published information ( Supplementary Table S3 ) . Although a huge variety of proteins was classified among the activators ( Figure 4A ) , known activators were included and both RBPs and translation factors were strongly enriched ( Figure 4A , Table 1 ) . We also found unexpected enrichment of cytoskeleton-associated proteins . Although a role for the cytoskeleton as translation regulator is possible [33] , more probably the enrichment is an artifact of the alignments because the proteins contain sequence repeats ( discussed in more detail later ) . Proteins containing Pfam domains related to RNA binding ( for example , RRM and zinc finger variants ) were the most significantly increased group ( Figure 4A ) . This is very encouraging since these proteins might be expected to have activity even when not tethered via lambda-N ( Table 1 ) . The 25 active proteins include 15 zinc finger proteins and 10 RRM domain proteins . Several of the RBPs that increased expression in our tethering screen have been shown to increase the abundance or translation of their target mRNAs , but have not previously been analyzed by tethering ( Supplementary Figure S3 ) . These include ZC3H20 [11] , PTB1/DRBD3 [34] , [35] , and PTB2/DRBD4 [35] . RBP42 , which was also found in the enhancing group , is associated with coding and non-coding regions of abundant mRNAs [36] , which would be consistent with a stabilizing function . Proteins involved in translation - especially initiation factors - were over-represented in the activating population . The trypanosome genome encodes multiple homologues for the eIF4A ( two ) , eIF4E ( four ) and eIF4G ( five ) subunits . Of these , only eIF4E3 and eIF4E4 , which are in complex with eIF4G4 and eIF4G3 respectively , are thought to be active in T . brucei translation initiation [37] . Supporting this , we found different fragments of the initiation factors eIF4E3 and 4 to increase reporter expression; but the full set of eIF4Gs appeared to be active ( Supplementary Figure S4 ) . Other high-scoring proteins related to translation are listed in Table 1 . We now examined the fusion protein fragments that were positively selected in the presence of PGKB-B-ACT ( for details , see Methods ) . The strongest suppressors of expression were also negatively selected in the BLA experiments after addition of 5–10 µg/ml ( 1–2x ) blasticidin ( Supplementary Tables S4 and S5 ) . We found 127 proteins that gave reproducible repression of gene expression ( Supplementary Table S3 , and S4 sheets 3 & 4; for definition of “reproducible” see methods ) . As for the activators , the discussion below refers to whole proteins although we have actually shown activity only for at least two fragments . Reassuringly , three components of the CAF1/NOT deadenylation complex NOT2 , NOT11 ( C2ORF29 ) and NOT5 [38] reproducibly suppressed PGKB expression , suggesting these subunits could recruit active NOT/CAF1 complex to associated mRNA . In contrast , neither CAF1 nor UBP1 , both of which are active as full-length proteins ( this paper , [28] , [29] ) were selected , and neither were the two CAF/NOT complex proteins CNOT10 or CAF40 , presumably because larger fragments or the intact protein are required . The ORF that encodes the 5′-3′ exoribonuclease XRNA [39] is over 4 kb long and three tethered fragments that gave really clear suppression all contained the C-terminal quarter of the protein ( Supplementary Figure S5 ) . As shown for both human and Drosophila melanogaster XRN1 [40] , we speculate that the C-terminus may recruit other proteins that initiate mRNA degradation . Multiple potential RBPs were able to suppress PGKB expression . These included RBP10 , which was previously shown to inhibits translation if tethered to a reporter [18] and DRBD12 ( Tb927 . 7 . 5380 ) , which was shown to destabilize ARE-containing targets [41] ( Supplementary Figure S5 ) . In total , we found 16 RBPs conferring a clear negative effect on reporter expression , from which 7 were zinc finger proteins , 8 had a RRM motif and one , PUF3 , a pumilio domain ( Table 2 ) . In mammalian cells , cytosolic pumilio domain proteins are mostly known as translational repressors [42] . One of the most dramatic and reproducible suppressors we found was the eIF4E-interacting protein . The Leishmania homologue of this protein was recently shown to interact with eIF4E1 in a stage-specific fashion [43] , and it was suggested that this keeps eIF4E1 inaccessible for translation . Finally , several proteases , peptidases and components of the ubiquitination machinery were found to suppress expression when tethered . We speculate that in these cases the nascent polypeptide was subject to proteolysis . For many of the “hits” in our experiments , there was no previous indication for involvement in post-transcriptional regulation . 68 proteins that decreased expression , and 88 that increased it , have no previously annotated function at all ( Figure 4B ) . There are also some metabolic enzymes in the list . It is tempting to speculate that these have regulatory function in addition to catalytic activity , as suggested for animal cells [44] , [45] . Mass spectrometry studies of proteins poly ( A ) + RNA in human cells catalogued more than 40 metabolic enzymes [4] , [5] . Because of the possibility of false-positives ( discussed below ) , it was important to validate some of our novel results using full-length clones . To do this , trypanosomes that constitutively expressed an mRNA encoding the CAT reporter with 5 boxB elements ( CAT-B-ACT ) were transfected with inducible lambda-N-myc fusion proteins . Most observed CAT activities were in qualitative agreement with the screening results ( Figure 5 ) . Quantitative agreement was not expected , since the measurement methods were so different . Cytidine deaminase was revealed as a false-positive , while ZC3H45 was a false-negative since tethering of full-length ZC3H45 decreased CAT expression . Remarkably , the full-length cap-binding protein eIF4E1 vastly decreased CAT expression , although it was negative in the PGKB experiment and a fragment had increased expression in the BLA screen ( Supplementary Figure S4 ) . Concentrating on the novel proteins , the Tb927 . 7 . 2780 fusion increased reporter CAT activity around 7-fold , as observed for PABP1 [28] . This protein has short poly-glutamine and -histidine tracts ( Figure 5A ) and was previously found to be essential in a high-throughput RNAi screen [31] . It also co-purifies with two proteins previously shown to increase mRNA stability , ZC3H11 [10] and MKT1 [19] . Tethering of the full-length pyrroline-5-carboxylate reductase ( P5CR , not in the list ) caused a modest but reproducible 1 . 6-fold CAT activity increase ( Figure 5 ) . The hypothetical proteins Tb927 . 11 . 14220 and Tb927 . 11 . 8020 decreased reporter expression; interestingly , SCOP analysis detected an RNA-binding domain in Tb927 . 11 . 14220 . Our screen had identified many proteins that affected mRNA translation or degradation when tethered artificially to an mRNA . In vivo , such proteins might be bound to mRNA directly , or via another protein . Alternatively , they might never be associated with mRNA , in which case the effect we had observed would be an artefact . To assess the RNA-binding properties of potential regulators , we constructed a custom protein array . We chose 384 proteins that are of interest in post-transcriptional regulation , including some translation factors , degradative enzymes , all proteins with identified RNA-binding domains and a subset of proteins that had been identified in the screen but had no known RNA-binding characteristics . For reasons of economy , we also excluded proteins of the mitochondrial inner membrane and matrix , cytoskeleton , vesicular transport pathway , surface proteins , nucleoporins , nuclear proteins , and a few proteins for which the ORFs were too large to be amplified by PCR . Gateway-compatible primers were designed ( Supplementary Table S6 ) , the ORFs were amplified , and the lengths of products confirmed by gel electrophoresis . The products were then re-amplified to create DNA templates for protein expression in a prokaryotic in vitro transcription-translation system . The resulting templates were spotted onto glass slides in triplicate , and then a transcription-translation mix was added to create protein arrays ( Supplementary Figure S6 ) . In order to minimise disruption of the protein structure our oligonucleotides included the native stop codon . This had the disadvantage that it was not possible to verify production of full-length protein by addition of a C-terminal tag . The templates did , however , encode an N-terminal His tag , production of which was verified ( not shown ) . The protein arrays were used for studies of RNA interactions by probing them with labelled total or poly ( A ) + RNA ( in each case two procyclic-form samples and one bloodstream-form sample ) . The results are summarized in Supplementary Table S7 . Negative results cannot be interpreted since we did not test whether the full-length proteins were made and there is no way to know whether RNA-binding domains , if they exist , were correctly folded . A negative result would also be obtained if a protein has very sequence specificity and its target mRNAs have low abundance . For example , the PABPs and pumilio domain proteins were negative . As a positive control , we probed ZC3H11 with its cognate recognition sequence of ( UAU ) repeats; it showed a clear signal whereas a C→S mutant in the zinc finger failed to bind ( not shown ) . 157 protein spots bound to total RNA in all three replicates , whereas 47 bound to poly ( A ) + RNA ( of which all but two were also positive on total RNA ) ; 148 spots bound in at least 4 of the 6 experiments . The small RNA binding protein RBP3 gave the strongest signal ( Supplementary Table S7 ) ; 43 other RNA-binding domain proteins were also positive in at least 5 slides . Proteins that showed some RNA binding included some with confirmed tethering activity ( Figure 5 ) but no known RNA-binding domain: Tb927 . 10 . 15760 ( 5 slides ) , Tb927 . 1 . 3070 ( 5 slides ) , Tb927 . 7 . 2780 ( total RNA only ) and pyrroline-5-carboxylate reductase ( 3 slides ) . Other rather surprising positives were the two peroxins , PEX13 ( 3 slides ) and PEX14 ( 5 slides ) , which are in the glycosomal membrane , and a variety of enzymes . From these results we concluded that many of the proteins that we had identified in our tethering screen - including those with no annotated RNA-binding domains - may indeed be capable of binding RNA . The remainder may influence mRNA or translation via interactions with other RNA-associated proteins .
In this study we have shown that genes regulating mRNA turnover and translation can be identified using a functional genomics approach . The screen identified mRNA regulation by canonical RBPs and by proteins that act via multi-protein complexes , and also revealed nearly 100 proteins that can bind RNA but had no previously known RNA-related function . The random shotgun approach has the additional power of allowing delineation of functional domains . The tethering approach is very powerful but , as in all high-throughput screens , false-positives and -negatives can occur: Despite these possible problems , the screen was overall highly informative . We envisage that this technique could be adapted to screen libraries for proteins that play a role in post-transcriptional control in other organisms where transfection methods with high efficiency are available . For T . brucei , this forward genetic approach could be easily used to identification of genes and pathways regulating other complex biological phenomena such as quorum sensing , antigenic variation , or drug target identification . Our set of 384 ORFs already provides a Gateway-compatible resource for proteins potentially involved in mRNA metabolism . The RNA binding of yeast [48] and mammalian [49] proteins has previously been examined using microarrays , but it is not clear which of the identified RNA binders was also found in the datasets obtained by poly ( A ) + mRNA precipitation . Protein microarrays have the advantage ( unlike poly ( A ) + mRNA precipitation ) of being unaffected by in vivo protein abundance; however , false negatives may be caused by incorrect protein folding and there is the danger of un-physiological interactions such as electrostatic binding of basic proteins to nucleic acids . Nevertheless , the results do provide preliminary indications which could be followed up by in vivo RNA cross-linking studies . In future , probing with more specific sequences may allow us to determine the RNA-binding specificities of the candidate proteins . The results described in this paper will be most useful in the context of other available datasets . A genome-wide RNAi screen provided a catalogue of T . brucei genes whose knock-down is detrimental to the parasite under a variety of developmental conditions [31] . This can be now be used , for example , to look for proteins that are essential in only one life-cycle stage , and are implicated in the post-transcriptional control of gene expression ( These results are included in Supplementary Table S7 ) . For bloodstream forms , these would include ZC3H5 , ZC3H15 , and numerous proteins of no known function . An RNAi screen for proteins associated with the AMP/cAMP response identified a number of proteins whose association with either signalling or control of gene expression were not apparent [50] . Of these , Tb927 . 11 . 2250 and Tb927 . 9 . 4080 are both up-regulators and Tb927 . 11 . 2250 binds RNA . Adenylosuccinate synthetase was suggested to affect differentiation via AMP metabolism: we also identified it as a weak expression down-regulator with RNA-binding properties . The catalogues of mRNA-associated proteins that are already available for mammals [4] , [5] , [7] and yeast [6] include numerous metabolic enzymes . Application of our methods to Opisthokont systems would clearly facilitate interpretation of these mRNA-associated protein datasets . We have now documented potential regulatory functions for trypanosome cystathione gamma lyase ( CTH ) , deoxyribose-phosphate aldolase , pyrroline-5-carboxylate reductase ( P5CR ) , dihydrofolate reductase-thymidylate synthase ( DHFR-TS ) , adenylosuccinate synthetase and tryparedoxin ( TXN ) ; all but the first two also bound RNA in the protein microarray assay . In Hela cells , two other oxidative stress-related enzymes - thioredoxin and peroxiredoxin - have been shown to bind RNA [4] . In animal cells , thymidylate synthase ( TS ) and dihydrofolate reductase ( DHFR ) bind to their own mRNAs , causing translational repression [51] . Here we showed that in trypanosomes , the bifunctional trypanosome DHFR-TS protein is also able to bind RNA and , as in animal cells , the functional consequence of its association with mRNA is inhibition of expression . Results from yeast-two-hybrid interaction screens and affinity purifications will also assist with mechanism prediction . Proteins that showed activity in the screen but cannot themselves bind to RNA , may nevertheless be associated with RNA via interactions with RNA-binding proteins . For example , although the Tb927 . 9 . 4080 protein may not bind to RNA , it is known to interact with both DRBD3 [34] and MKT1 [19] . Down regulators might be expected to interact with components of the degradation machinery , or to have inactivating interactions with the translation apparatus . Proteins that enhance gene expression might be expected to be found stably associated with mRNA , and those that enhance translation ought to be at least partially associated with polysomes , so additional datasets of this sort will greatly facilitate interpretation of our data .
The overexpression library was constructed in a derivative plasmid of the tetracycline-regulated pHD678 [52] . It contains a lambda-N peptide sequence cloned as a HindIII-ApaI fragment , and the NEO ( neomycin/G418 resistance ) gene replaced a HYG ( hygromycin resistance ) . Specific oligonucleotides were ligated into the ApaI and BamHI sites to add a stop codon in all three possible reading frames downstream of the unique XhoI cloning site . Reporter blasticidin plasmids are derivative of the pHD330 [53] . They contain a GFP gene positioned upstream of a blasticidin ( BLA ) resistance marker and were designed for targeting into the tubulin locus . For the screening , 5 copies of the boxB sequence element were inserted immediately between the reporter sequences and the actin 3′-UTR . A similar construct lacking the boxB element between the BLA and the 3′-UTR was used as control . For the PGKB experiment , the full-length PGKB CDS was cloned into the pHD2300 . It confers BLA resistance and was designed for targeting into the RRNA locus . As described , 5 copies of the boxB element were embedded between the PGKB sequence and the actin 3′-UTR . Details of all plasmids and oligonucleotides are provided in Supplementary Table S6 , sheet 2 . Complete sequences are available from us upon request . The overexpression library was essentially made as previously described [19] . The plasmid was linearized at the unique XhoI site , filled in with the large fragment of DNA polymerase I in the presence of dTTP , and ligated to the semi-Xho-adapted DNA ( size range from 0 . 7–3 Kbp ) . The ligation reaction was used to transform Escherichia coli NEB 5-alpha cells by electroporation to generate a library of approximately 3×106 ampicillin-resistant colonies . To assess the quality of the library , E . coli cells were re-transformed with purified library , plated and plasmids purified from individual colonies . After XhoI digestion , inserts were found in 99% of plasmids and the average size was 1 . 2 Kbp . Bloodstream-form T . brucei 2T1 cells were maintained and transfected as described [31] except that Tb-BSF buffer [54] was used for all transfections . The pRPaSce* plasmid was used to derive Sce* cells from 2T1 cells as described [31] . The Sce* strain expresses the tetracycline repressor ( TetR ) , and inducible homing endonuclease ( I-SceI ) which facilitates site-specific overexpression library integration . Upon transfection , overexpression plasmid library constructs replace the I-SceI gene and cleavage site . Parasite libraries were generated by several rounds of electroporation and for each series , an aliquot of the transfection was diluted to determine the transfection efficiency . The average efficiency was about 5×10−3 . To assay for stabilizing proteins , populations expressing the blasticidin resistance mRNA were pre-induced for 24 h in 1 µg/mL tetracycline and then grown with various concentrations of blasticidin ( 1x = 5 µg/mL ) for four days . The PCR-amplified DNA was fragmented and sequenced using Illumina HiSeq with multiplexing and data analyzed as described previously [10] , [55] . We sequenced input trypanosome population twice since the BLA-selected trypanosome libraries were expected to be less complex . SAMtools [56] and custom-made PERL scripts were used to select only coding region or 5′-UTR sequences that were in frame with the lambda-N peptide . The lambda-N sequence was removed , then the remainder was mapped to the T . brucei 927 reference genome ( http://tritrypdb . org/tritrypdb ) using Bowtie , allowing one base mismatch . To find proteins that increased blasticidin resistance we chose locations from position −36 relative to the ATG to position −18 relative to the stop codon ( Supplementary Table S1 ) . The counts ranged from 166000 to over a million ( Supplementary Table S1 ) . Then only positions for a unique gene set ( [32] , modified ) were chosen . For these we counted , for each experiment , the reads per million ( RPM ) . To avoid zero values , 1 was added to each value to give RPM+1 ( Supplementary Table S2 , sheet 2 ) . After that , for each location , the number of counts for 6-fold increased blasticidin ( BLA 6X ) in experiment A was divided by the counts for −tet , and separately by the counts for cells with tetracycline but without blasticidin . To reduce the likelihood of identifying PCR artifacts , the lower of these two values was taken to be the relative enrichment for this site . Similar calculations were done for all locations for BLA10X in experiments B and C ( Supplementary Table S2 , sheet 3 ) . Experiment C was the -boxB control , which should select only for fusions that enhance growth independent of the blasticidin resistance mRNA , and also for PCR artifacts . We therefore removed all locations that gave at least 3x increase in BLA10x in Experiment C . From the remaining list we selected genes for which at least two locations gave a 3x increase in either BLA6x ( A ) or BLA10x ( B ) ( Supplementary Table S2 , sheet 3 ) . The number of such locations per ORF was counted ( Supplementary Table S2 , sheet 4 ) . Separately , and using a similar procedure , numbers of total counts for the whole coding regions were extracted ( Supplementary Table S2 , Sheet 5 ) . For the unique genes , we computed the RPM in each experiment ( Supplementary Table S2 , Sheet 2 ) . The RPM+1 with 6x or 10x blasticidin , relative to amounts without blasticidin were calculated as before ( Supplementary Table S2 , Sheet 3 ) . We now deleted any CDS giving at least three-fold enrichment in counts for BLA10x in negative control experiment C . We retained CDSs that gave an overall count enrichment of at least 3-fold in both experiment A ( BLA6x ) and experiment B ( BLA10x ) . The resulting 197 genes are listed in Supplementary Table S2 , sheet S4 and ( for readers who want a smaller and simpler file ) in Supplementary Table S3 , Sheet 1 . Table 1 shows only the RBPs and translation factors that reproducibly increased BLA expression . To find proteins that increased survival after PGKB expression , we took the counts for each location ( Supplementary Table S4 , sheet 1 ) , computed the RPM and added 1 to every value to avoid zero values ( Supplementary Table S4 , sheet 2 ) . We then divided the number of counts for +tet by the counts for −tet ( Supplementary Table S4 , sheet 3 ) . The highest ratio for each location was found . Now , for each unique gene , we counted the number of locations that gave at least 3-fold enrichment , + tetracycline , in at least one experiment ( Supplementary Table S4 , sheet 3 , column D ) . Separately , numbers of total counts for the whole coding regions were extracted ( Supplementary Table S5 , Sheet 1 ) . For the unique genes , we computed the RPM in each experiment , added 1 as before , and ratios +tet to −tet were calculated for each experiment ( Supplementary Table S5 , Sheet 2 ) . These were transferred to Supplementary Table S4 , sheet 3 columns E–F . In addition , we looked for the effects of blasticidin selection . Counts per CDS with BLA1x ( experiment A ) or BLA2x ( experiment B ) were divided by the values with no blasticidin and the higher value taken . These values were also transferred to Supplementary Table S4 , sheet 3 , columns H , I and J . We now chose genes that were candidates as down-regulators . First , we selected genes for which at least two locations gave 3-fold enrichment in at least one of the PGK experiments ( Supplementary Table S4 , sheet 4 , column D ) . From these , we selected genes that gave a 2-fold RPM increase over the whole coding sequence in at least two of the three PGKB experiments . This list of 127 genes is in Supplementary Table S4 , sheet 4 . The best regulators are those which gave selective advantage in the PGK experiment , and were selected against in the BLA experiment . Supplementary Table S3 , sheet 2 has the same information , without the calculations and the raw data . Gene-specific primers were designed and synthesized in 96 well plates ( Supplementary Table S6 ) . The corresponding ORFs were then amplified directly from genomic DNA using Q5 high fidelity DNA polymerase ( New England Biolabs ) , with 2 min at 98°C , 35 cycles ( 98°C for 15 sec , 52°C for 15 sec and 72°C for 2 min ) then 10 min at 72°C . Each primer included additional sequence suitable for amplification and directional Gateway cloning . In a second PCR amplification , these extra sequences on the amplification products were used in order to reamplify the genes while adding a T7 promoter and Shine-Dalgarno sequence upstream of the ATG ( Supplementary Table S6 ) , with 0 . 02 µL of the first PCR as template ( 98°C for 2 min , two cycles of 98°C for 15 sec , 45°C for 15 sec and 72°C for 2 min; then 35 cycles as with the initial PCR ) . These templates were then used for the production of protein microarrays as previously described [57] . Briefly , 900 pL of DNA template was arrayed in triplicates on Nexterion Epoxy slide E using a GeSIM non-contact nanoplotter . In a subsequent spotting round , 3 . 6 nL of the S30T7 high yield protein expression system was spotted onto of each template spot . The entire assembly was placed in a deep humidified chamber containing 50 µL of nuclease-free water in each of its wells , and then incubated at 37°C for one hour and a 30°C overnight . Slides were removed from the chambers , cleaned and stored at −20°C until use . Poly ( A ) + RNA was isolated from total RNA using Nucleotrap mRNA purification kit as recommended by the manufacturer . Total and poly ( A ) + RNA were biotin-labeled using Pierce RNA 3′ End Biotinylation Kit following the manufacturer's instructions . Protein arrays were removed from the freezer and placed directly into a blocking solution containing Hepes-KOH pH 7 . 9 , 10% glycerol , 40 U of RNaseout RNase inhibitor , 1% BSA , 100 µM ZnCl2 , 1x Halt Protease and phosphatase cocktail inhibitor , 40 µg/mL heparin , 1 mM DTT 50 µg/mL E . coli tRNA , 50 mM glutamic acid potassium salt , 0 . 1% Triton X-100 and 8 mM Magnesium acetate . The slides were blocked for 1 hour at room temperature , and then washed in washing buffer ( Blocking buffer without BSA ) . Slides were then incubated with labeled RNA in blocking buffer , containing 0 . 5 mg/mL E . coli tRNA overnight at 4°C . The slides were removed from the incubation chambers and washed 3x twenty minutes in washing buffer , and 3x five minutes in Nuclease-free water . For detection of total and poly ( A ) + RNA , the arrays were probed with cy3-labelled extravidin at a dilution of 1∶100 . The slides were air-dried in a ventilated oven at 30°C for one hour and scanned in a Tecan power scanner with 75% laser power and 500% PMT gain . Images were saved as TIFF files and later loaded into Genepix 6 . 0 for data extraction . The mean background signal for each spot was subtracted from the mean spot intensity , and the average intensity for all triplicates of each sample calculated . As controls , a mutant zinc finger protein ( ZC3H11 C→S ) which has been shown to have no RNA binding activity [10] , as well as the PCR negative control and expression mix alone were used . | Survival and adaptation of trypanosomatids to new surroundings requires activation of specific gene networks . This is mainly achieved by post-transcriptional mechanisms , and proteins that bind to specific mRNAs , and influence degradation or translation , are known to be important . However , only few such proteins have been characterized to date . The trypanosome genome encodes over 150 proteins with conserved RNA-binding domains , and it is very likely that additional proteins that do not have such domains could also modulate mRNA fate . Here , we report the results of a genome-wide screen to identify mRNA-fate regulators in Trypanosoma brucei . We used a method called “tethering” to artificially attach protein fragments to an mRNA . Our findings confirmed the role of RNA-binding proteins in the regulation of mRNA fate , and also suggested such roles for many other proteins , including some metabolic enzymes . Our results should serve as a useful resource . Moreover , the tethering screen approach could readily be adapted for use in other organisms . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"biochemistry",
"cell",
"biology",
"biology",
"and",
"life",
"sciences",
"proteomics",
"microbiology",
"molecular",
"cell",
"biology"
] | 2014 | A Genome-Wide Tethering Screen Reveals Novel Potential Post-Transcriptional Regulators in Trypanosoma brucei |
Deciphering the architecture of the tRNA pool is a prime challenge in translation research , as tRNAs govern the efficiency and accuracy of the process . Towards this challenge , we created a systematic tRNA deletion library in Saccharomyces cerevisiae , aimed at dissecting the specific contribution of each tRNA gene to the tRNA pool and to the cell's fitness . By harnessing this resource , we observed that the majority of tRNA deletions show no appreciable phenotype in rich medium , yet under more challenging conditions , additional phenotypes were observed . Robustness to tRNA gene deletion was often facilitated through extensive backup compensation within and between tRNA families . Interestingly , we found that within tRNA families , genes carrying identical anti-codons can contribute differently to the cellular fitness , suggesting the importance of the genomic surrounding to tRNA expression . Characterization of the transcriptome response to deletions of tRNA genes exposed two disparate patterns: in single-copy families , deletions elicited a stress response; in deletions of genes from multi-copy families , expression of the translation machinery increased . Our results uncover the complex architecture of the tRNA pool and pave the way towards complete understanding of their role in cell physiology .
Messenger RNA translation is a central molecular process in any living cell and is among the most complicated and highly regulated of cellular processes [1] , [2] . The tRNA pool is a fundamental component in that process , serving as the physical link between the nucleotide sequence of mRNAs and the amino acid sequence of proteins . In the cycle of translation elongation , tRNA selection is considered the rate-limiting step [3] , therefore tRNA availability is one of the major factors that govern translation-efficiency and accuracy of genes [4] , [5] . Previous studies have established that efficient translation can increase protein levels and provide a global fitness benefit by elevating the cellular concentrations of free ribosomes [6] , [7] , while accurate translation benefits the cell by reducing the metabolic cost of mis-incorporation events [8] . The tRNA pool is composed of various tRNA isoacceptor families , each family carries a different anti-codon sequence that decodes the relevant codon by Watson-Crick base pairing , or codons with non-perfect base pairing of the third nucleotide by the wobble interaction . tRNA families are further classified to isotypes if they carry the same amino acid . In all eukaryotic genomes , each tRNA family can be encoded by a single or multiple gene copies [9] , [10] . It was previously shown for several organisms that the concentrations of various tRNA isoacceptors positively correlates with the tRNA family's gene copy-number [11] , [12] . These observations along with detailed analysis of the relationship between gene copy-number of tRNA families and codon-usage , established the notion that the multiplicity of tRNA genes in yeast is not functionally redundant . Such multiplicity might establish the correct balance between tRNA concentrations and the codon usage in protein-coding genes [13] , thus justifying the use of the tRNA gene copy-number as a proxy for actual tRNA amounts [12] , [14] , [15] . The transcription of tRNA genes is catalyzed by RNA polymerase III ( pol III ) , promoted by highly conserved sequence elements located within the transcribed region [16] . A genome wide analysis of pol III occupancy in yeast revealed that virtually all tRNA genes are occupied by the pol III machinery [17]–[19] , and are thus considered to be genuinely transcribed . This observation , combined with the fact that tRNA genes within a family are highly similar , led to the notion that all copies within a family contribute equally to the total expression level and hence to the tRNA pool . Although tRNAs have been extensively studied , until very recently many of the studies were performed on individual genes at the biochemical level . Only in recent years systematic genome-wide approaches started to complement the biochemical approach . These studies reveal a much more complex picture in which pol III occupancy , a proxy for tRNA transcription , varies within families and between tissues [17] , [20]–[22] . Expression however does not equal function , and so far no systematic study has been carried out to decipher the specific contribution of each tRNA gene to the tRNA pool and to the cell's fitness . To study the role of individual tRNAs and the architecture of the entire tRNA pool , we created a comprehensive tRNA deletion library in the yeast Saccharomyces cerevisiae . The library includes 204 deletions of nuclear-encoded tRNA genes out of the total 275 present in the yeast genome . In addition , we created double deletions of selected tRNA gene combinations and of specific tRNAs with a tRNA modifying enzyme . We developed a robotic method to screen and score various fitness parameters for these deletion strains , and applied it across various growth conditions . This systematic deletion library revealed an architecture of genetic interactions that feature extensive backup-compensations within and between tRNA families . Such compensation capacity endows the organism with robustness to environmental changes and to genetic mutations . We found that different copies within a tRNA family contribute differently to the organism's fitness , revealing a higher level of complexity in the tRNA pool's architecture , possibly at the regulatory level . Finally , we observed two distinct molecular signatures that underlie the cellular response to changes in the tRNA pool . First , the deletion of non-essential single-copy tRNA genes invoked proteotoxic stress responses , indicating a connection between aberrant tRNA availability and protein misfolding . Second , the deletion of representative tRNAs from multi-copy families triggered milder responses by up-regulating genes that are involved in the translation process . Together our results uncover the complex architecture of the tRNA pool reviling a profound effect on cellular fitness and physiology .
To gain a better understanding of the functional role of individual tRNA genes and their contribution to the tRNA pool , we created a comprehensive tRNA deletion library in S . cerevisiae , where in each strain a single nuclear–encoded tRNA gene was deleted . This methodology is based on recombining a selective marker into the genome at the expense of the deleted gene , as was done in the creation of the yeast ORF deletion library [23] ( Figure 1A ) . A particular challenge in targeting specific tRNA genes for deletion by such a method stems from the high degree of sequence similarity within tRNA families , which can share 100% sequence identity . Consequently , in order to create specific gene deletions , we relied on unique sequences that overlap or flank the tRNA genes ( see Supplemental text S1 ) . Our tRNA deletion library contained 204 deletions out of the 275 nuclear-encoded tRNA genes identified in S . cerevisiae ( see Materials and Methods ) . These deletions covered all 20 amino acids and 40 of the 42 anti-codon families . The remaining 71 tRNA genes were not deleted due to their complex genomic surrounding , since such deletions might affect neighboring potential features in their genomic vicinities . The library also consisted of 50 strains that represent various combinations of tRNA deletions , and co-deletions of selected tRNAs with the TRM9 gene which codes for an enzyme that post-transcriptionally modifies tRNA molecules . Although the majority of tRNA families contain multiple gene copies , there are six single-copy tRNA families in the S . cerevisiae genome . Out of these singleton families , four ( tS ( CGA ) , tR ( CCG ) , tQ ( CAG ) tT ( CGU ) ) were found to be essential in our analysis , which confirms previous reports [24]–[26] ( see Supplemental text S1 ) . The remaining two singleton families ( tR ( CCU ) , tL ( GAG ) ) were identified as non-essential upon deletion . All the tRNA genes that belong to multi-copy families were non-essential upon deletion . To assess the contribution of each tRNA gene to cellular growth , we attempted to accurately characterize the growth dynamics of each deletion strain by implementing a robotic method to screen and score growth phenotypes of all tRNA deletion strains in a given growth condition . This fitness measurement approach allowed us to differentiate between physiological effects of the deletion under different growth phases , unlike the competition approaches for fitness measurement [27] that typically integrated all growth phases . We characterized each deletion strain by two growth parameters: growth rate and growth yield , the latter is defined as the size of the population upon entering stationary phase ( Figure 1B and Supplemental figure S1A ) . We began the characterization of the tRNA deletion library by growing the strains in rich medium . Under this condition , 13% of the deletion strains demonstrated a phenotype in growth rate and 27% showed a growth yield phenotype ( Figure 1C–D and Supplemental figure S1B ) . Most strains exhibited a notable phenotype only in one of the two parameters . Strains that showed altered phenotypes in both growth rate and yield were rare ( Supplemental figure S1B ) . Overall , most tRNA deletion strains did not exhibit any altered growth phenotype in rich medium , indicating robustness to tRNA gene deletion . Seven percent of the tRNA deletion strains resulted in growth improvement , suggesting that for some genes the cost of retaining them in the genome and/or expressing them may exceed their benefit in this condition . Similar observations were also made on a selection of protein-coding genes in this species [28] . Apart from the singletons whose deletion strains were often dead or exhibit impaired growth , we could not explain the observed growth phenotypes in growth rate or yield by either tRNA family size or amino acid identity ( Supplemental figure S2 ) . To further examine the phenotypes of the tRNA deletion strains , we calculated the correlation to the mRNA expression level of adjacent genes and found none ( see , Supplemental table S1 figure S3 and Supplemental text S1 ) . Given that yeast cells are constantly exposed to varying environmental conditions , their tRNA repertoire should differentially accommodate growth in various environments . We next examine whether stressful conditions would retain the robustness observed in rich medium or reveal another set of condition-dependent growth phenotypes . We screened the deletion library under a diverse set of growth conditions including different metabolic challenges and stress-inducing reagents reported in previous studies [29]–[31] . The fact that the production of tRNA molecules is considered energetically costly [32] prompted us to explore the effect of carbon limitation , alternative carbon sources and minimal medium on tRNA essentiality . Growing the tRNA deletion library under stressful conditions revealed condition-specific phenotypes ( Figure 2A–D , Supplemental table S2 ) . In all but one of the examined conditions ( Dithiothreitol-DTT , a reducing agent that also inflicts a general protein-unfolding stress ) , robustness to tRNA gene deletion was maintained . In the DTT condition , the phenotypes were surprising: while multiple tRNA deletions exhibited impaired growth rates , many also demonstrated growth rate improvements ( Figure 2B , 2D ) . As in the rich medium condition , we could not explain the observed growth phenotypes by either the family size , or the amino acid identity in all of the examined stress conditions . Our observations of robustness to tRNA gene deletions in rich medium , as well as several stressful growth conditions , prompted us to further explore the genetic architecture conferring this phenotype . Given that most tRNA families contain multiple gene copies , we hypothesized that at least part of the observed robustness might be the outcome of compensation provided by the remaining genes in the family . In addition , due to wobble-interactions , robustness may also be the outcome of compensation between families of the same isotype . Focusing on rich medium conditions , we generated selected combinations of multiple tRNA deletions . To examine the first possibility we created deletions of entire two-member and three-member tRNA families . As shown in figure 3A such family deletions resulted in either lethality ( indicating a loss of the family's function ) , or viability with growth impairment ( indicating a partial compensation of the family's function by other families ) . We then turned to examine in more detail the interactions within these essential three-gene families by examining the growth of various double deletion strains . Contrary to the common notion that suggests little or no functional redundancy between tRNA gene copies [13] , we observed that in each of these families any one family member can sustain normal or near-normal fitness ( Figure 3A , Supplemental figure S4A–B and Supplemental table S3 ) . Similar observations were made for essential two-gene families upon one member's deletion ( Figure 3A ) . Such results can either imply that yeast cells carry more tRNA copies than are actually needed to sustain growth under optimal growth conditions , or that a responsive backup mechanism might be at work , one that provides compensation by increasing the transcription of the remaining copies , as was previously observed in protein-coding genes [33]–[35] . We thus decided to investigate the expression levels of certain tRNA families , using RT-qPCR ( Figure S5 ) . For each deletion , we compared the expression level of the remaining copies belonging to the designated family to that of a wild-type strain . We observed in most strains an expected reduction in expression of the respective family . These findings suggest that in these families , tRNA supply exceeds the demand under rich medium conditions ( Figure S5A ) . However in some cases there were no such decreases in expression , there were even observable increases , demonstrating that a responsive backup mechanism may have been at work , inducing the expression of the remaining family members following deletion of a certain member ( Figure S5B ) . Next , we turned to examine the surprising cases in which the deletion of an entire tRNA gene family resulted in a viable strain . We reasoned that in these cases a different type of compensation , which is based on wobble interactions across iso-acceptor families , came into play . To decipher this compensation mechanism we focused on the genetic interactions involving the two non-essential singleton families , tL ( GAG ) and tR ( CCU ) ( Figure 3A ) . In the absence of tL ( GAG ) , the members of the tL ( UAG ) family represent the sole tRNA that can decode the CUN Leucine codons , and might be a candidate for providing compensation upon deletion of tL ( GAG ) even though such decoding does not match the classic wobble rules [36] . Co-deletion of tL ( GAG ) with one of the tL ( UAG ) gene copies resulted in growth aggravation and negative epistasis . Deletion of the tL ( GAG ) together with two copies of the tL ( UAG ) family was lethal despite the fact that one copy of tL ( UAG ) still remained in the genome , indicating that a single tL ( UAG ) gene was insufficient to compensate for the loss of tL ( GAG ) ( Figure 3B ) . The genetic interaction between tL ( UAG ) and tL ( GAG ) appeared specific , since co-deleting one copy of the tL ( UAG ) family together with two additional tRNA genes ( tL ( CAA ) G3 and tW ( CCA ) G1 ) did not generate observable epistasis in either case ( Figure 3B ) . We thus concluded that the tL ( UAG ) family is partially redundant to the tL ( GAG ) family , yet such redundancy was not sufficient to completely compensate for the loss of tL ( GAG ) . Similarly , the viability of the tR ( CCU ) deletion strain could be due to compensation provided by the 11 copies of the tR ( UCU ) family . Indeed the wobble rules are consistent with this assumption , but such interaction was never functionally demonstrated . Formally , demonstrating that the tR ( UCU ) family can compensate for the loss of the singleton tR ( CCU ) would amount to co-deleting all 12 tRNA genes . Looking for simpler means , we decided on a more economic , albeit indirect way . We co-deleted the singleton tR ( CCU ) with the Trm9 enzyme , which is responsible for methylating the third anticodon position of tR ( UCU ) and tE ( UUC ) [37] . It was previously shown that such methylation is needed for supporting the wobble interaction between tR ( UCU ) tRNAs and the AGG codon ( the cognate codon of the CCU anti-codon ) [37] . The tR ( CCU ) –trm9 double deletion strain was viable , but exhibited an appreciable aggravation of growth yield ( Figure 3A and 3C ) . Thus our results confirm that the methylated tR ( UCU ) family can partially compensate for the loss of tR ( CCU ) . We attempted to define a more general role for the Trm9 modification enzyme in modulating the compensation mechanism between tRNA families . To this end we created 10 additional double deletions of the enzyme along with each of 10 tRNA genes from two glutamic acid families , one that is modified by the enzyme and one that is reportedly not modified by the enzyme [38] ( see Supplemental figure S6 ) . No epistasis was detected between the enzyme and any of these 10 tRNAs and hence , the data cannot support or exclude a putative similar role of the enzyme beyond the tR ( UCU ) family . We thus conclude that there are two mechanisms that can account for the observed robustness for tRNA deletions under favorable growth conditions . The first is redundancy within a family , and its efficiency appears to be independent of the number of remaining tRNA gene copies . The second is compensation between families , which operates via wobble interactions . We then asked whether all copies within a family contribute equally to the tRNA pool . It is often implicitly assumed that all tRNA copies contribute similarly to the cellular tRNA pool . However , comparison of the growth parameters of tRNA deletions from the same family revealed marked differences between seemingly identical family members . In particular , under rich medium , 21 out of the 32 deletions examined from multi-copy families showed growth yield differences spanning a broad range of at least 10% ( Figure 4A ) . Such differences were also detected in the growth rate parameter ( Supplemental figure S7A ) although they were less pronounced . We thus focus on the growth yield parameter in all further analysis . The phenomenon of differential contribution to fitness by different family members was further enhanced when we grew the deletion strains on more challenging conditions such as low glucose ( Figure 4B and Supplemental figure S7B ) . To further investigate the genetic interactions between differentially contributing tRNA copies within a given family , we focused on the tR ( UCU ) family . The tR ( UCU ) family contains 11 identical copies in the genome , 5 of which were represented in our library . In rich medium , two copies ( tR ( UCU ) E and tR ( UCU ) M2 ) showed appreciable reduction in growth yield ( termed Major copies ) , while deletions of the other three copies ( tR ( UCU ) M1 , tR ( UCU ) G1 and tR ( UCU ) K ) grew essentially as the wild-type ( termed Minor copies ) . Introducing a plasmid with the appropriate tRNA gene copy complemented the growth of all deleted copies ( Supplemental figure S7C ) . To further assert the separation between the Major and Minor copies , we examined various pair-wise deletion combinations of these members . All pairs that included at least one Major member exhibited growth impairment upon deletion , while pairs that consisted of only Minor copies demonstrated either a slight growth defect or none at all ( Figure 4C ) . Further analysis of genetic interactions of these family members with either the TRM9 gene , or with the above mentioned tR ( CCU ) gene that belongs to a different Arginine family , revealed a similar effect ( Figure 4C ) . These results indicate that the loss of different tR ( UCU ) genes in the same genetic background does not affect the phenotype equally , Major copies are more essential than Minor copies and as such are also more essential in providing compensation within the family . We next turned to examine whether the hierarchy of Major and Minor copies is preserved across various stress conditions ( Figure 4D ) . Examining essentiality in several conditions , we observed the same phenomenon in which Major copies demonstrated a stronger effect on growth compared to Minor copies in most stress conditions . We also noted that the Minor copies showed a diverse response ranging from slight growth improvement , wild-type level growth to observable growth impairment . A potential scenario may be one in which the Major copies always actively contribute to the pool , while the Minor copies might be recruited at times of need to maintain efficient translation . Thus , the loss of a Major copy could only be partially compensated by the remaining copies . Following these observations , we turned to examine possible genetic elements that might promote the phenomenon of differential contribution . Since all family members have identical sequence , we hypothesized that differential contribution should be due to differences in the vicinity of tRNA genes . To demonstrate this notion we performed a complementation assay , introducing different tRNAs from the UCU family , along with 200 bp of their flanking sequences , to the tR ( UCU ) M2 deletion strain . We observed different degrees of complementation . Given that different constructs differ only in the region flanking the tRNA gene , the variation in complementation capability can be attributed to the different sequences flanking the tRNA ( Supplemental figure S7D ) . The effect of sequences that flank tRNA genes on their transcription was reported in multiple studies [39]–[42] . In one such study Giuliodori et al . [42] preformed an analysis of conserved sequence elements upstream of S . cerevisiae tRNA genes . They identified four conserved sequence elements located at positions −53 ( T-rich ) , −42 ( TATA-like ) , −30 ( T-rich ) and −13 ( pol III TSS ) with respect to the first nucleotide of the mature tRNA . We used these results to examine the entire tRNA deletion library and checked whether tRNA deletions that exhibited or that did not exhibit altered phenotype in rich medium revealed enrichment for any particular motif ( Figure 4E ) . We found that deletions exhibiting phenotypes of growth impairment were significantly enriched for the presence of specific motifs . In particular , deleted strains that exhibited impairment in growth yield had an enrichment for the TATA-like motif at position −42 . In addition , the TSS motif at position −13 was enriched in deletion strains that exhibited impairment in both growth rate and yield . To reinforce these observations , we ran the MEME motif search algorithm [43] to screen the upstream sequences of tRNA deleted strains exhibiting impaired growth yield for enriched motifs ( see Materials and Methods ) . Two significant motifs were found that resemble those reported by Giuliodori et al . in both sequence and location ( Figure 4F ) . Together these results indicate that the contribution to the tRNA pool and cellular fitness of different copies of the same tRNA family are far from equal . We provide one possible explanation , which can account for the differential essentiality , implying that the sequences flanking tRNA genes play a role in determining their expression level . As mentioned above , screening the tRNA deletion library in the presence of the reducing agent Dithiothreitol ( DTT ) , a drug that exerts a proteotoxic stress in the cell , showed severe phenotypic defect in many deletion mutants ( Figure 2A , B ) . Yet , many of the strains that demonstrated growth reduction in other conditions were less sensitive than wild-type to this drug ( Figure 2C , D ) . These findings point towards a connection between tRNA functionality and the protein folding state in the cells . To further explore this connection , we turned to thoroughly characterize a selection of tRNA deletions in the presence of various proteotoxic agents . We chose two deletion mutants that exhibited either impaired or wild-type growth under DTT , namely ( tR ( UCU ) M2 and tH ( GUG ) G1 ) , both members of multi-copy families designated the MC group . In addition to the two viable single gene deletions ( tR ( CCU ) J and tL ( GAG ) G ) , the initiator methionine tiM ( CAU ) C also demonstrated improved growth; we thus designated these three strains the SC group . The various strains were treated with either DTT , Azetidine 2 carboxylic acid ( AZC ) - a toxic analog of proline [44] , or Tunicamycin- a drug used to induce the unfolded protein response ( UPR ) in the endoplasmic reticulum ( ER ) [45] . The growth of each strain was characterized under each proteotoxic agent applied at several concentrations . The strains in the MC group demonstrated either growth impairment or wild-type growth under all examined conditions . However , the deletions of single-copy tRNAs and to some extent the imitator methionine demonstrated reduced sensitivity to all three proteotoxic agents ( Figure 5A–C ) . The differences in relative growth for all the examined strains were apparent even at low concentrations and were consistent upon increase in the concentrations of these proteotoxic agents ( Figure 5A–C ) . The fact that the tRNA deletion strains from the SC group are resistant to proteotoxic agents led us to hypothesize that deleting these genes might inflict intrinsic and chronic misfolding stress , even at the absence of the drug . This stress results in the activation of relevant cellular response that protects cells from the aggravating effect of extrinsic proteotoxic stress . Such an effect is reminiscent of the cross protection effect observed between environmental stressors [46] , yet here it is manifested between a genetic perturbation and an environmental stress . To directly examine whether changes in the tRNA pool induce proteotoxic stress in these strains , we examined the state of the protein quality control machinery using the naturally unstructured human protein VHL as a proteotoxic stress reporter [47] . In this system , the VHL protein can be destined to one of two cellular localizations . If the cell experiences protein-folding stress , the heterologous protein VHL will aggregate in inclusions ( or puncta ) due to saturation of the protein quality control machinery . In contrast , under normal conditions , the quality control machinery is available to properly deal with this naturally unfolded heterologous protein , thus it remains soluble in the cytoplasm and no inclusions are formed . For each of the five deletion strains , we quantified the number of VHL inclusions ( puncta ) in populations of yeast cells . This analysis revealed that indeed the tRNA deletions in the SC group exhibited a significant increase in the number puncta containing cells relative to the wild-type ( Figure 5D and 5F ) , indicating saturation of the quality control machinery caused by intrinsic proteotoxic stress . The MC group did not exhibit inherent proteotoxic stress; their puncta containing cells count resembled that of the wild-type . The inherent chronic proteotoxic stress observed for the SC deletions might provide them with the capacity to respond better to an additional external proteotoxic stress . To further explore this possibility we examined the state of the protein quality control machinery upon extrinsic proteotoxic stress induced by treatment with AZC . Treating the wild-type cells with AZC resulted in a rapid accumulation of the VHL protein in stress foci , indicated by increase in the occurrence of multiple inclusions [48] . As anticipated , the behavior of the SC group demonstrated a significant increase in the presence of a single punctum upon AZC treatment , however the appearance of stress foci ( multi-puncta ) was significantly lower compared to the wild-type and to the MC group ( Figure 5E and 5G ) . As in the previous experiment , the deletions of the MC group responded in a similar manner to that of the wild-type , displaying increased number of stress foci . These results thus indicate that the deletion of some tRNA genes induced an inherent proteotoxic stress in the cell , demonstrating a physiological role of proper tRNA supply in protein folding by an undetermined mechanism . Such physiological response renders these cells relatively less sensitive , compared to other tRNA deletion strains and the wild-type , from the otherwise harmful effect of proteotoxic drugs . To determine whether changes in the tRNA pool result in a distinct molecular signature , we examined the same set of tRNA deletions ( SC and MC groups ) using mRNA microarrays . For each strain , we measured genome-wide changes in mRNA levels compared to the wild-type , under rich growth conditions . The expression changes we observed were modest and demonstrated a correlation between the essentiality of the tRNA gene and the extent of changes in mRNA expression upon its loss . Hierarchical clustering of the strains according to similarity in expression changes ( Fig . 6A and supplemental figure S8 ) , revealed that the strains could be divided into two groups recapitulating the division to the SC and MC groups . An example for this division can be found in the pronounced effect observed for the COS8 gene . This gene was extremely up-regulated ( about 16 fold ) in the SC group while unchanged in the MC group ( Figure 6B ) . These results suggest different molecular signatures for the two groups , which are also related to the proteotoxic stress response . To determine the responses and the underling molecular pathways that differentiate these two groups , we examined which KEGG pathways [49] , [50] differentiate between them . We used Gene Set Enrichment Analysis ( GSEA ) , a computational software which determines whether a defined set of genes shows statistically significant differences between two biological states [51] , [52] . This analysis revealed a somewhat opposite signature between the two groups ( Table 1 and supplemental figure S8 ) . Pathways which are responsive to proteotoxic stress such as the Proteasome ( FDR q-value<1E-5 ) and Protein processing in endoplasmic reticulum ( FDR q-value 2E-3 ) were significantly induced in the SC group relative to the MC group . While in the MC groups , translation-related pathways such as Ribosome biogenesis ( FDR q-value<1E-5 ) and Ribosome ( FDR q-value 1E-4 ) were significantly induced compared to the SC group . To further characterize these differences we focused on specific pathways . A more detailed examination of the expression changes observed for all the genes that constitute the proteasome complex revealed an up-regulation to various extents in response to deletion of tRNAs from the SC group . The MC group demonstrated no change and even a slight down-regulation of these genes ( Figure 6C ) , a trend which was further verified using RT-qPCR ( Figure 6D ) . These observations establish the notion that cells experience proteotoxic stress upon deletion of members of the SC group . A further indication of proteotoxic stress in these deletion strains is the up regulation of COS8 . The exact biological function of this gene is still unclear , it was however found to interacts with IRE1 , which is a hallmark regulator of the unfolding stress response [53] . An interesting distinction between the groups was also observed in the pathway consisting of the RNA polymerase machinery . Expression of genes that belong to this pathway were up-regulated only in the MC group ( Table 1 ) . Separating the RNA polymerase genes into modules corresponding to the different polymerases , revealed an interesting pattern . While the genes that encode RNA Pol II subunits did not change in any of the tRNA deletion strains ( Supplemental figure S9 ) , the genes encoding RNA Pol III machinery ( the polymerase responsible for tRNA gene transcription ) demonstrated up-regulation in the MC group and no change or even down regulation in the SC group ( Figure 6E ) . These results were further verified by RT-qPCR ( Figure 6F ) . Up-regulation of the pol III machinery for the MC group may suggest that in some MC deletion strains , the transcription of the remaining tRNA genes could increase , thus providing a possible molecular mechanism for backup compensation within families . Such response to deletions of tRNAs from the MC group could indicate the presence of a negative feedback loop , allowing the cell to respond to changes in the tRNA pool in the attempt to regain steady state levels .
In this study , we investigated the genetic architecture of the tRNA pool and its effect on cellular fitness using a comprehensive tRNA deletion library . We found extensive dispensability of many tRNA genes , especially under optimal growth conditions . Such lack of essentiality has been studied in protein-coding genes , and is often interpreted to reveal a role for partially redundant genes and pathways providing backup compensation for the deleted gene [33] , [34] , [54]–[56] . Similar design principles are displayed in the architecture of tRNA genes , which exhibited significant gene redundancy and compensation ( either partial or complete ) among family members . An additional reason for apparent lack of essentiality of genes is the limited set of examined environmental challenges , and it was indeed shown for protein-coding genes that challenging gene deletion libraries to less favorable conditions exposes more essentiality [57] , [58] . We showed that a similar situation holds for tRNA genes . We found condition-specific functional roles for tRNAs , demonstrating increased demand for certain tRNA genes under certain defined conditions . This implies that the compensation within tRNA families changes across conditions . Such changes in the essentiality of tRNA genes can imply that the tRNA pool is dynamic and changes across conditions to accommodate cellular needs , as was recently suggested [59] . Further , we have discovered interesting architecture within families , which questions the prior notion that all tRNA gene copies contribute equally to the pool . Previous work has shown that Pol III transcription machinery displays different occupancy levels at various copies of the same tRNAs in the genome [21] , [22] , [60] . However , the potential phenotypic consequences of such transcriptional differences have not been previously explored . We report that the flanking sequences around each tRNA gene contains motifs that are predictive of the deletion phenotypic consequences , potentially affecting pol III transcription machinery . We further speculate that some tRNA genes , i . e . the Major copies , might be active across all conditions and with only partial functional redundancy , thus their loss cannot be fully compensated . Minor copies on the other hand are either not transcribed or have a modest contribution to the tRNA pool , with complete functional redundancy by other copies , thus their loss can be fully compensated . Such architecture could provide the cell with means to respond in a dynamic manner to changes in the environment , by transcribing varying portions of the members of each tRNA family depending on demand . As such , differential contribution within tRNA families exposed an additional novel mean to regulate the tRNA pool and as a consequence to regulate the translation process . An interesting finding was that changes in the tRNA pool elicit molecular changes in the cells even when no severe phenotype is detected . Our results demonstrated two distinct molecular signatures which can be attributed to the family architecture and the severity of the changes in the pool . Upon deletion of the two viable single copy tRNAs , and also upon deletion of one of the initiator tRNA methionine copies , the cell exhibited a response reminiscent of a proteotoxic stress . We were able to identify such a stress in these mutant cells . Although the exact mechanisms by which changes in the tRNA pool induces proteotoxic stress remains to be determined , we hypothesize that the elimination or reduction in these tRNAs may lead to events of amino acid misincorporation , ribosome frame-shifting or stalled protein synthesis terminations . Such events would have a clear impact on the protein quality control machinery of the cell by titrating chaperons to deal with misfolded or misassembled proteins . Translation errors such as incorrect tRNA selection and incorrect tRNA aminoacylation have been shown to induce proteotoxic stress in yeast [61] , [62] . Given that cells exploit chaperon availability as a sensing mechanism to induce a stress response [63] , [64] , translation errors may lead to the onset such a response . On the other hand , deletions of tRNAs from multi-copy families results in milder effects on the tRNA pool due to the extensive redundancy or backup-compensation , and they indeed elicit a different cellular response from the one invoked upon deletion of single-member families . In the response to deletion of members from multi-gene families , the pol III transcription machinery seems to be up regulated . Such up-regulation would bring about induced transcription of tRNAs , this would act as a feedback mechanism to bring the tRNA pool closer to its normal state [65] . At least in one case ( Supplemental figure S5 ) our results suggest the existence of such responsive backup among tRNA genes from the same family . Yet , a clearer relationship between changes in the tRNA pool , pol III activation , and tRNA transcription is still lacking . Regardless of the actual mechanism that determines the exact cellular response to tRNA deletions , the fact that such a response wiring exists may be beneficial for maintaining cellular robustness upon environmental changes and mutations . This work provides for the first time a systemic tool to study the functional role of individual tRNA genes . Using this deletion library , we discovered a much more complex picture than was previously known . We anticipate that a high throughput mapping of all genetic interactions between pairs of tRNA genes ( as done for protein-coding genes ) [66] , [67] would reveal the full genetic network . In addition , it might reexamine and potentially refine the wobble interaction rules from a genetic , rather than the traditional biochemical/structural perspective . The design principles defined in this study , consisting of massive gene redundancy as well as differential contribution of gene copies may provide cellular plasticity and allow the tRNA pool to accommodate various growth conditions and developmental planes . Deciphering the effects of tRNA variations as is found in some diseases such as cancer [68] and Huntington [69] can provide possible routes for future treatment . We provide this novel set of minimalist genetic perturbations in the translation machinery as a resource to the yeast community towards further characterization of this highly complex process as well as additional cellular processes .
The complete tRNA pool of S . cerevisiae was obtained from the tRNA genomic database [70] , where 286 tRNA genes are annotated . 13 tRNA genes are encoded by the mitochondrial genome and the remaining are nuclear-encoded . Here we focused on the nuclear-encoded tRNAs . Two tRNA genes that are annotated in this database as not determined , belong to the tS ( GCU ) family . Thus , the tS ( GCU ) family contains two additional members , tS ( GCU ) L and tS ( GCU ) D , both verified by PCR , bringing the total number of nuclear encoded tRNA genes to 275 . Deletion strains were constructed using a PCR-based gene deletion [71] , [72] , in the genetic background of the Y5565 strain ( MATα , can1Δ::MFA1pr-HIS3 , mfα1Δ::MFα1pr-LEU2 , lyp1Δ , ura3Δ0 , leu2Δ0 ) . The S . cerevisiae strain S288C reference genome sequence R57-1-1 downloaded from the Saccharomyces Genome Database was used for primer design . Each deletion construct contained 45 bp flanking or overlapping a tRNA sequence for specific recombination event , a unique barcode and the HPH antibiotics ‘cassette’ , conferring resistance to the antibiotic hygromycin B , [73] . PCR products were transformed into yeast cells and single colonies were verified by PCR . Three colonies from each strain were used to verify phenotypes in growth analysis . A wild-type strain in which the same antibiotic marker was integrated 200 bp upstream of the tL ( CAA ) L3 locus was created as a control and was used in all analyses as wild-type . A complete list of all plasmids , yeast strains and PCR fragments can be found in Supplemental text S1 and Supplemental table S5 . Strains were grown for two days at 30°C in YPD ( 1% yeast extract , 2% peptone , 2% glucose ) , diluted ( 1∶50 ) into the appropriate medium in U-bottom 96-well plates and grown at 30°C ( using TECAN Freedom EVO robot ) . The OD of the population in each plate was monitored every 30 minutes using a spectrophotometer at 600 nm ( INFINITE200-TECAN ) . Each plate contained a wild-type strain to which the growth parameters of the deletions strains were normalized . The OD reads served for growth analysis and extraction of growth parameters . At least 3 biological repeats and 36 technical repeats were performed for each strain in each condition . Complete description of analysis and normalization procedures are provided in the Supplemental text S1 . Library strains were screened in the following growth conditions: YPD , SCD ( 0 . 67% Bacto-yeast nitrogen base w/o amino acids 2% glucose supplemented with amono acids ) , YP supplemented with 0 . 025% glucose , YP supplemented with 1% galactose , YPD supplemented with 0 . 5 M NaCl , SCD supplemented with 1 . 5 mM DTT . Growth measurements were also performed on YPD supplemented with increasing concentrations of the proteotoxic agents DTT , AZC and Tunicamycin . A sequence motif analysis was performed using the MEME online software [43] . The motif search was done on the upstream sequence of tRNA genes which exhibited a yield impairment phenotype in rich medium upon deletion ( 42 genes ) versus the upstream sequence of tRNA genes which exhibited a phenotype in no more than two out of the six conditions ( 99 genes ) . To apply location constrains on the motifs , the MEME analysis was done in windows of size 9 bp , looking for motifs of 4–8 bp in length . Wild-type and tRNA deletion strains harboring the pGAL-VHL-mCherry ( CHFP ) fusion were grown overnight on SCD+2% raffinose , diluted into SCD+2% galactose and grown at 30°C for 6 hours . Cells were visualized using an Olympus IX71 microscope controlled by Delta Vision SoftWoRx 3 . 5 . 1 software , with ×60 oil lens . Images were captured by a Photometrics Coolsnap HQ camera with excitation at 555/28 nm and emission at 617/73 nm ( mCherry ) . Images were scored using the ImageJan Image Processing and Analysis software . The percentage of cells harboring VHL-CHFP foci was determined by counting at least 500 cells for each strain in three biological repetitions . Protein un-folding stress was induced with AZC at a concentration of 2 . 5 mM AZC ( Sigma ) following induction with galactose . Cultures were grown in YPD medium at 30°C to a cell concentration of 1 . 5*107 cells/ml . Cells were then harvested , frozen in liquid nitrogen , and RNA was extracted using MasterPure™ ( EPICENTER Biotechnologies ) . The quality of the RNA was assessed using the BIOANALYZER 2100 platform ( AGILENT ) ; samples were then processed and hybridized to Affymetrix yeast 2 . 0 microarrays using the Affymetrix GeneChip system according to manufacturer's instructions . The background adjustment was done using the Robust Multi-array Average ( RMA ) procedure followed by quintile normalization . For each strain , the fold change in expression for all genes was calculated by comparing the wild-type measurement in the same batch and averaged over two biological repeats . The cluster tree is based on the correlation between the mRNA fold change of the different strains . For the clustering we used the top 50% of the sorted genes based by the gene variance across the strains . The data from this study have been submitted to the NCBI Gene Expression Omnibus ( GEO ) under accession number GSE47050 . A list of the measured fold changes for all genes in each strain can be found in Supplemental table S4 . Cultures were grown in YPD medium at 30°C to a cell concentration of 1*107 cells/ml . RNA was extracted using MasterPure™ ( EPICENTER Biotechnologies ) , and used as a template for quantitative RT–PCR using light cycler 480 SYBR I master ( Roche ) ( LightCycler 480 system ) according to the manufacture instructions . A list of the primers can be found in Supplemental table S6 . | Transfer RNAs are an important component of the translation machinery . Despite extensive biochemical investigations , a systems-level investigation of tRNAs' functional roles in physiology , and genetic interactions among them , is lacking . We created a comprehensive tRNA deletion library in yeast and assessed the essentiality of each tRNA in multiple conditions . The majority of tRNA deletions showed no appreciable fitness defect when such strains were grown on rich medium . More challenging environmental conditions , however , revealed a richer set of specific-tRNA phenotypic defects . Co-deletion of tRNA combinations revealed that tRNAs with essential function can be compensated by members of the same or different anti-codon families . We often saw that identical tRNA gene copies contribute deferentially to fitness , suggesting that the genomic context of each gene can affect functionality . Genome-wide expression changes in response to tRNA deletions revealed two different responses . When a deleted tRNA belongs to a family which contains multiple genes with the same anti-codon , the affected cells responded by up-regulating the translation machinery , but upon deletion of singleton tRNAs , the cellular response resembled that of proteotoxic stress . Our tRNA deletion library is a unique resource that paves the way towards fully characterizing the tRNA pool and their role in cell physiology . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"systems",
"biology",
"rna",
"cellular",
"stress",
"responses",
"model",
"organisms",
"molecular",
"cell",
"biology",
"rna",
"processing",
"nucleic",
"acids",
"gene",
"regulation",
"genetics",
"gene",
"expression",
"molecular",
"genetics",
"population",
"genetics",
"biology",
"yeast",
"and",
"fungal",
"models",
"saccharomyces",
"cerevisiae",
"gene",
"networks",
"gene",
"pool"
] | 2014 | A Comprehensive tRNA Deletion Library Unravels the Genetic Architecture of the tRNA Pool |
In humans and rodents , stress promotes habit-based behaviors that can interfere with action–outcome decision-making . Further , developmental stressor exposure confers long-term habit biases across rodent–primate species . Despite these homologies , mechanisms remain unclear . We first report that exposure to the primary glucocorticoid corticosterone ( CORT ) in adolescent mice recapitulates multiple neurobehavioral consequences of stressor exposure , including long-lasting biases towards habit-based responding in a food-reinforced operant conditioning task . In both adolescents and adults , CORT also caused a shift in the balance between full-length tyrosine kinase receptor B ( trkB ) and a truncated form of this neurotrophin receptor , favoring the inactive form throughout multiple corticolimbic brain regions . In adolescents , phosphorylation of the trkB substrate extracellular signal-regulated kinase 42/44 ( ERK42/44 ) in the ventral hippocampus was also diminished , a long-term effect that persisted for at least 12 wk . Administration of the trkB agonist 7 , 8-dihydroxyflavone ( 7 , 8-DHF ) during adolescence at doses that stimulated ERK42/44 corrected long-lasting corticosterone-induced behavioral abnormalities . Meanwhile , viral-mediated overexpression of truncated trkB in the ventral hippocampus reduced local ERK42/44 phosphorylation and was sufficient to induce habit-based and depression-like behaviors . Together , our findings indicate that ventral hippocampal trkB is essential to goal-directed action selection , countering habit-based behavior otherwise facilitated by developmental stress hormone exposure . They also reveal an early-life sensitive period during which trkB–ERK42/44 tone determines long-term behavioral outcomes .
Goal-directed actions are defined as behaviors directed towards achieving a specific outcome . By contrast , habits are stimulus elicited and insensitive to action–outcome relationships . Individuals who experience early-life stress have an increased incidence of behaviors that can lead to addiction and obesity as adults , and Patterson et al . [1] provided evidence that these behaviors may result from an overreliance on outcome-insensitive habits . In rats , chronic stressor exposure similarly biases behavioral response strategies towards habits [2] , and the primary glucocorticoid corticosterone ( CORT ) is sufficient to induce habit biases in both rats and mice [3] . Exogenous glucocorticoids similarly enhance habit-based learning and memory in humans [4] . And like early-life stress [1] , prenatal stress in humans and maternal separation in neonatal rats also induce inflexible habit behavior [5 , 6] . Despite these convergences across species , how elevated glucocorticoids , particularly during specific developmental periods , cause long-term biases towards habit-based behavior remains unclear . To address this issue , we elevated CORT in mice during a timespan equivalent to early adolescence in humans , which induced habit biases in adulthood . We hypothesized that adolescent CORT exposure may have long-term behavioral consequences by impacting tyrosine kinase receptor B ( trkB ) , the high-affinity receptor for Brain-derived Neurotrophic Factor ( BDNF ) , in corticolimbic regions . We were motivated by evidence that corticohippocampal trkB levels increase during early postnatal development and adolescence [7] and are stress sensitive [8] . Our investigations focused on the ventral hippocampus ( vHC ) , medial prefrontal cortex ( mPFC ) , striatum , and amygdala , brain regions implicated in action–outcome decision-making—that is , in selecting behaviors based on expected consequences , rather than familiar habit-based strategies [9–13] . Specifically , inactivation of the mPFC or connected regions of the striatum causes failures in selecting actions based on their outcomes or on outcome value [9–13] . Similar behavioral impairments follow amygdala inactivation ( in particular , inactivation of the basolateral compartment [13] ) . Meanwhile , vHC inactivation disrupts goal encoding in the mPFC [14 , 15] , and the vHC appears to route contextual and task-relevant information to the mPFC and amygdala in order to influence reward-related decision-making and response selection [16 , 17] . Habit biases that occur due to disruptions in corticolimbic networks may be associated with depression-like behavior . Depressive rumination in humans can be habit-like—stimulus elicited , resistant to change , and precipitated by stressor exposure [18] . A sense of helplessness in depression has also been conceptualized as a habit-based weakness in awareness of action–outcome contingency [19] . To investigate habit biases following adolescent CORT exposure , we used an instrumental contingency degradation procedure in which a familiar behavior was uncoupled from reward . To investigate depression-like behavior , we turned to the progressive ratio task , a classical assay of reward-related motivation . Using these separable strategies , our findings suggest that vHC trkB is necessary for goal-directed action ( occluding habits ) and that compromised trkB signaling induces habit-based and depression-like behavior . They also reveal a sensitive period during which enhancing ERK42/44 activity during adolescence can interfere with CORT-induced habit-based and depression-like behavior later in life .
In both humans and rodents , glucocorticoid exposure can induce biases towards habit-based behavior , at the expense of goal-directed action [3 , 4] . Furthermore , stressor exposure during early developmental periods appears to confer long-term habit biases across rodent–primate species [1 , 5 , 6] . Despite these homologies , mechanisms remain unclear . To address these issues , we exposed mice to CORT in the drinking water from postnatal day ( P ) 31–42 , equivalent to early adolescence in humans [20] . The timing of experimental events is provided in Table 1 , and timelines are also provided in the figures . We first confirmed that exogenous CORT exposure elevated blood serum CORT late in the active period ( that is , nighttime ) when mice had been active and ingesting CORT for several h . Notably , levels did not differ between CORT-exposed and control groups during the early active period when mice were just waking ( interaction F ( 1 , 27 ) = 22 . 3 , p < 0 . 001 ) ( Fig 1a ) . This pattern indicates that exogenous CORT disrupts typical diurnal changes in blood serum CORT levels . Correspondingly , adrenal and thymus glands atrophied during the CORT exposure period , as expected ( t8 = 5 . 7 , p < 0 . 001; t8 = 4 . 24 , p = 0 . 003 ) ( Fig 1b ) . Also as expected , gland weights recovered when exogenous CORT was removed ( t10 = −0 . 25 , p = 0 . 8; t10 = 0 . 099 , p = 0 . 9 ) ( Fig 1b ) . Despite this recovery , break point ratios in a progressive ratio test , an assay of reward-related motivation , were reduced in mice with a history of CORT exposure ( t9 = 2 . 39 , p = 0 . 04 ) ( Fig 1c ) . This pattern is consistent with amotivation in depression and provides evidence of long-term behavioral consequences of adolescent CORT exposure . We next compared our oral CORT exposure procedure to daily forced swim stress . Forced swimming increased blood serum CORT as expected , although this effect appeared to habituate with repeated exposure ( F ( 2 , 20 ) = 4 . 0 , p = 0 . 04 ) ( Fig 1d ) . Nonetheless , adolescent stressor exposure decreased progressive ratio break points , as with the oral CORT procedure ( t13 = 3 . 56 , p = 0 . 003 ) ( Fig 1e ) . Another well-characterized consequence of repeated stressor exposure is the elimination of dendritic spines on pyramidal mPFC neurons [21] . Thus , as another experiment validating our CORT exposure method , we enumerated dendritic spines on excitatory deep-layer pyramidal neurons in the mPFC using thy1-yellow fluorescent protein ( YFP ) –expressing transgenic mice . Spines were eliminated in the anterior-most sections ( interaction F ( 1 , 73 ) = 4 . 9 , p = 0 . 03 ) ( Fig 1f ) . The effect size ( Cohen’s d ) was 0 . 92 , signaling that approximately 80% of dendrites in CORT-exposed mice had fewer dendritic spines than the control mean . Dendritic spines were also reconstructed in 3D , revealing increased volume following CORT ( main effect F ( 1 , 72 ) = 6 . 2 , p = 0 . 02 ) ( Fig 1g ) . This phenomenon could not be accounted for by an increase in the head size ( Fs < 1 ) ( Fig 1h ) , suggesting that CORT induced dysmorphic spines with aberrantly large necks . This pattern was detected even several weeks following CORT exposure ( S1 Fig ) . Lastly , we exposed adult mice with a history of adolescent CORT treatment to the forced swim test . In this test , attempting to escape has been termed “active coping , ” while immobility has been termed “passive coping” [22] . Prior CORT exposure did not impact baseline immobility scores; however , an acute stressor challenge prompted an active coping style in control mice , reducing immobility . CORT-exposed mice were , by comparison , more immobile , favoring a passive response ( CORT × stress F ( 1 , 21 ) = 6 . 3 , p = 0 . 02 ) ( Fig 1i ) . Thus , exposure to exogenous CORT in adolescence modified stressor reactivity in adulthood . Having characterized our model , we next determined whether the same subchronic CORT exposure procedure would impact habit biases . We first trained control and CORT-exposed mice to nose poke 2 separate recesses for food reinforcers . We detected no side biases nor group differences in instrumental response acquisition rates ( F ( 1 , 15 ) = 2 . 7 , p = 0 . 12; interaction F ( 9 , 135 ) = 1 . 6 , p = 0 . 12 ) ( Fig 2a ) . We next decreased the likelihood that 1 response would be reinforced . A “goal-directed” response strategy is to then preferentially engage the remaining behavior , which remains likely to be reinforced , while habit-based responding is insensitive to action–outcome contingency [23] . In this case , mice engage both responses ( both “non-degraded” and “degraded” ) equivalently . Following an initial test , both groups inhibited the response that was unlikely to be reinforced in a goal-directed ( nonhabitual ) fashion ( main effect F ( 1 , 15 ) = 11 . 6 , p = 0 . 004; no interaction ) ( test 1 , Fig 2b ) . Response rates were also lower overall in the CORT-exposed mice , consistent with diminished break point ratios in the progressive ratio test ( Fig 1c ) . With additional training using random interval ( RI ) schedules of reinforcement that can bias responding towards habits , mice with a history of CORT exposure indeed assumed habit-based strategies , failing to differentiate between the behaviors that were more ( or less ) likely to be reinforced . Meanwhile , control mice differentiated between the responses , retaining goal-oriented response strategies ( interaction F ( 1 , 13 ) = 6 . 0 , p = 0 . 03 ) ( test 2 , Fig 2b ) . Thus , subchronic CORT exposure in adolescence causes a bias towards habit formation in adulthood . Notably , we discovered the same patterns when we tested female , rather than male , mice ( S2 Fig ) . In separate mice , test 2 occurred in a distinct context . In this case , both groups generated the response that was likely to be reinforced ( main effect F ( 1 , 12 ) = 16 . 7 , p = 0 . 001; no interaction ) ( Fig 2c ) . Thus , adolescent CORT-induced habits were context dependent . Next , we assessed whether subchronic CORT exposure similarly impacted adult mice . CORT exposure decreased body weights as expected ( Table 2 ) . While instrumental response rates during training were generally lower than in our younger cohorts , control versus CORT-exposed groups did not differ ( Fs < 1 ) ( Fig 2d ) . And unlike with subchronic CORT exposure in adolescence , both groups consistently generated the response most likely to be reinforced in a goal-directed fashion ( main effect test 1: F ( 1 , 12 ) = 8 . 4 , p = 0 . 01; main effect test 2: F ( 1 , 12 ) = 36 , p < 0 . 001; no interactions ) ( Fig 2e ) . Thus , adolescents were more vulnerable to the habit-inducing influence of elevated glucocorticoids . Insensitivity to instrumental contingencies is commonly associated with insensitivity to reinforcer value . However , when we tested the same mice for sensitivity to reinforcer devaluation , we found no impairment in response inhibition following ad libitum access to the reinforcer pellets prior to test ( “devalued” condition ) , relative to prefeeding with regular chow ( “non-devalued” condition ) ( main effect F ( 1 , 26 ) = 28 , p < 0 . 001; no interaction ) ( Fig 2f ) . Thus , subchronic CORT exposure during adolescence , but not adulthood , induced failures in selecting actions based on their consequences . This failure was context dependent , while value-based action selection was intact . In a final experiment , we exposed adolescent mice to CORT , then trained them to respond for food reinforcers following a prolonged , 4-wk washout period . This procedure doubled the “recovery” period following CORT and matched the age of testing in our adult CORT-exposed population . Mice acquired the nose poke responses without group differences ( F ( 1 , 21 ) = 2 . 7 , p = 0 . 12; interaction Fs < 1 . 2 ) ( Fig 2g ) . Mice were sensitive to action–outcome contingency degradation at test 1 , as in all other groups ( main effect F ( 1 , 21 ) = 5 . 8 , p = 0 . 03; no interaction ) ( Fig 2h ) . In a second test , CORT-exposed mice again preferentially generated the response most likely to be reinforced; however , this preference decayed ( group × response × time interaction F ( 1 , 20 ) = 4 . 2 , p = 0 . 05 ) ( Fig 2h ) . Thus , adolescent CORT-exposed mice recovered some function with a prolonged washout period ( as opposed to our shorter washout period tested above ) , but habitual response biases remained detectable nearly 1 mo following exposure . The transition from goal-directed to habit-based modes of response has been characterized as a decline in behavioral control by specific prefrontal cortex ( PFC ) -limbic structures ( e . g . , Fig 3a ) , in favor of sensorimotor circuits [11–13 , 23–25] . Within the hippocampus , the ventral compartment provides the primary inputs to the PFC [26] . For these reasons , we assessed levels of the stress-sensitive neurotrophin receptor trkB and phosphorylation of its substrate ERK42/44 in a mPFC–vHC–amygdala–striatal network . Results are reported in Table 3 . Key findings are also displayed graphically . Specifically , adolescent CORT exposure decreased the ratio of full-length trkB/trkB . t1 in the mPFC ( t12 = 4 . 3 , p < 0 . 001 ) ( Fig 3b ) . This pattern was not anatomically selective—in both the vHC and amygdala , developmental CORT also decreased trkB/trkB . t1 ratios ( main effect of CORT F ( 1 , 23 ) = 10 . 7 , p = 0 . 003 ) ( Fig 3c , top ) , as was also observed in the ventral striatum ( t10 = 3 . 9 , p = 0 . 003 ) ( Table 3 ) . CORT elevated overall levels of trkB . t1 in the vHC ( interaction F ( 1 , 21 ) = 5 . 8 , p = 0 . 025 ) ( Fig 3c , bottom ) and ventral striatum ( t10 = −2 . 3 , p = 0 . 046 ) ( Table 3 ) , but only in the vHC were levels of phosphorylated ( active ) ERK42/44 also decreased by CORT ( Fig 3d; Table 3 ) . Of note , phosphorylated ERK ( p-ERK ) levels were also generally higher in the vHC than the amygdala ( p-ERK42 F ( 1 , 18 ) = 16 . 4 , p < 0 . 001; p-ERK44 F ( 1 , 18 ) = 8 . 5 , p = 0 . 009 ) ( Fig 3d ) . To determine whether this effect was long lasting , we immunoblotted for p-ERK42/44 in the vHC 12 wk following adolescent CORT exposure , again revealing lower phosphorylated levels of both ERK isoforms ( p-ERK42: t18 = 2 . 7 , p = 0 . 01; p-ERK44: t18 = 3 , p = 0 . 008 ) ( Fig 3e ) . We next tested adult mice exposed to CORT . CORT again decreased amygdalo–hippocampal ratios of full-length trkB/trkB . t1 ( main effect F ( 1 , 18 ) = 9 . 4 , p = 0 . 007 ) and also elevated trkB . t1 ( main effect F ( 1 , 20 ) = 21 . 2 , p < 0 . 001 ) ( Fig 3f ) ; however , we identified no effect of CORT on p-ERK42/44 ( main effects and interactions p > 0 . 18 ) ( Fig 3g ) . Thus , a unique consequence of subchronic CORT exposure in adolescence appears to be the elevation of trkB . t1 and concurrent suppression of p-ERK42/44 within the vHC . Next , we attempted to block adolescent CORT-induced habits with the trkB agonist 7 , 8-dihydroxyflavone ( 7 , 8-DHF ) . We first aimed to identify a dose range that stimulated p-ERK42/44 in the vHC . Three and 10 mg/kg , intraperitoneal ( i . p . ) , induced ERK42 phosphorylation ( F ( 2 , 38 ) = 3 . 9 , p = 0 . 03 ) ( Fig 4a ) , though p-ERK44 levels were not affected ( F ( 2 , 38 ) = 1 . 7 , p = 0 . 2 ) ( Fig 4b ) . 7 , 8-DHF also blocked the chronic p-ERK42 deficit due to CORT exposure ( Table 4 ) while having no long-term consequences for adrenal and thymus gland weights ( S1 Table ) , and the 10 mg/kg dose increased levels of the postsynaptic marker , postsynaptic density 95 ( PSD95 ) ( F ( 2 , 24 ) = 4 . 8 , p = 0 . 02 ) ( Fig 4c ) . Next , separate mice were exposed to CORT during early adolescence , and half were treated with 7 , 8-DHF ( 3 mg/kg ) from P39–47 , overlapping with the end of the CORT exposure period and extending into late adolescence [20] ( Fig 4d ) . As adults , mice acquired the nose poke responses with no group differences ( main effect CORT and 7 , 8-DHF Fs < 1; CORT × 7 , 8-DHF interaction F ( 1 , 19 ) = 3 . 8 , p = 0 . 065; all other interactions Fs ≤ 1 . 1 ) ( Fig 4e ) . Following an initial instrumental contingency degradation test , all groups preferentially generated the response most likely to be reinforced in a goal-directed fashion as in our experiments described above ( main effect F ( 1 , 20 ) = 24 . 6 , p < 0 . 001; no interaction ) ( test 1 , not shown ) . With more training , CORT-exposed mice developed habit-based behavior , also as expected , but critically , 7 , 8-DHF blocked CORT-induced habit biases ( CORT × response interaction F ( 1 , 19 ) = 7 . 2 , p = 0 . 015; 7 , 8-DHF × response interaction F ( 1 , 19 ) = 5 . 3 , p = 0 . 03 ) ( test 2 , Fig 4f ) . This effect of 7 , 8-DHF was also detectable in female mice ( S2 Fig ) . We next tested these mice in the forced swim test . As in Fig 1 , prior CORT did not impact immobility in the absence of stressor exposure in adulthood . However , a history of 7 , 8-DHF treatment reduced time spent immobile , an antidepressant-like effect that was notably detectable multiple weeks following treatment ( main effect of 7 , 8-DHF F ( 1 , 20 ) = 8 . 3 , p = 0 . 008; no interaction ) ( Fig 4g ) . The reduction in time spent immobile could not obviously be attributable to general hyperactivity following 7 , 8-DHF treatment ( S2 Table ) . We also quantified responding on a progressive ratio schedule of reinforcement . 7 , 8-DHF dose-dependently blocked CORT-induced deficits in break point ratios , and this blockade was detectable when either total responses or break points were compared ( interactions F ( 2 , 56 ) = 3 . 9 , p = 0 . 03; F ( 2 , 54 ) = 3 . 4 , p = 0 . 04 ) . Break points are shown ( Fig 4h ) . To summarize , adolescent CORT exposure increases levels of trkB . t1 and decreases p-ERK42/44 in the vHC and also induces biases towards habit-based behaviors . These biases are blocked by the putative trkB agonist 7 , 8-DHF at doses that increase p-ERK42 in vHC . To determine whether selectively elevating trkB . t1 and decreasing p-ERK42/44 in the vHC is sufficient to recapitulate the behavioral effects of CORT exposure , we overexpressed Trkb . t1 in the vHC and a major projection target , the central nucleus of the amygdala ( CeA ) ( Fig 5a ) . vHC-targeted infusions were mostly restricted to the ventral Cornu amonis ( CA ) 1 region , with some spread into the intermediate hippocampus ( Fig 5a; see also S3 Fig ) . Amygdala-targeted infusions were largely contained within the CeA as intended ( Fig 5a; see also S3 Fig ) . Seven mice were excluded due to mistargeted infusions infecting white matter tracts , and control green fluorescent protein ( GFP ) -expressing mice did not differ and were combined . At the infusion site , p-ERK42/44 levels were reduced by 15% in Trkb . t1-expressing mice relative to mice expressing GFP ( t13 = 2 . 6 , p = 0 . 02 ) ( Fig 5b ) , mimicking the long-term consequences of adolescent CORT exposure ( compare to Fig 3d ) . Mice acquired the food-reinforced instrumental responses without group differences ( Fs ≤ 1 ) ( Fig 5c ) . GFP-expressing control mice were sensitive to instrumental contingency degradation , preferentially engaging the response most likely to be reinforced . By contrast , Trkb . t1 overexpression induced inflexible habits , indicated by a failure to respond in a selective fashion following instrumental contingency degradation ( interaction F ( 2 , 34 ) = 5 . 6 , p = 0 . 008 ) ( Fig 5d ) . Thus , Trkb . t1 overexpression recapitulated the long-term effects of adolescent CORT exposure . Next , we expanded these studies , deviating from our protocol used thus far , to assess whether Trkb . t1 overexpression in the vHC also interfered with the ability to “break” habits . We accordingly trained separate mice using an RI schedule of reinforcement . Instrumental response acquisition curves are segregated according to whether the action–outcome contingency associated with each response would ultimately remain intact or be “degraded , ” highlighting equivalent response rates throughout ( Fs < 1 ) ( Fig 5e ) . We modified our typical instrumental contingency degradation procedure to determine whether responses could be inhibited once habits formed . Specifically , we exposed mice to alternating training sessions in which 1 response was reinforced or the contingency between the other response and its outcome was degraded . Initially , mice responded equivalently during these 2 types of training sessions , exhibiting habit-based behavior . With repeated testing , control mice were ultimately able to inhibit the response that was unlikely to be reinforced . By contrast , response strategies in Trkb . t1-overexpressing mice were unchanged and habit-based ( group by contingency interaction F ( 1 , 32 ) = 6 . 6 , p = 0 . 02 ) ( Fig 5f ) . Thus , Trkb . t1 overexpression caused significant behavioral inflexibility . Finally , we also confirmed that vHC Trkb . t1 overexpression decreased break point ratios in a progressive ratio test ( t16 = 2 . 8 , p = 0 . 01 ) ( Fig 5g ) , as with adolescent CORT exposure . This finding is consistent with evidence that vHC-targeted knockdown of the trkB ligand BDNF also induces depression-like behavior [28] .
Adolescent CORT exposure also induced a long-term bias towards habit-based response strategies . Specifically , mice were initially able to select actions ( left/right nose poke ) based on their consequences ( food ) , but with repetition , these behaviors assumed habitual qualities such that they were insensitive to response–outcome contingency . Subchronic CORT exposure did not impact adult mice , indicating that adolescents are more vulnerable to developing CORT-induced habits . Moreover , when we doubled the “recovery” period duration following adolescent CORT exposure , all mice could initially select actions based on their consequences , but response preferences faded over time in CORT-exposed mice . We interpret this as uncertainty in response selection , resulting in a deferral to familiar , habit-based behaviors that are insensitive to response–outcome associations . Behavioral insensitivity to response–outcome contingency is often associated with insensitivity to reinforcer value [23] . This was not the case here , however , in that all mice reduced response rates following prefeeding with the reinforcer pellets , which decreases their value . Why might this be ? One possibility is that subchronic CORT exposure particularly impacted hippocampal function . Lesions of the entorhinal cortex , a primary input to the hippocampus , reduce sensitivity to response–outcome contingencies but not reinforcer value , as with CORT here [37 , 38] . This may be because organisms form an association between the context and response–outcome contingency during training . When that contingency is later violated , the hippocampus detects the discrepancy between “the context where I typically work for reward” and noncontingent pellet delivery and facilitates response inhibition . This model predicts that response strategies should be intact if instrumental contingency degradation occurs in a contextually distinct environment relative to the training environment , which was indeed the case here . In contrast , behavioral sensitivity to reinforcer devaluation is context-independent because it relies on an animal’s ability to prospectively calculate reinforcer value . Accordingly , it is unaffected by entorhinal cortex lesions [37 , 38] , while lesions/inactivation of other structures , such as the dorsal mPFC , basolateral amygdala , and dorsomedial striatum impair sensitivity to both response–outcome contingency and reinforcer value [23] . Based on these findings , we next quantified levels of the stress-sensitive neurotrophin receptor trkB in the vHC and other regions that counter habit-based behavior ( dorsal mPFC , amygdala , dorsomedial striatum , and ventral striatum; see [11–13 , 23] ) . CORT caused widespread modifications in the ratio of full-length:truncated isoforms , favoring the inactive isoform . ( Indeed , of the brain regions tested , only the dorsomedial striatum was spared . ) This is significant because trkB . t1 dimerization with full-length trkB reduces the receptor’s ability to stimulate ERK42/44 , PI3-kinase , and PLCγ . trkB . t1 is also linked to neurodegeneration [39] and excitotoxicity [40] . Additionally , increasing trkB . t1 decreases cell-surface levels of full-length trkB , further reducing opportunities for trkB-mediated signaling [41] . TrkB . t1 levels remained particularly robust in the vHC even following the CORT exposure period . These findings are in general agreement with prior investigations using social [42] ( though not physical [43] ) stress . Also , transgenic mice overexpressing Trkb . t1 are insensitive to classical antidepressants [44] , further motivating us to investigate whether direct trkB stimulation could have antidepressant-like properties following CORT . Indeed , 7 , 8-DHF given during adolescence ( P39–47 ) corrected CORT-induced habit behavior and amotivation and increased mobility in the forced swim test , an antidepressant-like effect . 7 , 8-DHF derivatives also have antidepressant-like consequences in adult mice [45 , 46] and restore the reinforcing properties of cocaine following stress [47] , but ours is the first evidence , to our knowledge , of antidepressant-like efficacy in adolescents . And importantly , effects were detectable well beyond the treatment period . How might adolescent 7 , 8-DHF treatment confer long-term benefits ? Enhancing trkB-mediated signaling could correct the suppressive effects of stressor or CORT exposure on neurogenesis in the vHC or restore BDNF–trkB interactions in the hippocampus , both of which would be associated with antidepressant-like efficacy [29 , 48 , 49] . Also of note , vHC innervation of the anterior mPFC develops during adolescence [50] . We found that adolescent CORT exposure caused dendritic spine elimination in the anterior mPFC . Additionally , remaining spines were irregularly enlarged , as also occurs following Trkb ablation [51] . Future studies could determine whether these modifications result from aberrant vHC input and whether 7 , 8-DHF corrects abnormalities . Finally , while we find a robust p-ERK42/44 response to repeated 7 , 8-DHF injections , the ability of acute application to stimulate trkB , ERK42/44 , and other factors has been questioned [52] . Identifying off-target actions of 7 , 8-DHF could reveal novel mechanisms by which acute application benefits animal models of depression ( e . g . , [46] ) and Alzheimer disease ( discussed in [52] ) . Despite broad-spread changes in trkB:trkB . t1 ratios following subchronic CORT exposure , p-ERK42/44 was detectably reduced only in the vHC of adolescent CORT-exposed mice . These findings led us to overexpress Trkb . t1 in the vHC , reducing local p-ERK42/44 and causing habit-based behavior . Trkb . t1 overexpression impairs both long-term potentiation and long-term depression ( [53 , 54] , but see [55] ) and reduces BDNF [56] , which could potentially account for a transition away from hippocampal-dependent action selection to habit-based behaviors , which are instead associated with dorsolateral striatal and cortical sensorimotor systems [11–13 , 23 , 25] . The vHC innervates the CeA [57] and likely provides BDNF , given that the CeA expresses little Bdnf mRNA but abundant BDNF from non-cortical sources , and among the highest levels of amygdalar trkB [58] . Axonal BDNF transport can be trkB-dependent [59]; thus , we hypothesized that vHC Trkb . t1 overexpression may render the CeA BDNF-deficient and that Trkb . t1 overexpression in the CeA may similarly influence decision-making strategies . Indeed , Trkb . t1 overexpression in the posterior CeA caused a deferral to habit-based strategies . This finding provides novel evidence that the healthy posterior CeA is involved in selecting behaviors according to their consequences . This may occur via regulation of appetitive arousal , rather than encoding specific action–value information per se [60] . By contrast , the anterior ( and not posterior ) CeA is essential to habit-based behavior , potentially due to interactions with the dorsolateral striatum [61] . Importantly , trkB . t1 is expressed in neurons [40 , 62 , 63] and glia [64–66] . Lentiviruses , as used here , preferentially infect excitatory neurons , but moderate glial infection would be anticipated [67] . Future investigations could elucidate cell-type–specific effects of Trkb . t1 overexpression . Also , it is notable that we did not overexpress Trkb . t1 in the mPFC . We were motivated by evidence that selective reduction of its ligand BDNF in this region fails to induce habit-based behavior [3 , 68] . mPFC-selective Bdnf knockdown does cause depression-like amotivation , however , and targeted BDNF infusions provide partial recovery from CORT-induced amotivation [3] . Thus , systemic 7 , 8-DHF treatment here may have ameliorated CORT-induced depression-like amotivation by acting in multiple brain regions , not strictly the vHC . To summarize , subchronic CORT exposure in adolescence imbalances trkB/trkB . t1 throughout several brain regions and selectively decreases vHC p-ERK42/44 . A Trkb . t1 overexpression procedure that reduces p-ERK42/44 recapitulates CORT-induced behavioral abnormalities . Interestingly , the “pro-habit” effects of vHC Trkb . t1 overexpression were not age dependent , given that viral vector infusion at both P31 and P56 induced behavioral inflexibility . We argue that adolescents are not necessarily uniquely vulnerable to Trkb . t1 overexpression but are rather more vulnerable to a corticosteroid-induced triggering of neurobiological factors associated with depression-like and habit-based behaviors ( in this case , the concomitant elevation of trkB . t1 and reduction in p-ERK42/44 in the vHC ) .
Group-housed wildtype C57BL/6 mice ( Jackson Labs ) were used , except for dendritic spine imaging experiments , in which case mice were thy1-YFP–expressing ( C57BL/6 background ) [69] . Mice were provided a 12-h light cycle ( 0800 on ) and food and water ad libitum except during instrumental conditioning when body weights of all mice were reduced to 90%–93% of baseline to motivate food-reinforced responding . Mice were males unless otherwise explicitly noted . The timing of experimental events is provided in Table 1 , and timelines are also provided in the figures . Procedures were approved by the Emory University Institutional Animal Care and Use Committee , licenses 2000973 , 2002802 , and 4000010 , and the Guide for the Care and Use of Laboratory Animals in Research . In cases of euthanasia , mice were deeply anesthetized with isoflurane prior to rapid decapitation . CORT hemisuccinate ( 4-pregnen-11β 21-DIOL-3 20-DIONE 21-hemisuccinate; Steraloids ) was dissolved in tap water ( 25 μg/ml free base ) according to an established protocol [3 , 29 , 30] . Mice were given CORT in place of normal drinking water . Water bottles were weighed daily , and mice were weighed every other day ( Table 2 ) . Average doses ( mg/kg ) of CORT were calculated by normalizing daily consumption values per cage to the total body weight of the animals in the same cage . Every 3 d , water bottles were emptied and refilled with fresh water or newly-prepared CORT solution . Mice were exposed to CORT from P31–42 or 56–67 ( in 1 cohort , P68 due to experimental error ) , resulting in approximately 5–9 mg/kg/d . These periods correspond to early adolescence and early adulthood in rodents [20] . Mice were euthanized at the end of the CORT exposure period , or they experienced a 2- , 4- , or 12-wk washout period as indicated . To compare blood serum CORT levels between CORT-exposed versus stressor-exposed mice , naive mice were exposed to forced swim stress at P31 or daily from P31–42 . Mice were placed in a glass cylinder ( 24 cm × 15 . 5 cm diameter ) filled to 10 cm with 22–25°C water in a dimly lit room . After 6 min , mice were allowed to dry in a warm cage lined with paper towels before being returned to the home cage . Water was changed between mice . Control mice were handled but not exposed to swim stress . Groups were also housed separately . Mice were weighed every other day ( Table 2 ) . We collected trunk blood at P31 or P42 . Mice were briefly anaesthetized with isoflurane and then decapitated either early in the active , dark cycle ( 2000 h ) or late in the active cycle ( 0600 h ) . In the case of swim stress , mice were euthanized 30 min following swimming [70] . Blood was centrifuged in chilled Eppendorph tubes at 4°C for 30 min , and serum was extracted . CORT levels were analyzed in duplicate by ELISA ( Assay Designs ) in accordance with manufacturer’s instructions with the exception of the extraction step , which was excluded . Adrenal and thymus glands were extracted following euthanasia by midline dissection and weighed in pairs . A widely documented consequence of repeated stressor exposure is the elimination of dendritic spines in the mPFC . As part of our efforts to test the possibility that our CORT exposure procedure recapitulated aspects of stressor exposure , brains from YFP-expressing mice were collected at the end of CORT exposure at P42 and submerged in chilled 4% paraformaldehyde for 48 h , then transferred to 30% w/v sucrose . Brains were sectioned into 40 μm-thick sections at −15°C . Dendrites on deep-layer mPFC neurons , prelimbic/medial orbital compartments , were imaged using confocal microscopy and reconstructed in 3D using Imaris software . Methods are described elsewhere [71] , the only modification being that a Leica TSC SP8 microscope was used . Eight dendrites/mouse , 16–25 μm in length and located between Bregma +1 . 98–+1 . 70 , were imaged and reconstructed by a single blinded rater/experiment . In our adolescent population , dendritic spine densities were more variable than expected based on our prior investigations of prelimbic cortical neurons [3 , 7 , 35] , and unblinding revealed considerable variance based on rostrocaudal positioning . Most dendrites ( 73% ) were imaged at roughly Bregma +1 . 94 or +1 . 78 . Thus , we next compared dendritic spine densities and morphological metrics by 2-factor ( CORT × anatomical position ) analysis of variance ( ANOVA ) , total = 4–8 dendrites/mouse , considering each dendrite an independent sample . Values +/− 2 standard deviations from the mean were considered outliers and excluded , and the results of these comparisons are reported here . To investigate long-term consequences of adolescent CORT exposure , dendritic spines were imaged and classified from mice exposed to CORT during adolescence , then behaviorally tested in adulthood ( methods described immediately below ) . Mice were food-restricted and trained to nose poke for 20-mg grain-based food reinforcers ( Bio-Serv Precision Pellets ) using Med-Associates conditioning chambers equipped with 2 nose poke recesses and a food magazine . Training was initiated with a fixed ratio 1 ( FR1; also called “continuous reinforcement” ) schedule of reinforcement; 30 reinforcers were available for responding on each aperture ( 60 reinforcers/session ) . Sessions ended when mice acquired all reinforcers or at 70 min . 5–7 training sessions were conducted ( 1/d ) . Unless specified , response acquisition curves represent both responses/min; there were no response biases throughout . To assess decision-making strategies , a modified version of classical instrumental contingency degradation was used , as in our prior reports ( e . g . , [3 , 35 , 72] ) . In a 25-min “non-degraded” session , 1 nose poke recess was occluded , and responding on the other was reinforced using a variable ratio 2 schedule of reinforcement . In a 25-min “degraded” session , the opposite aperture was occluded , and responding on the available aperture produced no programmed consequences . Instead , reinforcers were delivered into the magazine at a rate matched to each animal’s reinforcement rate on the previous day ( that is , food pellets were delivered independently of the animal’s actions ) . Thus , responding on 1 aperture became significantly less likely to be reinforced than the other . The order of the sessions and which response–outcome contingency was “degraded” were counterbalanced . The following day , a 5-min probe test was conducted in extinction . Both apertures were available . Mice that are sensitive to instrumental contingencies preferentially generate the response that is likely to be reinforced , a goal-directed response strategy; meanwhile , mice that have developed habits are insensitive to instrumental contingency degradation and generate both familiar responses equally , habitually ( for further discussion of this task , see [23] ) . Following test 1 , responding was reinstated using an RI30-sec schedule of reinforcement for 4 d to promote the formation of stimulus–response habits [73] . 30 reinforcers were again available ( 60 reinforcers/session , 1 session/d ) . Sessions ended when mice acquired all reinforcers or at 70 min . Then , the 3-day contingency degradation and probe test protocol was repeated ( “test 2” ) . In a separate experiment , mice were trained to nose poke using FR1 and then RI30 schedules of reinforcement . Then , mice were tested in the contingency degradation procedure 3 times ( 1 session/d ) to quantify the development of response inhibition . To determine whether insensitivity to instrumental contingency degradation was context dependent , we utilized a “context shift” [72]: the “non-degraded” and “degraded” training sessions and probe test 2 occurred in unique chambers located in a separate room in the laboratory relative to training and test 1 . The chambers were contextually distinct ( containing a recessed lever and distinct odors ) and configured differently ( nose poke ports and house light located on different walls ) . Following instrumental contingency degradation , nose poking was reinstated using an RI30 schedule of reinforcement during 3 daily training sessions . Then , prefeeding devaluation was used to assess value-based response selection . Mice were placed individually in empty shoebox-style cages for a 1-hr habituation period . Then , mice were allowed access for 30 min to either standard chow or the food pellets used during instrumental conditioning . Immediately following this prefeeding , mice were placed in the conditioning chambers , and responding in a probe test conducted in extinction was measured for 10 min . This procedure was repeated the following day with the opposite food item . Prefeeding with the reinforcer pellets , but not standard chow , reduces response rates if mice select actions based on outcome value . Mice consumed more reinforcer pellets than chow during the prefeeding period; thus , we tested mice in a third condition in which the number of pellets available to each mouse was matched to the amount of chow that the group consumed during the prior prefeeding period . Subsequent response rates during the probe test did not differ ( that is , when the pellets were restricted or not ) , and those following restricted access—controlling the amount of food ingested—are shown . All intake data are provided in S3 Table . In separate mice tested in the instrumental contingency degradation procedure , nose poke responding on 1 recess was reinstated using an FR1 schedule for 2 50-min sessions ( 1/d ) . Then , responding on a progressive ratio schedule , in which the response requirement increased by 4 with each reinforcer delivery , was measured . Sessions ended after 180 min or when mice executed no responses for 5 min . The “break point ratio” refers to the highest number of responses:reinforcers generated . Following instrumental conditioning , mice were fed ad libitum . Within 1 wk , mice were placed in a glass cylinder ( 24 cm × 15 . 5 cm diameter ) filled to 10 cm with 25°C water , as previously used to detect an increase in immobility following CORT [29] . Ten-min sessions were videotaped under dim light , and time spent immobile , defined as only movements necessary to keep the head above water , was scored by a single blinded rater . In 1 experiment , an acute stressor ( 19 hr water deprivation ) preceded the forced swim test in half of the group; “unstressed” mice in this experiment were left undisturbed . It is important to note that while mice in this report had ad libitum food access at the time of forced swim testing , they had all experienced modest food restriction during instrumental conditioning experiments; this could conceivably influence mobility scores ( for review , [74] ) . Following forced swim testing , locomotor activity was monitored for 24 hr using a custom-built Med-Associates locomotor monitoring system equipped with 16 photocells . Locomotor activity was quantified in photobeam breaks across the 24-hr period , which were summed into 6-hr bins ( S2 Table ) . 7 , 8-DHF ( Sigma; dissolved in 17% dimethylsulfoxide ( DMSO ) and saline; 3-10mg/kg [46] ) was administered i . p . daily from P39-47 . This period overlapped with the end of the adolescent CORT exposure period and was determined based on pilot studies . Control mice received 17% DMSO and saline . Mice were euthanized at the end of the CORT exposure procedure; 12 wk following CORT ( and following instrumental contingency degradation testing ) ; or 30 min following the last of 8 daily injections of 7 , 8-DHF . Mice were briefly anaesthetized by isoflurane and euthanized by rapid decapitation . Brains were extracted and frozen at −80°C . Brains were sectioned into 1-mm sections using a chilled brain matrix , and the mPFC , vHC , amygdala , dorsomedial striatum , and ventral striatum were extracted using a 1-mm tissue core by a single experimenter . Tissues were homogenized by sonication in lysis buffer ( 200 μl: 137 mM NaCl , 20 mM tris-Hcl [pH = 8] , 1% igepal , 10% glycerol , 1:100 Phosphatase Inhibitor Cocktails 2 and 3 [Sigma] , 1:1 , 000 Protease Inhibitor Cocktail [Sigma] ) , and stored at −80°C . Protein concentrations were determined using a Bradford colorimetric assay ( Pierce ) . Equal amounts of protein were separated by SDS-PAGE on 7 . 5% gradient tris-glycine gels ( Bio-rad ) . Following transfer to PVDF membrane , blots were blocked with 5% nonfat milk for 1 hr . Membranes were incubated with primary antibodies at 4°C overnight and then in horseradish peroxidase secondary antibodies for 1 hr . Immunoreactivity was assessed using a chemiluminescence substrate ( Pierce ) and measured using a ChemiDoc MP Imaging System ( Bio-rad ) . Densitometry values were normalized to the control sample mean from the same membrane in order to control for fluorescence variance between gels . vHC and amygdala samples were loaded on the same gels to allow for comparisons within and between brain regions and tested at least twice . Primary antibodies were anti-trkB ( Rb , Cell Signaling , 4603s , lot 3; 1:375 ) , anti-ERK42/44 ( Rb , Cell Signaling , 9102s , lot 26; 1:2 , 000 ) , anti-p-ERK42/44 ( Ms , Cell Signaling , 9106s , lot 43; 1:1 , 000 ) , and anti-PSD95 ( Rb , Cell Signaling , 3450s , lot 2; 1:1 , 000 ) . In our initial comparisons of vHC–amygdala ERK42/44 , a loading control ( GAPDH; Ms , Sigma , G8795 , lot 044M4808V; 1:5 , 000 ) was additionally applied to further confirm equivalent loading . Naïve mice were infused with a lentivirus expressing a CMV promoter and truncated trkB receptor isoform , Trkb . t1 , with an HA tag ( titer = 5 . 8 × 108 iu/ml; virus described in [75 , 76] ) . Control mice were infused with lenti-GFP , also bearing a CMV promoter . Mice were anaesthetized with ketamine/xylazine and placed in a digitized stereotaxic frame ( Stoelting ) . The scalp was incised , skin retracted , bregma and lambda identified , the head leveled , and stereotaxic coordinates corresponding to the vHC or CeA were located ( −3 . 0 AP/−4 . 0 DV/±2 . 75 ML and −1 . 5 AP/−4 . 9 DV/±3 . 0 ML , respectively ) . Viral vectors were infused over 5 min , with 0 . 5 μl/side . Needles were left in place for 5 additional min prior to withdrawal and suturing . Three wk later , instrumental conditioning began . Following testing , fixed brain tissue was imaged for GFP or immunostained for HA as described [76] . Mice were infused at P31 or P56 at the same coordinates; timing did not impact behavioral outcomes . Tissue sections from mice expressing viral vectors were blocked in a PBS solution containing 2% normal goat serum , 1% bovine serum albumin ( BSA ) , and 0 . 3% Triton X-100 ( Sigma ) for 1 hr at room temperature . Sections were then incubated in a primary antibody solution containing 0 . 3% normal goat serum , 1% BSA , and 0 . 3% Triton X-100 at 4°C for 48 hr . p-ERK42/44 ( Rb , Cell Signaling , 9102s , lot 26; 1:400 ) served as the primary antibody . Sections were incubated in secondary antibody solution containing 0 . 5% normal goat serum and 0 . 3% Triton X-100 , with Alexa Fluor 633 ( 1:200; Life Technologies ) serving as the secondary antibody . Sections were imaged on a Nikon 4550s SMZ18 microscope with settings held constant . Integrated intensity ( normalized to the size of the sampling area ) was measured where HA or GFP staining was also detected . Sections were compared in 2 cohorts , and fluorescence values were normalized to the control mean from each respective cohort . We analyzed 1–10 sections from each mouse , with each animal contributing a single mean value to statistical analysis . Body weights , blood serum CORT , gland weights , response rates , response counts , break point ratios , and densitometry values were compared by 2-tailed ANOVA or t test using SPSS with p < 0 . 05 considered significant . Following interactions , post hoc comparisons utilized Tukey’s HSD tests , and results are indicated graphically . In mice bearing Trkb . t1-expressing viral vectors , exclusions due to mistargeted infusions and the combination of control groups resulted in considerably uneven sample sizes ( reported in the captions ) ; thus , we compared these groups using type III ANOVA . Statistical approaches to comparing dendritic spine densities and morphologies are outlined in the corresponding section above . Values lying >2 standard deviations above the means were considered outliers and excluded . | Stress can impair the ability of individuals to select actions based on their consequences , promoting instead the acquisition of habits—automatic behaviors that are insensitive to action–outcome relationships . Stress in adolescence can have particularly long-lasting consequences , affecting behavior and mood regulation even in adulthood . Here , we show that adolescent mice exposed to a stress hormone developed biases towards habits and amotivation ( a symptom of depression ) in adulthood . We investigated the role of the stress-sensitive tyrosine kinase receptor B ( trkB ) receptor , a protein that regulates synaptic strength and plasticity in the mammalian nervous system . This receptor was modified by adolescent experience throughout multiple brain regions , but the changes were particularly long lasting in the ventral hippocampus—a brain region involved in mood regulation . Experimentally reducing the activity of the receptor recapitulated the behavioral abnormalities induced by the stress hormone . Meanwhile , pharmacological stimulation of the receptor corrected these abnormalities and had long-lasting antidepressant-like effects that persisted even after drug treatment ended . Thus , pharmacological interventions that augment trkB activity may be particularly efficacious in treating neurobehavioral deficits in adolescents exposed to chronic stress . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"swimming",
"medicine",
"and",
"health",
"sciences",
"brain",
"vertebrates",
"mice",
"neuroscience",
"habits",
"mammals",
"animals",
"biological",
"locomotion",
"animal",
"behavior",
"zoology",
"neuronal",
"dendrites",
"animal",
"cells",
"behavior",
"neostriatum",
"amygdala",
"rodents",
"cellular",
"neuroscience",
"eukaryota",
"cell",
"biology",
"anatomy",
"physiology",
"neurons",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"amniotes",
"organisms"
] | 2017 | Regulation of actions and habits by ventral hippocampal trkB and adolescent corticosteroid exposure |
Yaws is endemic in Ghana . The World Health Organization ( WHO ) has launched a new global eradication campaign based on total community mass treatment with azithromycin . Achieving high coverage of mass treatment will be fundamental to the success of this new strategy; coverage is dependent , in part , on appropriate community mobilisation . An understanding of community knowledge , attitudes and practices related to yaws in Ghana and other endemic countries will be vital in designing effective community engagement strategies . A verbally administered questionnaire was administered to residents in 3 districts in the Eastern region of Ghana where a randomised trial on the treatment of yaws was being conducted . The questionnaire combined both quantitative and qualitative questions covering perceptions of the cause and mechanisms of transmission of yaws-like lesions , the providers from which individuals would seek healthcare for yaws-like lesions , and what factors were important in reaching decisions on where to seek care . Chi-square tests and logistic regression were used to assess relationships between reported knowledge , attitudes and practices , and demographic variables . Thematic analysis of qualitative data was used to identify common themes . A total of 1 , 162 individuals participated . The majority of individuals ( n = 895 , 77% ) reported that “germs” were the cause of yaws lesions . Overall 13% ( n = 161 ) of respondents believed that the disease was caused by supernatural forces . Participants frequently mentioned lack of personal hygiene , irregular and inefficient bathing , and washing with dirty water as fundamental to both the cause and the prevention of yaws . A majority of individuals reported that they would want to take an antibiotic to prevent the development of yaws if they were asymptomatic ( n = 689 , 61 . 2% ) , but a substantial minority reported they would not want to do so . A majority of individuals ( n = 839 , 72 . 7% ) reported that if they had a yaws-like skin lesion they would seek care from a doctor or nurse . Both direct and indirect costs of treatment were reported as key factors affecting where participants reported they would seek care . This is the first study that has explored community knowledge , attitudes and practices in relation to yaws in any endemic population . The belief that ‘germs’ are in some way related to disease through a variety of transmission routes including both contact and dirty water are similar to those reported for other skin diseases in Ghana . The prominent role of private healthcare providers is an important finding of this study and suggests engagement with this sector will be important in yaws eradication efforts . Strategies to address the substantial minority of individuals who reported they would not take treatment for yaws if they were currently asymptomatic will be needed to ensure the success of yaws eradication efforts . The data collected will be of value to the Ghana Health Service and also to WHO and other partners , who are currently developing community mobilisation tools to support yaws eradication efforts worldwide .
Yaws , caused by Treponema pallidum subsp . pertenue , is endemic in Ghana . The disease is reported in all districts but predominantly in the south of the country . The majority of clinical disease is seen in young children , with the organism being transmitted by skin to skin contact with infectious lesions . In 2012 , the World Health Organization ( WHO ) launched a new global eradication campaign based on total community mass treatment ( analogous to mass drug administration—MDA ) with azithromycin [1] . This involves the treatment of all individuals in an endemic community regardless of the presence or absence of clinical yaws . Achieving high coverage of mass treatment will be fundamental to the success of this new strategy [2]; coverage is dependent , in part , on appropriate community mobilisation . The success of interventions may be jeopardized by non-participation within endemic communities [3] , because it may leave reservoirs of infection from which disease may re-emerge . Studies in Ghana have established that amongst residents of rural communities there is a strong conceptual connection between lack of cleanliness and disease causation [4–6] . In a study conducted within Bono ( Akan ) society , sickness and disease may be attributed to many causes , including water ( nsuo ) , dirt ( efi , ) lack of personal hygiene , cleanliness , the environment in which one sleeps , and bad air or wind ( mframa ) [6] . Another study reported that beliefs about dirty bodies ( shoefi ) are central to the understanding of disease for the Akan , contributing to the importance attached to daily bathing [5] . Many traditional belief systems incorporate a strong correlation between microorganisms ( mmoa ) and dirt ( efi ) , in which efi is “synonymous with contamination and disease , and implies the necessity of cleansing and purification” [6] . Community perceptions of disease causation may play an important role in access to or utilization of health services [7 , 8] . There are no publications on community knowledge , attitudes and practices ( KAP ) related to yaws in Ghana or other endemic countries . However , an understanding of these issues will be vital in designing effective community engagement strategies to support the yaws eradication effort . This study was conducted alongside a randomised control trial investigating the optimal dose of azithromycin for treatment of yaws in Ghana . The study was conducted in three districts ( total population 319 , 898 ) which had reported 1 , 505 cases of yaws in the preceding 36 months , and aimed to collect both quantitative and qualitative data on local knowledge , attitudes and practices concerning the causation and treatment of yaws .
A verbally administered questionnaire was developed , based on existing information about yaws and previous studies on community beliefs around skin disease in West Africa [9 , 10] . Prior to commencing this study , the questionnaire was pre-tested with a small number of participants from a group of communities not included in recruitment of the definitive sample of respondents . The questionnaire was refined based on this experience to ensure clarity of questioning . This study was conducted in 3 districts in the Eastern region of Ghana namely Ayensuanor , West Akyem and Upper West Akyem in 2015 . These contiguous districts were selected out of 4 districts where a randomised trial on different doses of azithromycin for the treatment of yaws was being conducted ( ClinicalTrials . gov identifier NCT02344628 ) . The fourth district was 400 km away and could not be included in the KAP study because of distance and costs reasons . None of these districts had previously received azithromycin mass treatment for yaws or trachoma . Within the chosen districts , multi-stage sampling was used to select participants for the adjunctive investigations described here . The first level of sampling was communities in the study districts which were selected using simple random sampling . In each selected community , an estimate of the total adult population was obtained from the district directorate of health services . In selected communities , systematic sampling was used to select participating households by randomly selecting an initial house and then utilising a quasi-random clockwise walking method [11 , 12] . In each house , either the oldest male or female resident was interviewed . If no adults were present at the time of visit , the house was excluded . In order to obtain a gender-balanced survey , the selection of respondent was alternated by gender between houses where possible . Questionnaires were administered by local staff from each district . Participants were shown photos of typical yaws lesions as examples of the disease being discussed . Data were collected on demographics , including gender , ethnicity , religion , age , highest level of education achieved and occupation . The questionnaire combined both quantitative and qualitative questions covering several themes , including perceptions of the cause and mechanisms of transmission of yaws-like lesions , the providers from which individuals would seek healthcare if they themselves had a yaws-like lesions , and what factors were important in reaching decisions on where to seek care . Prior to conducting fieldwork , all staff received training on qualitative techniques and on administering the questionnaire , with training provided by the lead member of the KAP study team ( MMA ) . Questionnaires were administered in local languages , and recorded on a standard form . Data were entered into a study database in EpiInfo by one of two members of the study team ( either RD or BO ) . Quantitative data were analysed with the use of descriptive statistics . Categorical variables were summarized using absolute numbers and percentages . Chi-square tests and logistic regression were used to assess relationships between reported knowledge , attitudes and practices , and demographic variables , including age , ethnicity , religion and level of education . These analyses were performed using STATA 13 . 1 ( Statacorp ) . For the purposes of qualitative analysis , we considered knowledge to be participants’ reported beliefs around the causation and transmission of yaws , attitudes to reflect what individuals believed could or should be done to prevent disease , and practices to encompass the healthcare seeking behaviour of individuals with yaws and the underlying reasons for presenting . Qualitative data were assessed across each of these domains . Data were manually codified to identify common themes . Representative examples of major themes were identified in each domain . The studies were approved by the World Health Organization ( RPC 720 ) , London School of Hygiene & Tropical Medicine ( LSHTM 8832 ) and Ghana Health Service ( GHS 13/11/14 ) ethics committees . Written informed consent was obtained from all participants .
A total of 1 , 162 individuals participated . Slightly more women ( n = 600 , 51 . 6% ) were included than men , and the median age of participants was 36 years ( IQR 27–46 years ) . The majority of participants were Christians ( n = 966 , 83 . 1% ) and of Akan ethnicity ( n = 593 , 51 . 0% ) ; other reported religious affiliations and ethnicities are shown in Table 1 . A minority of participants reported a personal history of a lesion consistent with yaws ( n = 214 , 18 . 4% ) . The majority of individuals ( n = 895 , 77% ) reported that “germs” were the cause of yaws lesions . Contact with an individual with a similar lesion was reported to be an important cause by 532 participants ( 45 . 8% ) . Many individuals ( n = 816 , 70 . 2% ) incorrectly believed that washing in or drinking dirty water were possible routes of transmission ( Table 2 ) . There was substantial overlap in beliefs with the majority of those who believed contact with an infected individual was a route of transmission also believing that washing or drinking water played a role ( n = 457 , 85 . 9% ) . In open-ended questions many respondents ( 38 . 1% ) also suggested transmission was more broadly related to hygiene , contact with dirty water or a lack of washing . 13% ( n = 161 ) of respondents believed that the disease was caused by supernatural forces such as witchcraft , curses or a punishment from god . Belief in a supernatural cause was not associated with gender , religion or ethnic group ( p > 0 . 05 for all comparisons ) . Participants frequently mentioned lack of personal hygiene , irregular and inefficient bathing , and washing with dirty water as fundamental to both the cause and the prevention of yaws . More broadly , many participants asserted that bathing and personal hygiene were closely associated to wider notions of longevity of life and health , stating that bathing was fundamental to healthy living . Contact with individuals and/or sharing of sponges or other items used in bathing were also reported as playing an important role in disease transmission , while some individuals reported a belief that transmission could be caused by airborne spread of germs . The over-riding theme emerging from the interviews was that dirt , dirtiness and germs were inextricably connected to notions of cleanliness in understanding the mechanism of transmission , and that personal hygiene was believed to be the key intervention to protect individuals from germs and dirt—and therefore from infection . A majority of individuals reported that they would want to take an antibiotic as part of a mass treatment campaign to treat yaws , even if they themselves were asymptomatic ( n = 689 , 61 . 2% ) , but a substantial minority reported they would not want to do so ( 31 . 8% ) . After adjustment for confounders , neither religion nor level of education were associated with individuals reporting that they would accept treatment for yaws ( p = 0 . 06 and p = 0 . 09 respectively ) . After controlling for other factors individuals of Ga-Adangbe were slightly more likely to report willingness to accept treatment ( OR 1 . 6 95% CI 1 . 05–2 . 45 , p = 0 . 01 ) . After controlling for demographic factors , belief in supernatural causation of yaws-like lesions was associated with a non-significant lower likelihood of accepting medication ( OR 0 . 53–95% CI 0 . 07–3 . 99 , p = 0 . 541 ) . A majority of individuals ( n = 839 , 72 . 7% ) reported that if they had a yaws-like skin lesion they would seek care from a doctor or nurse . A substantial minority of individuals reported that they would seek care through the private sector , most commonly a pharmacy ( n = 426 , 37% ) or a private shop ( n = 28 , 2 . 4% ) ( Table 2 ) . Many individuals also reported that they would seek care from a traditional healer or witchdoctor ( n = 496 , 42 . 7% ) . Belief in a supernatural cause for yaws was strongly associated with the individual reporting that they would seek care from a traditional healer or witchdoctor ( OR 2 . 93–95%CI 2 . 1–4 . 2 , p <0 . 0001 ) . Many individuals ( n = 276 , 36 . 8% ) reported that they would seek care both from the formal healthcare sector ( doctor , nurse , hospital ) and a traditional healer or witchdoctor . Cost was the factor most commonly reported by respondents as being important in determining where they sought care ( n = 1 , 080 92 . 9% ) , with both the direct cost of treatment ( n = 1 , 029 , n = 88 . 6% ) and the cost of travel ( n = 646 , 55 . 6% ) reported as considerations . The opinions of family members or an individual’s previous experience were relatively infrequent considerations ( n = 118 , 10 . 2% ) in guiding treatment seeking behaviours ( Table 2 ) . The qualitative data confirmed that cost was a major driver in decision making about where to seek care: Participants reported receiving a wide range of different treatments for yaws-like lesions . The most commonly reported treatments were tablets ( n = 505 , 43 . 6% ) or injectable antibiotics ( 46 . 2% ) but many people also reported receiving topical treatments ( Table 2 ) .
We have explored community knowledge , attitudes and practices with regards to yaws in southern Ghana , a region of the country where yaws is endemic . Whilst most participants reported that infection was in some way related to “germs” , they reported a variety of perceived routes of transmission , including both correct ( contact with infected individuals ) and incorrect ( e . g , contact with dirty water ) routes of transmission . These findings are in keeping with studies undertaken in Ghana concerning knowledge , attitudes and practices in relation to Buruli ulcer , which have emphasised the perceived role of contact with infected individuals and drinking dirty water as important routes of transmission for that disease [9 , 10] . In the study reported here , almost 15% of participants believed that witchcraft or curses could be responsible for skin ulcers . Attributing an illness to supernatural forces such as witchcraft and curses is a common phenomenon and is often due to a lack of knowledge of the true aetiology of illness , particularly when the illness is prolonged [9] . Many individuals in Buruli ulcer endemic areas of Ghana report similar beliefs about causation of disease [9 , 10] . Whilst the majority of individuals reported that they would seek care from a doctor or a nurse for the management of yaws-like lesions , the prominent role of private healthcare providers is an important finding of this study . In Ghana , most individuals are eligible for free health care via a national health insurance programme and most individuals have access to a local primary health care facility . However indirect costs and time to travel to state providers may still represent a barrier to accessing ‘free’ health-care and may be a reason for seeking alternative care provider if that is more readily available locally [13] . As in other parts of sub-Saharan Africa private pharmacies are also common . There individuals may purchase a wide-range of medications directly without the need for medical consultation . More than a third of individuals reported that they would seek care from private providers: specifically staff at private pharmacies or traditional healers . Acceptance of a given intervention can depend on a variety of factors , including understanding of causation , perceived effectiveness of the intervention and both direct and indirect cost factors [7] . Our findings that cost and distance represent barriers to accessing healthcare are in keeping with other studies from Ghana [14 , 15] . Whilst community-based MDA programmes should address cost issues , health education will be needed to address beliefs around causation and the effectiveness of azithromycin against yaws . Yaws eradication programmes should also consider how best to engage with these other providers , who operate outside of the formal healthcare system , and who may be an essential means for detecting cases of yaws in endemic communities as the eradication endpoint is approached . They should be considered both in Ghana and in other settings where such informal providers play important roles alongside formal healthcare systems . As the majority of yaws cases occur amongst children , consideration could also be given to training school teachers to identify signs and symptoms of yaws , and engaging them in the yaws eradication effort and further studies exploring this possibility should be undertaken . A sizeable minority of individuals in our study reported that they would not be willing to take medicine for yaws if they were asymptomatic at the time that it was offered to them . This is potentially worrisome , as estimates suggest that for each symptomatic case of yaws , there are 5–10 individuals with asymptomatic infection , and that latent yaws can reactivate up to a decade after primary disease has resolved in an individual; this , in fact , is part of the basis for the use of an initial strategy of total community treatment , rather than one involving identification and treatment of those with disease [16] . If our unwillingness rate was translated into non-participation rates in yaws eradication campaigns , it would represent a significant challenge for interrupting transmission in Ghana and likely increase the number of rounds of treatment required to reach that goal [2] . Studies of the acceptability of azithromycin MDA for trachoma have reported associations with gender and beliefs around disease causation [7] . In our respondents , we did not identify any clear demographic factors associated with the likelihood of treatment refusal when asymptomatic . The majority of cases of yaws occur in children ( who were not included in this survey ) but in the context of MDA for trachoma , the decision of the household head is often important in determining whether any members of a household receive azithromycin [3] . Interventions that increase the acceptability of treatment to household adults might also result in improved uptake of treatment amongst children . Individuals who reported a supernatural belief about causation , such as witch-craft or a curse , were more likely to report that they would not accept treatment , but this association did not reach statistical significance . More in-depth qualitative work should be considered to explore this issue in greater detail . This study has a number of limitations . First , we did not conduct in-depth interviews alongside the questionnaires . This may have limited our ability to explore issues arising from the questionnaire in more detail , or to fully explore the nuances of individual or community beliefs around yaws . Second , the clinical lesions of yaws are similar to many other skin diseases and we cannot be certain of the extent to which the beliefs expressed by community members are specific to yaws rather than applying more broadly to locally endemic skin diseases . Clinically important differentiations between diseases such as scabies , Buruli ulcer , yaws and leprosy may not be easily discernible to community members , although Buruli ulcer is not known to be endemic in any of the districtsin which this study was conducted . Although MDA with azithromycin may have ancillary benefits in treating some other skin diseases [17 , 18] it will not affect many common skin infections , such as Buruli ulcer , scabies or tinea . Health education programmes will need to clearly emphasise that whilst azithromycin will have a significant impact on yaws and some other diseases , it is not a panacea for all skin complaints . Finally we cannot be certain that individuals would always seek care for their children given how common skin disease is in many communities , nor that they would seek care for their children in the same place that they stated they would seek care for themselves . Data from other studies suggests that the attitude of the adult or head of household dictates engagement with healthcare services for all household members [3] so we believe that the viewpoints of adults would be key in determining healthcare entry points for children . Despite these limitations , this is the first study of which we are aware that has explored community knowledge , attitudes and practices in relation to yaws in any endemic population . We were able to collect quantitative and qualitative data from a broad cross-section of individuals across a range of ages , genders , and ethnic and religious groups in three districts in Ghana . Our data highlight the need to address barriers to accessing care including both direct and indirect healthcare costs . The data collected will be of value to the Ghana Health Service and also to WHO and other partners , who are currently developing community mobilisation tools to support yaws eradication efforts worldwide . | Yaws , a bacterial skin infection , is endemic in Ghana . WHO has launched a campaign to eradicate yaws based on community mass treatment with the antibiotic azithromycin . Community perceptions of disease are an important contributor to participation in mass treatment interventions . This study used questionnaires to understand beliefs about yaws amongst individuals living in endemic communities in Ghana . Most individuals reported that ‘germs’ were the cause of yaws although the route of transmission was less well understood with many individuals reporting that dirty drinking or washing water was responsible for transmission . Participants reported they would normally seek care within the formal healthcare sector although many individuals reported they would visit traditional healthcare providers or pharmacies . Cost of care was the key factor for many participants . A majority of individuals reported they would be happy to take an antibiotic to prevent infection but a large minority ( 38 . 8% ) reported that they would not . This study provides the first data on community beliefs about yaws in Ghana and will be valuable in helping the Ghana Health Service and partners develop community mobilisation tools to support yaws eradication efforts worldwide . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"dermatology",
"medicine",
"and",
"health",
"sciences",
"demography",
"tropical",
"diseases",
"geographical",
"locations",
"health",
"care",
"treponematoses",
"bacterial",
"diseases",
"health",
"care",
"providers",
"skin",
"infections",
"global",
"health",
"neglected",
"tropical",
"diseases",
"africa",
"skin",
"diseases",
"public",
"and",
"occupational",
"health",
"infectious",
"diseases",
"hygiene",
"yaws",
"people",
"and",
"places",
"ghana"
] | 2017 | Knowledge, attitudes and practices towards yaws and yaws-like skin disease in Ghana |
The Eastern North American monarch butterfly , Danaus plexippus , is famous for its spectacular seasonal long-distance migration . In recent years , it has also emerged as a novel system to study how animal circadian clocks keep track of time and regulate ecologically relevant daily rhythmic activities and seasonal behavioral outputs . However , unlike in Drosophila and the mouse , little work has been undertaken in the monarch to identify rhythmic genes at the genome-wide level and elucidate the regulation of their diurnal expression . Here , we used RNA-sequencing and Assay for Transposase-Accessible Chromatin ( ATAC ) -sequencing to profile the diurnal transcriptome , open chromatin regions , and transcription factor ( TF ) footprints in the brain of wild-type monarchs and of monarchs with impaired clock function , including Cryptochrome 2 ( Cry2 ) , Clock ( Clk ) , and Cycle-like loss-of-function mutants . We identified 217 rhythmically expressed genes in the monarch brain; many of them were involved in the regulation of biological processes key to brain function , such as glucose metabolism and neurotransmission . Surprisingly , we found no significant time-of-day and genotype-dependent changes in chromatin accessibility in the brain . Instead , we found the existence of a temporal regulation of TF occupancy within open chromatin regions in the vicinity of rhythmic genes in the brains of wild-type monarchs , which is disrupted in clock deficient mutants . Together , this work identifies for the first time the rhythmic genes and modes of regulation by which diurnal transcription rhythms are regulated in the monarch brain . It also illustrates the power of ATAC-sequencing to profile genome-wide regulatory elements and TF binding in a non-model organism for which TF-specific antibodies are not yet available .
Organisms have evolved daily and seasonal rhythms in behavior , physiology and metabolism that are driven by biological clocks to adapt to a temporally dynamic environment . The Eastern North American migratory monarch butterfly , Danaus plexippus , has emerged as a new system to study circadian clocks and their role in regulating daily and seasonal biology [1 , 2] . The sequencing of its genome [3] , along with the development of CRISPR/Cas9-mediated reverse-genetics to generate monarch clock gene knockouts [4–6] , have unlocked the potential of the monarch for studying animal clockwork mechanisms . Yet , in contrast to conventional model organisms like the fruit fly and the mouse , the identity of the genes expressed rhythmically over the course of the day and their cis-regulatory regions in the brain , the anatomical site driving daily rhythms in locomotor activity , remain unknown in the monarch system . Migratory monarchs are famous for their iconic seasonal long-distance migration and the use of a sun compass that allows both fall migrants and spring remigrants to navigate to their respective destinations [7–13] . Circadian clocks are part of the monarch navigational toolkit as they time-compensate the sun compass output in the brain , enabling migrants and remigrants to maintain a fixed flight direction throughout the day [7–9 , 11 , 12] . These clocks may also contribute to the seasonal induction of the migratory physiology and behavior and the timing of their migration [2] . The molecular oscillator that drives monarch circadian rhythms has been deciphered and relies , similar to other eukaryotes , on a transcriptional-translational core feedback loop running with a period of ~24 hours . Interestingly , the monarch possesses a hybrid clock that contains a mammalian-like core feedback loop and a Drosophila-like entrainment pathway [14] . Heterodimers of the transcription factors CLOCK ( CLK ) :BMAL1 drive the rhythmic transcription of the cryptochrome 2 ( cry2 ) , period ( per ) , and timeless ( tim ) genes [14 , 15] . Upon translation , CRY2 , PER , and TIM form complexes that translocate back into the nucleus where CRY2 , the orthologue of the mammalian CRY1 transcriptional repressor , inhibits CLK:BMAL1-mediated transcription [5 , 6 , 14 , 15] . Photic entrainment of the clock is similar to that found in the fruit fly , occurring through the circadian blue-light photoreceptor CRYPTOCHROME 1 ( CRY1 ) and the CRY1-TIM pathway [14 , 16–18] . The molecular circadian clock does not just regulate the expression of core clock components . It also drives the rhythmic expression of thousands of transcripts in a tissue-specific manner such that physiological functions are tuned to optimally perform at the most appropriate time of the day . The genes rhythmically expressed in light:dark or constant dark conditions and the transcription factors ( TFs ) that regulate their rhythmic expression have been identified in different tissues of several organisms from insects to mammals [19–29] . However , their identity remains unknown in the monarch . Given the potential of the monarch as a new genetic model to study how circadian clocks regulate daily and seasonal physiological and behavioral outputs , we aimed in this study at 1 ) identifying genes expressed rhythmically and under the control of circadian core clock components over the 24-hour light:dark cycle in the monarch brain , 2 ) identifying the cis-regulatory elements that contribute to rhythmic gene expression , and 3 ) characterizing the effect of clock disruption on cis-regulatory element accessibility and on the putative TFs that bind to them .
To determine the genes expressed rhythmically in the brain that may regulate physiological and behavioral rhythms in the monarch butterfly , we profiled genome-wide gene expression in the brains of wild-type adults collected every 3 hours in a 15-hr light:9-hr dark ( LD ) cycle using RNA-sequencing , with two independent replicates per time point . LD conditions were chosen instead of constant dark conditions because they reflect natural environments that drive diurnal rhythms of monarch activity and the 15:9 LD regime was used because it is the ecologically relevant photoperiod experienced by summer reproductive monarchs , and the condition under which we maintain our colony in the laboratory . Expression levels for each gene were examined for rhythmic variation using RAIN [30] and MetaCycle [31] , and genes with a maximum/minimum fold-change ≥ to 1 . 3 and an adjusted p value ( i . e . , corrected for multiple testing ) ≤ 0 . 05 , as defined by either method , were considered rhythmic ( see Methods for details ) . Using these criteria , we identified 431 rhythmic genes with peaks of expression distributed throughout day and night in the monarch brain ( Fig 1A; S1 and S2 Tables ) . To determine whether the rhythmic expression of these genes was dependent on core clock genes , we also profiled genome-wide temporal expression patterns every 6 hours in 15:9 LD cycle in the brains of two monarch knockout strains bearing a non-functional circadian clock , lacking either a functional circadian activator CLK or a functional circadian repressor CRY2 [4 , 5] . An analysis of differential rhythmicity between rhythmic gene expression in wild-type and gene expression in clock-deficient mutants using DODR [32] revealed that the expression of most rhythmic genes in wild-type monarchs was affected in Cry2 and Clk knockouts , and that this effect was significantly higher than background ( calculated using genes arrhythmically expressed in brains of wild-type ) ( Fig 1B ) . Consequently , only genes with an adjusted p-value cutoff of ≤ 0 . 05 were considered in subsequent analysis . Under these stringent conditions , the expression of 126 and 163 rhythmic genes was respectively affected in Cry2 and Clk knockouts ( Fig 1B; S3 and S4 Tables ) , 72 of which overlapped in both datasets ( Fig 1C; S5 Table ) . Interestingly , while the expression of the rhythmic genes differentially regulated in Cry2 knockouts peaked throughout the day , the genes differentially regulated in Clk knockouts mostly peaked during the first half of the light phase ( Fig 1C ) . Among the rhythmic genes identified , we found the CLK:BMAL1 direct target core clock genes Per and Tim , which exhibited similar profiles to those previously shown by qPCR in wild-type monarch brains [5] , as well as Cry2 ( Fig 1D ) . Some clock genes such as vrille ( Vri ) and clockwork orange ( Cwo ) , whose products respectively function in Drosophila as a regulator of clk expression [33 , 34] and as a repressor of CLK-mediated activation [35–38] , were also rhythmically expressed ( Fig 1D ) . The rhythmicity and phase of Per , Tim , Cry2 , Vri and Cwo expression in the brain of wild-type monarchs was confirmed by qRT-PCR on independent samples ( S1 Fig ) , validating our RNA-seq data . The mRNA levels of other known clock genes did not however show cyclic expression in the monarch brain . This includes Clk and Bmal1 , the blue-light photoreceptor Cry1 , and the homologue of par domain protein 1 ( Pdp1 ) that is required for behavioral rhythmicity in Drosophila [39] ( Fig 1D ) . If the rhythmic expression of core clock genes was solely controlled by the rhythmic activation of CLK:BMAL1 , these genes would be predicted to be expressed at constitutively low levels in the absence of transcriptional activation ( i . e . , in Clk knockouts ) and at constitutively high levels in the absence of transcriptional repression ( i . e . , in Cry2 knockouts ) . As expected , we found that all clock genes rhythmically expressed in wild-type brains were expressed at constitutive low levels in the brain of Clk knockouts ( Fig 1D ) . However , in the brains of Cry2 knockouts , only Per and Cwo were found expressed at constitutive high levels ( Fig 1D ) ; Vri was expressed at constitutively low levels and Tim at mid-levels ( Fig 1D ) , suggesting a complex regulation of their expression , either through the action of other factors on the regulation of their transcription or via post-transcriptional events . Similar to rhythmic transcriptome analysis in brains of two other insects , Drosophila and the mosquito Anopheles gambiae [21 , 40–42] , we found that rhythmic genes in the monarch brain belong to diverse biological processes that include transmembrane transport , several metabolic processes , and regulation of DNA binding ( S2 Fig; S6 Table ) . KEGG pathway enrichment analysis further revealed that besides circadian rhythms , glycolysis , the biosynthesis of amino acids , carbohydrate metabolism , and other metabolic pathways were among the most enriched rhythmic pathways in the monarch brain ( S2 Fig; S7 Table ) . Rhythmic expression of genes involved in metabolic pathways are thought to temporally orchestrate metabolic processes over the course of the 24-hour day [43 , 44] . In the monarch brain , we found that many genes encoding key enzymes and regulatory proteins involved in glucose metabolism , which is essential to fuel basic brain physiology , were expressed rhythmically ( Fig 2A and 2B ) . Two trehalose transporters , Tret1-1 and Tret1-2 , which act in glial cells to uptake trehalose ( the main metabolite supplying energy in insects ) and provide the brain with energy and protection from neurodegeneration [45] , were rhythmically expressed with a peak of expression during the day ( Fig 2B and 2C ) . To provide the brain with energy and nurture neurons , trehalose must be metabolized through glycolysis , which is toxic in neurons [46] but essential in glia [45] . Consistent with the idea that glial glycolysis is rhythmically regulated upon conversion of trehalose to glucose , several genes involved in glycolysis were also rhythmically expressed with peaks of expression within a 4-hr window in the middle of the day ( Fig 2B and 2C ) . These included genes encoding ( i ) the enzyme involved in the rate-limiting step of glycolysis , 6-phosphofructokinase ( Pfk-1 ) , and ( ii ) other enzymes in the glycolytic pathway such as phosphoglucose isomerase ( Pgi ) , aldolase ( Ald ) , triose phosphate isomerase ( Tpi ) , glyceraldehyde 3 phosphate dehydrogenase 2 ( Gapdh2 ) , phosphoglycerate kinase ( Pgk ) , and enolase ( Eno ) ( Fig 2A and 2B ) . Interestingly , genes encoding two regulatory proteins , 6-phosphofructo-2-kinase/Fructose 2 , 6 biphosphatase ( Pfk-2/FBPase-2 ) , which activates Pfk-1 to increase the glycolytic rate , and pyruvate dehydrogenase kinase ( Pdk ) , which inhibits the activity of pyruvate dehydrogenase blocking the entry of pyruvate into the TCA cycle , were found to cycle in phase with one another ( Fig 2B and 2C ) . These results strongly suggest that pyruvate catabolism is inhibited in phase with the regulation of glycolysis in glia . Among genes expressed rhythmically with a peak of expression in the middle of the day , we also found three mRNAs of key players of glycogenesis/glycogenolysis , glycogen synthase ( GlyS ) and 1 , 4-alpha-glucan branching enzyme ( AGBE ) , which generate highly branched glycogen molecules to store glucose for quick energy use , and glycogen phosphorylase , the rate-limiting step in glycogenolysis ( Fig 2A , 2B and 2C ) . Interestingly , phosphoglucose mutase ( Pgm ) , which links glycolysis to glycogenesis ( Fig 2A ) , also cycled in a phase similar to those of genes regulating both pathways ( Fig 2B and 2C ) . Finally , a gene involved in de novo synthesis of trehalose , trehalose 6-phosphate synthase ( Tps1 ) was also rhythmically expressed in the monarch brain with a peak of expression during the day ( Fig 2A , 2B and 2C ) , suggesting a possible local rhythmic synthesis of trehalose . Importantly , the rhythmicity of the genes encoding key enzymes in glycolysis and glycogenesis appeared to be under the regulation of core clock components as their daily rhythms were , for the most part , disrupted in the brains of Cry2 and Clk knockouts ( S3 Fig ) . Together , our data provide evidence of a tight coordination in the temporal expression of genes involved in glucose metabolism in the insect brain , likely optimizing the timing at which energy substrates are produced to fuel the neurons and sustain their daily activities . Brain neuronal activity and communication within neural networks , which ultimately regulates physiology and behavior , occur via neurotransmission and G- protein coupled receptor ( GPCR ) signaling . In Drosophila and mammals , circadian rhythms are regulated by a number of neurotransmitters , neuropeptides and corresponding receptors that are under circadian clock control to temporally relay information across the clock network [47–51] . Not surprisingly , genes involved in cholinergic , glutamatergic , GABAergic and GPCR signaling were also found rhythmically expressed in the monarch brain ( Fig 3 ) . However , we noted interesting features of regulation of gene expression consistent with a temporal separation of chemically antagonistic processes . mRNAs encoding acetylcholinesterase ( AChE ) , the enzyme that breaks down acetylcholine ( ACh ) released by presynaptic terminals to decrease Ach signaling [52] , and the choline transporter , which transports choline in the presynaptic terminal for the synthesis of ACh , were expressed rhythmically in anti-phase to one another ( Fig 3A ) . Rhythms of choline transporter were abolished in the brains of Cry2 and Clk knockouts ( S4A Fig ) , and those of AchE were abolished in the brains of Cry2 knockouts ( S4A Fig ) . These data suggest a temporal partitioning of the synthesis and degradation of ACh over the course of the day that could be controlled by the circadian clock . Genes involved in the recycling and reception of the neurotransmitters glutamate and γ-aminobutyric acid ( GABA ) , which are known to drive circadian rhythms in Drosophila and mammals [53 , 54] , were also found rhythmically expressed in the monarch brain . In neurons , glutamate levels can be decreased through its conversion to alpha-ketoglutarate by the enzyme glutamate oxaloacetate transaminase ( GOT ) ( Fig 3B ) . We found that mRNA levels of GOT cycled in the monarch brain , with a peak of expression during the day ( Fig 3B ) , and that the rhythms were disrupted in clock-deficient monarch strains ( S4B Fig ) . Glutamate and GABA modulate the neuronal excitability of post-synaptic neurons through their respective action on glutamate receptors and GABA receptors ( GABAR ) . Extracellular glutamate at the synapse is recycled by increased glial activity of glutamate transporters ( excitatory amino acid transporters; EaaT ) [55] . Interestingly , GABAR and EaaT2 mRNA levels were also rhythmic in the monarch brain ( Fig 3B ) and significantly disrupted in Cry2 and Clk knockouts , with the exception of GABAR in Cry2 knockout where significance was not reached despite an apparent loss of rhythm ( S4B Fig ) . Together , these data suggest that the circadian clock may rhythmically regulate neuronal excitability by modulating both the expression of genes involved in glutamate clearance at the synapse and of GABA receptors . Finally , we also found daily rhythms in mRNA levels of a G-coupled protein receptor ( GPCR ) of unknown function ( Fig 3C ) that are abolished in Cry2 and Clk knockouts ( S4C Fig ) , suggesting that it could be involved in modulating circadian behaviors . Similar to Drosophila [27] , rhythmic genes cycled in the monarch brain with phases of expression distributed throughout the day ( Fig 1A ) . How the expression of these rhythmic genes is temporally orchestrated is however not fully understood . As shown in flies and mice , both transcriptional and post-transcriptional processes contribute to the generation of rhythmic mRNA levels [56–60] . Transcriptional regulatory mechanisms such as rhythmic accessibility of DNA regulatory elements [22] or rhythmic activities of distinct TFs [61–64] have been shown to control diurnal rhythms in transcription and could also contribute to the generation of different phases of gene expression in the monarch brain . To investigate whether rhythmic chromatin accessibility plays a role in the control of rhythmic gene expression in the monarch , we profiled open chromatin regions using ATAC-seq in the brains of wild-type monarchs and of two mutants deficient in circadian activation ( Clk knockout and Cyc-like , a Bmal1 mutant lacking the C-terminal transactivation domain that mimics the Drosophila cycle gene [6] ) at Zeitgeber Time 04 ( ZT04; i . e . , 4 hours after lights on ) and at ZT16 , with two independent replicates per time point ( Fig 4 ) . ZT04 was chosen as the day time point because it corresponds to the trough of Per and Tim RNA expression levels and thus the likely time at which transcription is activated for these genes , and ZT16 was chosen as the night time point such that the two would be separated by 12 hours . Across all samples , 13 , 555 to 23 , 222 ATAC-seq peaks were identified throughout the genome ( S8 Table ) . These numbers are within the same order of magnitude as that obtained in Drosophila , which harbors a comparable genome size ( ~138 Mb for Drosophila versus ~249 Mb for monarch ) [65] . We also found good replicate concordance of the ATAC-seq peaks between biological replicates ( Fig 4A and S5A Fig; R2 ranging from 0 . 76 to 0 . 85 ) . As expected , Tn5 tagmentation of purified naked genomic DNA from brain did not display characteristic ATAC-seq peaks ( Fig 4B ) . These results strongly indicate that ATAC-seq can reliably and reproducibly measure chromatin accessibility in the monarch brain . Visual inspection of the ATAC-seq peaks from different genotypes and time-of-day revealed broadly similar profiles of chromatin accessibility ( Fig 4B ) , even in regions associated to clock genes that display robust rhythms of expression ( Fig 4C; see Fig 1D for mRNA expression patterns ) . A hierarchical clustering of consensus ATAC-seq peaks for all genotypes and time points also showed that they were strongly correlated with one another ( Fig 4D ) . The lack of significant differences in ATAC-seq peaks was further validated by quantifying differential Tn5 integration signals using DESeq2 [66] with Log2 fold-change > 0 . 3785 and FDR < 0 . 05 as thresholds . Even at this relatively low fold-change cutoff , less than 0 . 64% of ATAC-seq peaks genome-wide or associated with rhythmic genes differentially regulated in Clk knockouts were found to be significantly altered between time-of-day ( ZT04 vs . ZT16 ) and genotypes ( wild-type vs . clock impaired mutants ) ( Fig 4E; S9 Table ) . While a lack of statistically significant differences is not evidence of absence of differences , previous reports showing that ATAC-seq is sensitive enough to identify differences in chromatin accessibility in Drosophila [67] support the idea that there is no significant temporal regulation of significant opening/closing of the chromatin in the monarch brain . To further characterize our ATAC-seq datasets , we mapped the genomic location of each ATAC-seq peak and found that genome-wide peaks were enriched by ~ 3 to 4-fold in promoter regions ( Fig 4F ) . A similar enrichment was found for peaks in the promoter regions of genes rhythmically expressed in the brain and differentially regulated in Clk knockouts ( Fig 4F ) . Many genome-wide peaks were also found in introns and intergenic regions , as well as in exons and transcription termination sites ( S5B Fig ) , likely representing the full set of enhancer elements in the monarch brain . The lack of cyclic chromatin accessibility as measured by ATAC-seq does not exclude the possibility that the DNA regulatory elements within ATAC-seq peaks could be involved in driving rhythmic gene expression through the rhythmic binding of TFs . In addition to identifying chromatin accessibility , ATAC-seq can also be used to reveal DNA footprints , i . e . genomic regions protected from Tn5 integration because of DNA-bound proteins like TFs [68 , 69] ( Fig 5A ) . To test whether rhythms in TF occupancy within DNA regulatory elements in promoters and enhancers could underlie rhythmic gene expression , we performed a footprinting analysis within the open chromatin of 339 ATAC-seq peaks associated with the 163 rhythmic genes differentially regulated in Clk knockouts using the Wellington TF footprinting algorithm [70] . We detected 202 and 125 statistically significant footprints in brains of wild-type monarchs at ZT04 and at ZT16 , respectively ( p-value ≤ 10−10; FDR ≤ 0 . 01 ) . Of these , 130 and 53 were respectively specific to ZT04 and to ZT16 ( Fig 5B ) . Most of these time point-specific footprints exhibited footprint signals below the statistical threshold at the other time point rather than being completely absent . This therefore suggested that the depth of TF footprints could be regulated in a time-of-day dependent manner in monarch brains . To test this idea , we averaged the signal of footprints specific to a given time point and compared it to the average signal observed at the other time point ( Fig 5C and 5D ) . Averaged footprint signal and quantification confirmed a time-of-day dependent regulation of TF binding . For TF footprints specific to ZT04 , we observed a significant increase in footprint depth at ZT04 relative to ZT16 ( Fig 5D ) that appeared to be caused by both increased Tn5 integration in regions directly flanking the footprint and decreased Tn5 integration within the footprint ( Fig 5C ) . These results suggest that TF binding at ZT04 may occupy their DNA binding sites with a longer residence time and may also increase to some extent chromatin accessibility in their immediate surrounding regions . In contrast , TF footprints specific to ZT16 did not exhibit a significant increase in footprint depth at ZT16 compared to ZT04 ( Fig 5C and 5D ) . To verify that time-of-day dependent TF binding is clock-dependent , we measured the footprint signal at the same genomic regions in brains of Clk knockouts and Cyc-like mutants at ZT04 and ZT16 ( Fig 5B , 5C and 5D ) . We found that the time-of-day differences in TF footprint signal specific to ZT04 observed in wild-type monarch brains were abolished in Cyc-like mutants , in which the footprint signal was not significantly different than the signal observed at ZT16 in wild-type ( Fig 5C and 5D ) . Although the signal in Clk knockouts was not significantly different to the one observed at ZT16 in wild-type , there was a significant difference between time points , with a moderate increase at ZT16 compared to ZT04 ( Fig 5C and 5D ) . Together , these data suggest that the time-of-day differences in TF binding found in wild-type monarch may underlie rhythmic gene expression . Furthermore , the differences in footprint signal between ZT04 and ZT16 suggested that distinct classes of TFs could differentially bind DNA at different times of the day . To identify these TF classes , we performed a motif enrichment analysis in footprints found specifically at ZT04 and at ZT16 in brains of wild-type monarchs . The analysis revealed a large number of predicted TF families preferentially binding at both and either time points ( Fig 5E ) . Among the predicted TF classes , the bHLH , forkhead , HTH and MADS TF families appeared as among the most overrepresented in footprints at ZT04 in comparison to those at ZT16 . These data suggest that TFs belonging to these families could be responsible for the increased DNA accessibility observed in footprints at ZT04 .
Numerous studies over the past 15 years have profiled daily and circadian transcriptional rhythms in various organisms and tissues [19–29] . Most genome-wide rhythmic gene expression studies performed to identify cycling mRNAs in the insect brain have however largely relied either on the use of whole heads [19 , 21 , 25–29] , from which mRNAs derive mainly from the compound eyes , or on the use of specific subsets of clock neurons in the brain [40 , 71] . To date , a single study has profiled the circadian and daily transcriptome in the Drosophila brain [72] . With 217 robust rhythmic mRNAs whose expression was affected in Cry2 and Clk knockouts , our daily transcriptome in the monarch brain provides a useful complement to that of Drosophila for future comparative approaches . Consistent with previous reports [5 , 11 , 14] , Per and Tim were identified in our RNA-seq study as core clock genes that cycle robustly in the monarch brain , validating the quality of the RNA-seq datasets . Our study extended the set of core clock genes with robust rhythms to Vri , Cwo , and to the mammalian-like circadian repressor Cry2 , although the latter cycled with lower amplitude . While all cycling core clock genes were expressed at constitutive low levels in Clk knockouts , as anticipated in the absence of activation , their expression in Cry2 knockouts ranged from low to high levels , suggesting a complex regulation of their expression . Although the mechanistic underpinnings are not yet understood , this phenomenon is not without precedent as the levels of CLK-CYC direct targets in tim and per Drosophila mutants are also intermediate [25] . In contrast to the above-mentioned core clock genes , neither the circadian activators Clk and Bmal1 , the blue-light photoreceptor Cry1 , or Pdp1 were expressed rhythmically in the monarch brain . Because the brain is a heterogeneous tissue , it is possible that these genes could be expressed both in non-clock cells and in a subset of clock cells in which they could oscillate . Alternatively , post-transcriptional and/or post-translational mechanisms may be responsible for their rhythmic function . Future studies using in situ hybridization in the brain and/or a population of homogeneous clock cells , such as the DpN1 monarch cell line containing a light-driven clock [14] , could help distinguish between these possibilities . A number of genes relevant to brain physiology were also found to be rhythmic in brains of wild-type monarchs . Similar to a study in the Antarctic krill [73] , our data revealed a temporal regulation of key enzyme-encoding genes involved in glucose , trehalose and glycogen metabolism; however , the expression of these genes appeared to be coordinated in the monarch , peaking within a 7-hr window in the middle of the day . Although our data do not provide information on the rhythmic production of glycolytic metabolites , they indicate that fuel production in the brain is under tight rhythmic regulation at the level of gene expression , and likely increased during the monarch active phase . Pronounced rhythms in the abundance of metabolites and metabolic fluxes have however been recently proposed to emerge from phase lags rather than from coordinated expression of key enzymes and/or of regulatory proteins at the transcriptional levels [74] . This does not seem to be the case for glucose metabolism in the monarch brain because the genes that respectively encodes the main regulatory proteins Pfk-2 ( which stimulates glycolysis ) and Pdk ( which blocks the entry of pyruvate into the TCA cycle ) cycle in phase with one another . One interesting possibility could be that instead of being converted to acetyl-CoA , pyruvate is converted into lactate and alanine in brain glial cells , which once secreted , are taken up by neurons to fuel the neuronal TCA cycle and to generate the ATP needed for synaptic transmission , as previously shown in Drosophila [45] . If this is the case , the coordination of rhythmic gene expression in glial glucose metabolism could be a mechanism to fuel the neurons during the active phase of the monarch , i . e . during the day . Likewise , glycolysis and glycogenesis appeared to be temporally coordinated through phosphoglucose mutase . Because this enzyme acts in glycogenesis when glucose levels are high and in glycogenolysis when glucose levels are low , the temporal coordination of these pathways may serve as a homeostatic mechanism ensuring proper , rhythmic fuel production even in period of fasting . Importantly , we found that the rhythms of almost all of the genes involved in glucose metabolism were abolished in Cry2 and Clk monarch mutants with impaired clock function . Together with a report showing circadian regulation of metabolic genes in Drosophila heads [75] , our data indicate that the circadian clock likely plays an important role for regulating daily rhythm of glucose metabolism and glucose homeostasis in the insect brain . By revealing anti-phase rhythms of expression of key genes involved in the transport/synthesis and degradation of Ach , which are dependent on the presence of both circadian activator and repressor , our data support the idea that the circadian clock also plays a role in temporally separating chemically antagonistic processes . Like ACh , GABA is a widespread neurotransmitter in the insect brain , and similar to Drosophila [71] , a GABA receptor is expressed rhythmically in the monarch brain . Together , our findings that both cholinergic and GABAergic signaling are rhythmic and dependent on a functional circadian clock may suggest the existence of a more complex neuronal clock network than the four clock cells described in each hemisphere of the pars lateralis [14 , 76] . Developing CRISPR/Cas9-assisted methods to tag neurons in vivo would facilitate a comprehensive mapping of the neuronal clock network in the monarch brain . Our study also establishes ATAC-seq as an applicable method for the discovery of both genome-wide accessible chromatin regions and TF binding in a non-traditional model organism for which species-specific antibodies against TFs are not available . Although thousands of genome-wide regulatory elements were identified in the monarch brain with ATAC-seq , a surprising result was the absence of a significant temporal change in chromatin accessibility between day and night , as well as between wild-type and clock impaired mutants . This was surprising because diurnal changes in chromatin accessibility measured by DNase-seq have recently been reported in the mouse liver [22] . The different nature of tissues sampled could account for this difference . If so , our data would support the idea that the chromatin accessibility landscape in the adult brain could be fully programmed and not sensitive to changes in light:dark conditions . Alternatively , the contrasting results between our study and the study in the mouse liver could stem from differences in sensitivity between ATAC-seq and DNase-seq approaches . Although DNase-seq could be a more sensitive approach to quantify diurnal changes in chromatin accessibility , applying it to samples like the insect brain may not be achievable due to the large amount of material required . Regardless of the reasons underlying the lack of diurnal changes in chromatin accessibility in our study , performing a TF footprinting analysis in accessible chromatin regions associated with genes rhythmically expressed revealed time-of-day regulation of TF occupancy in the brain of wild-type monarchs . Interestingly , time-of-day dependent TF occupancy was completely abolished in Cyc-like mutants that lack the transactivation domain necessary for CLK:BMAL1-mediated transcriptional activation [6] . The situation found in Clk mutants was slightly different in that time-of-day dependent TF occupancy was impaired , but the impairment consisted of a reversed trend compared to TF occupancy in wild-type . Because the CLK partner BMAL1 is present in this mutant , it is possible that in absence of a functional CLK , BMAL1 heterodimerizes with another bHLH-PAS domain-containing protein to activate transcription in a different phase than that of CLK:BMAL1 . The juvenile hormone ( JH ) -receptor methoprene-tolerant ( MET ) could be a candidate , as it has been shown to form a heterodimer with BMAL1 in mosquitoes to activate the circadian transcription of JH-induced genes [77] . In addition , a motif enrichment analysis within open chromatin regions associated with rhythmic genes differentially expressed in Clk mutants revealed a large number of predicted TF families preferentially binding at either time point , indicative of the complexity of circadian regulation of gene expression in the monarch brain . Among the predicted TF classes , the bHLH , forkhead , HTH and MADS TF families appeared among the most overrepresented in footprints at ZT04 in comparison to those at ZT16 . These data suggest that TFs belonging to these families could be responsible for the increased DNA accessibility observed in footprints at ZT04 in brains of wild-type monarchs , consistent with the reported pioneer activities of bHLH and forkhead TFs [78 , 79] . Together , these data show that ATAC-seq can be used in non-traditional organisms to not only identify open chromatin regions but also predict dynamic TF binding . It is of particular importance because the overall lack of species-specific antibodies against TFs in non-model organisms continues to preclude the systematic use of chromatin-immunoprecipitation to comprehensively identify TF binding sites . Taken all together , our results represent the first analysis of daily transcriptome , DNA regulatory elements , and time-of-day dependent TF occupancy in the monarch butterfly brain . Given the central role of circadian clocks in the seasonal migration of this iconic insect , our datasets will be valuable resources to further our understanding of the molecular basis of seasonality and of migratory behavior .
Wild-type , Cry2 knockout , Clk knockout and Cyc-like mutant monarch butterflies were raised in the laboratory on semi-artificial diet under 15-hour light , 9-hour dark ( 15:9 LD ) conditions in Percival incubators at 25°C and 70% humidity , as previously described [4–6] . 15:9 LD was used because it is the ecologically relevant photoperiod experienced by summer reproductive monarchs , and the condition under which we maintain our colony in the laboratory . Adults were housed in glassine envelopes in the same lighting and temperature conditions , and were manually fed a 25% honey solution daily . Adult monarchs were entrained for a minimum of 7 days after eclosion in 15:9 LD cycles at 25°C and brains free of eye photoreceptors were dissected under a microscope in 0 . 5X RNA later ( Ambion ) to prevent RNA degradation , and stored at -80°C until use . For wild-type monarchs , three pooled brains were collected in two replicates at ZT1 , ZT4 , ZT7 , ZT10 , ZT13 , ZT16 , ZT19 , and ZT22 . Three pooled brains of Cry2 and Clk knockouts were each collected in two replicates at ZT4 , ZT10 , ZT16 , and ZT22 . For each sample , total RNA was extracted using RNeasy Mini kit ( Qiagen ) , polyA+ RNA was isolated from 2 μg of total RNA with NEBNext Oligo d ( T ) magnetic beads ( New England Biolabs ) , and libraries were prepared using the NEBNext Ultra Directional RNA Library Prep kit for Illumina and NEBNext Multiplex Oligos ( New England Biolabs ) and amplified with 12 PCR cycles , following the manufacturer’s recommendations . Library quality and size distribution was verified on a Bioanalyzer , libraries were quantified by real-time quantitative PCR , and 16 libraries were mixed in equimolar ratios for multiplexing and sequenced on a Hi-seq 2500 ( Illumina ) using 50bp single end reads . The resulting sequencing files were checked for quality control and demultiplexed by the Texas A&M AgriLife Genomics and Bioinformatics Facility . Reads were mapped to the monarch genome ( assembly v3; [80] ) using TopHat2 [81] with parameters “—read-realign-edit-dist 2 -g 1 —b2-sensitive” . On average , ~88% of the reads were mapped uniquely to the genome even in absence of rRNA depletion ( S10 Table ) . The total number of reads , mapped reads and mapping rate for each library are summarized in S10 Table . After mapping , gene expression levels were quantified in the brains of wild-type monarchs , Cry2 knockouts and Clk knockouts at each time point and for each replicate using Cufflinks [82 , 83] . Only genes with three or more reads per kilo base per million mapped reads ( RPKM ) in at least one time point were classified as expressed and further considered for subsequent analysis . To identify cycling mRNAs in the brain of wild-type monarchs , RAIN [30] and MetaCycle [31] were used with parameters “period = 24 , deltat = 3 , nr . series = 2” for RAIN and “adjustPhase =“predictedPer” , combinePvalue =“fisher” , timepoints = seq ( 1 , 46 , by = 3 ) , minper = 24 , maxper = 24” for MetaCycle . To consider oscillations determined by either method , the resulting p-values from RAIN and MetaCycle were combined using the minP method [84] and adjusted for multiple testing using the Benjamini-Hochberg ( BH ) procedure to control for false discovery rate ( FDR ) . Genes were considered rhythmically expressed when meeting the following criteria: ( 1 ) adjusted p-value ≤ 0 . 05 , and ( 2 ) fold-change ( maximal/minimal RPKM expression values within a time series ) ≥ 1 . 3 . Of the 15 , 130 genes in the monarch genome , ~ 69% ( 10 , 412 ) were expressed in the brain of wild-type monarchs , and ~4% ( 431 ) of these were determined as being rhythmically expressed . To determine if the oscillations of the rhythmic genes identified in wild-type were altered in Cry2 and Clk knockouts , a differential rhythmicity analysis was performed using robust DODR method [32] . After adjusting the p-values using the BH procedure , 126 and 163 genes rhythmically expressed in wild-type were respectively found to be differentially expressed in Cry2 and Clk knockouts . Phases were estimated using the R package for harmonic regression [85] . Heatmaps and phase plots were respectively generated using the R packages gplots [86] and ggplot2 [87] . Gene ontology ( GO ) terms for biological processes and KEGG pathway enrichment analysis were performed using Metascape ( http://metascape . org; [88] ) against the expressed genes in the wild-type monarch brains as background . Sequencing data are available at https://www . ncbi . nlm . nih . gov/geo/query/acc . cgi ? acc=GSE122447 ( Accession number GSE122442 ) . Brains from adult wild-type monarch butterflies entrained to seven 15:9 LD cycles after eclosion were dissected every 3-hours over a 24-hour day starting at ZT1 . Dissections were performed in 0 . 5X RNA later ( Ambion ) to prevent RNA degradation , and brains free of eye photoreceptors were immediately frozen and stored at -80°C until use . Total RNA was extracted using an RNeasy Mini kit ( Qiagen ) , treated with RQ1 DNase ( Promega ) , and random hexamers ( Promega ) were used to prime reverse transcription with Superscript II Reverse Transcriptase ( Thermo Scientific ) , all according to the manufacturers’ instructions . Quantifications of gene expression were performed on a QuantStudio 6 Flex Real-Time PCR System ( Thermo Scientific ) using iTaq Universal SYBR Green Supermix ( Bio-Rad ) , as previously described [4] . Monarch Per , Tim , and Rp49 expression levels were quantified using previously validated primers [4] . Monarch Cry2 , Vri , Cwo , GOT1 , GPCR , AGBE , Tps1 , and Eno expression levels were quantified using the following primers: Cry2F , 5’-TGGCTCTCATGCTCGTCTTTC-3’; Cry2R , 5’-ACCGCACTGGACAGTAGCAAT-3’; VriF , 5’-CGGACAGCGTAAGCAGAGAGA-3’; VriR , 5’-TCCCAGTAACCGTCGTCCTT-3’; CwoF , 5’-GCGCGCGCGCTTCAC-3’; CwoR , 5’-TCGATGCAGGGTTGGAAGTT-3’; GOT1F , 5’-CACAACCCCACAGGCATAGA-3’; GOT1R , 5’-CCATGACATCAGCGATCTTCTC-3’; GPCRF , 5’-GGGTACGAGCGGTATAGACATTG-3’; GPCRR , 5’- CTGCAAAGGACACTGGTCGAT-3’; AGBEF , 5’-CGGATGGCTCGCATCAA-3’; AGBER , 5’-TTTGTCGCCTTCATGCTTACA-3’; Tps1F , 5’-GACGGCGGGAAAAACAGA-3’; Tps1R , 5’-GCCTTGAGGAACGCCTTCA-3’; EnoF , 5’-GACTGTCGACCGGCCAGAT-3’; EnoR , 5’-ATTTGGCGAGACGCTCAGA-3’ . The near 100% efficiency of each primer set was validated by determining the slope of Ct versus dilution plot on a 3 x 104 dilution series . Individual reactions were used to quantify each RNA level in a given cDNA sample , and the average Ct from duplicated reactions within the same run was used for quantification . The data for each gene at a given time-point were normalized to Rp49 as an internal control , and normalized to the mean of one sample within a set for statistics . P-values were calculated using one-way ANOVA in R . Adult monarchs were entrained for a minimum of seven days after eclosion in 15:9 LD cycles at 25°C and brains free of eye photoreceptors were dissected in ice-cold ringer’s solution . Three pooled brains of wild-type , Clk knockout , and Cyc-like mutant monarchs were each collected in two replicates at ZT04 and at ZT16 . Each sample was resuspended twice in 600 μl of NP-40 lysis buffer ( 10mM Tris-HCl at pH 7 . 5 , 10 mM NaCl , 3 mM MgCl2 , and 0 . 1% NP-40 ) . Crude nuclei were prepared by gently homogenizing the brains in an ice-cold 2-ml Dounce homogenizer with two strokes of a loose-fitting pestle . After centrifugation , the pellet was directly subjected to transposition by Tn5 transposase for 30 min at 37°C using the Nextera DNA Library Preparation kit ( Illumina ) , and the tagmented DNA was then purified using a Zymo DNA Clean & Concentrator-5 kit , all according to a previously published protocol [69] . To generate a control naked DNA library , 1 ng of genomic DNA extracted with phenol/chloroform from six pooled wild-type brains was also subjected to transposition by Tn5 transposase and the tagmented DNA was purified following the same procedures . Barcoded libraries were PCR amplified , each using a common custom primer and a unique custom Nextera barcoded primer , as in [69] . For each library , after an initial round of five PCR cycles , the optimal number of PCR cycles needed to stop amplification prior to saturation was estimated by real-time quantitative PCR . Five to ten additional PCR cycles were then performed bringing the total number to 10 ( for control DNA ) and to 15 cycles ( for other samples ) , and the libraries were purified using a Qiagen MinElute PCR purification kit . Library quality and size distribution was assessed on a Bioanalyzer , libraries were quantified by real-time quantitative PCR and mixed in equimolar ratios before sequencing on a single lane of Hi-seq 2500 ( Illumina ) using 50bp single end reads . The resulting sequencing files were checked for quality control and demultiplexed by the Texas A&M AgriLife Genomics and Bioinformatics Facility . After removing the adapter sequence using fastx clipper ( hannonlab . cshl . edu/fastx_toolkit/ ) with options “-Q33 -n -v” , the clipped reads were mapped to the monarch genome ( v3; [80] ) using Bowtie2 [89] with parameters “—phred33—local” . Unmapped reads and mapped reads with mapping quality below 10 after sorting with SAMtools version 0 . 1 . 19 [90] were discarded . Mapped reads were adjusted such that those aligned to the positive strand and negative strand were shifted by +4 bp and –5 bp , respectively , as described in [69] . ATAC-seq peaks were called using MACS2 [91] with parameters “-q 0 . 01—nomodel—shift -100—extsize 200—keep-dup all” , using reads from naked DNA as control . For each genotype and time point , consensus peaks between biological replicates were generated by merging peaks with overlapping coordinates , using HOMER [92] . Pairwise comparison of the biological replicates was performed by quantifying the density of reads contained within consensus peaks in each replicate , using HOMER [92] . To compare replicates from all genotypes and time points to one another , peaks from all libraries were merged into a set of 37 , 642 consensus peaks . Read densities were quantified in each sample , and correlation was determined using the Pearson correlation coefficient [93] . Given that high level of reproducibility was found between biological replicates , ATAC-seq peaks from replicate libraries of wild-type , Clk knockouts and Cyc-like mutants at ZT04 and at ZT16 were merged by calling peaks on combined replicates using MACS2 [91] with the same parameters as described above . The peaks called , which ranged from 13 , 555 to 23 , 222 , were used in subsequent analysis . With the exception of a few peaks that could not be annotated because they were located in scaffolds containing no annotated genes , all peaks in the genome were assigned using HOMER [92] based on their locations relative to a gene as follows: ( 1 ) promoter-TSS if present within –1kb to +100 bp of the transcription start site ( TSS ) , ( 2 ) TTS if within -100 bp to +1kb of the transcription termination site ( TTS ) , ( 3 ) exon if within any exon , ( 4 ) intron if within any intron , or ( 5 ) intergenic . Peaks associated with genes rhythmically expressed in wild-type and differentially expressed in Clk knockouts were determined from these annotations . Differential peak analysis between conditions was performed on sets of overall merged genome-wide peaks and peaks associated with genes rhythmically expressed in wild-type and differentially expressed in Clk knockouts using HOMER [92] and its R package DESeq2 [66] wrapper with cutoff thresholds for FDR of < 0 . 05 and log2 fold-change of > 0 . 3785 and < -0 . 3785 . To visualize ATAC-seq peaks , bigWig files normalized to 10 million reads were generated from BAM files and visualized using the Integrative Genomics Viewer ( IGV; [94 , 95] ) . Sequencing data are available at https://www . ncbi . nlm . nih . gov/geo/query/acc . cgi ? acc=GSE122447 ( Accession number GSE122445 ) . Occupied TF DNA binding sites within ATAC-seq peaks associated with genes rhythmically expressed in wild-type and differentially expressed in Clk knockouts were identified in the brains of wild-type monarchs at ZT04 and at ZT16 using the Wellington TF footprinting algorithm in pyDNase [70] with p-value < 10−10 and FDR < 0 . 01 . This program uses the imbalance in the reads aligned to the positive and negative strands surrounding the protein-DNA interactions to accurately predict which binding sites are protected from Tn5 integrations . Footprints specific to either time point were identified using BEDTools [96] . Footprints from each group were further analyzed in wild-type , Clk knockouts and Cyc-like mutants at ZT04 and at ZT16 using the dnase_to_treeview . py script in pyDNase [70] to obtain Tn5 integration counts from the center of the footprints flanked by 100 bp on either side . Visualization was performed by generating heatmaps using Java Treeview [97] . To determine the effect of clock disruption on time-of-day dependent footprints , average profiles were generated from the average of Tn5 integration counts for each bp of the 200 bp regions . Data smoothing was applied using 3-bp rolling averages . To compare the depth of footprints found to be specific to ZT04 or specific to ZT16 in wild-type between wild-type , Clk knockouts and Cyc-like mutants at ZT04 and at ZT16 , a footprint occupancy score ( FOS ) was quantified using the formula ( C + 1 ) /L + ( C + 1 ) /R [98] . C corresponds to the average of Tn5 integrations within the central region of the footprints where a transcription factor is directly engaged , while L and R correspond to the average of Tn5 integrations respectively on the left and right flanking regions on each side of the footprints . Since Wellington footprinting algorithm identified footprints with 11–25 nucleotides in length and flanked by 35 nucleotides on each side [70] , the average of Tn5 cuts on both positive and negative strands corresponding to these lengths were used in the calculations . Data were represented as inverse FOS such that higher inverse FOS values indicate stronger footprints , i . e . a higher difference between the central and flanking regions of footprints . Inverse FOS values were visualized through box plots using ggplot2 [87] and the statistical analysis was performed using the base R function kruskal . test and the R package dunn . test [99] . Motif enrichment analysis was performed on TF footprints ( FPs ) located within peaks of genes rhythmically expressed in wild-type but differentially expressed in Clk knockouts that were specific to either ZT04 or ZT16 in wild-type monarch brains using findMotifsGenome script in HOMER [92] with parameters “-size given -mask” . For each motif , a fold-change > 1 . 5 was used as a measure of enrichment . Fold-change was expressed as the proportion of a given motif in FPs associated with rhythmically expressed genes over background , divided by its corresponding proportion in genome-wide FPs over background . Genome-wide FPs were identified using the combined 22 , 004 FPs at ZT04 and 14 , 101 FPs at ZT16 . Background sequences were randomly selected from a pool of sequences that ( 1 ) did not contain sequences in which FPs were found , ( 2 ) did not contain repeats sequences , and ( 3 ) whose length and GC-content matched those of input FPs . Sequence-biases and plant-specific motifs were filtered out from the analysis . Enriched known motifs were assigned to the corresponding FPs using HOMER’s findMotifsGenome script [92] with option “-find <motif matrix file>” . In cases where several TFs belonging to the same class ( i . e . , TFs with similar DNA binding domains ) were identified to occupy the same FP , only one was retained for further analysis . The percentage of motif classes and their distribution was calculated for FPs specific to ZT04 and to ZT16 in wild-type monarch brains . | With a rich biology that includes a clock-regulated migratory behavior and a circadian clock possessing mammalian clock orthologues , the monarch butterfly is an unconventional system with broad appeal to study circadian and seasonal rhythms . While clockwork mechanisms and rhythmic behavioral outputs have been studied in this species , the rhythmic genes that regulate rhythmic daily and seasonal activities remain largely unknown . Likewise , the mechanisms regulating rhythmic gene expression have not been explored in the monarch . Here , we applied genome-wide sequencing approaches to identify genes with rhythmic diurnal expression in the monarch brain , revealing the coordination of key pathways for brain function . We also identified the monarch brain open chromatin regions and provide evidence that regulation of rhythmic gene expression does not occur through temporal regulation of chromatin opening but rather by the time-of-day dependent binding of transcription factors in cis-regulatory elements . Together , our data extend our knowledge of the molecular rhythmic pathways , which may prove important in understanding the mechanisms underlying the daily and seasonal biology of the migratory monarch butterflies . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"carbohydrate",
"metabolism",
"invertebrates",
"gene",
"regulation",
"glucose",
"metabolism",
"animals",
"dna",
"transcription",
"animal",
"models",
"drosophila",
"melanogaster",
"model",
"organisms",
"experimental",
"organism",
"systems",
"chronobiology",
"molecular",
"biology",
"techniques",
"epigenetics",
"chromatin",
"drosophila",
"genetic",
"footprinting",
"research",
"and",
"analysis",
"methods",
"chromosome",
"biology",
"animal",
"studies",
"gene",
"expression",
"molecular",
"biology",
"insects",
"arthropoda",
"circadian",
"rhythms",
"genetic",
"fingerprinting",
"and",
"footprinting",
"biochemistry",
"eukaryota",
"cell",
"biology",
"genetics",
"biology",
"and",
"life",
"sciences",
"metabolism",
"organisms"
] | 2019 | Genome-wide discovery of the daily transcriptome, DNA regulatory elements and transcription factor occupancy in the monarch butterfly brain |
Oral cholera vaccines are primarily recommended by the World Health Organization for cholera control in endemic countries . However , the number of cholera vaccines currently produced is very limited and examples of OCV use in endemic countries , and especially in urban settings , are scarce . A vaccination campaign was organized by Médecins Sans Frontières and the Ministry of Health in a highly endemic area in the Democratic Republic of Congo . This study aims to describe the vaccine coverage achieved with this highly targeted vaccination campaign and the acceptability among the vaccinated communities . We performed a cross-sectional survey using random spatial sampling . The study population included individuals one year old and above , eligible for vaccination , and residing in the areas targeted for vaccination in the city of Kalemie . Data sources were household interviews with verification by vaccination card . In total 2 , 488 people were included in the survey . Overall , 81 . 9% ( 95%CI: 77 . 9–85 . 3 ) of the target population received at least one dose of vaccine . The vaccine coverage with two doses was 67 . 2% ( 95%CI: 61 . 9–72 . 0 ) among the target population . The vaccine coverage was higher during the first round ( 74 . 0 , 95%CI: 69 . 3–78 . 3 ) than during the second round of vaccination ( 69 . 1% , 95%CI: 63 . 9–74 . 0 ) . Vaccination coverage was lower in male adults . The main reason for non-vaccination was to be absent during the campaign . No severe adverse events were notified during the interviews . Cholera vaccination campaigns using highly targeted strategies are feasible in urban settings . High vaccination coverage can be obtained using door to door vaccination . However , alternative strategies should be considered to reach non-vaccinated populations like male adults and also in order to improve the efficiency of the interventions .
Cholera is endemic [1] in several countries in Africa and Asia [2 , 3] and can also cause large national wide epidemics , often involving several countries [4 , 5] . The World Health Organization ( WHO ) with financial support from GAVI , the Vaccine Alliance , has established an oral cholera vaccine ( OCV ) stockpile for emergency response conceived to respond to outbreaks and humanitarian crises where the risk of cholera is high . In addition , the Global Task Force for Cholera Control manages a non-emergency reserve that is expected to be used in endemic countries or in outbreak and humanitarian crises if the emergency stockpile is insufficient to cover the demand . However , the number of cholera vaccines currently produced is very limited and examples of OCV use in endemic countries , and especially in urban settings , are scarce . Kalemie , a city located in Tanganyika province in the Democratic Republic of Congo ( DRC ) , is considered one of the urban hotspot in the National Cholera Control Plan of DRC and it is mentioned a priority area to carry out control interventions in this country . Médecins Sans Frontières ( MSF ) has provided support to the Ministry of Health ( MoH ) in Kalemie in different cholera control activities using an integrated approached as it is recommended in the National Cholera Control Plan . These activities include: treatment of cholera cases , diseases surveillance , improvements in access to safe water and health promotion . The most ambitious part of this intervention was the building of a new water supply network to bring water to the health areas of Undugu and Kituku where the access to safe water was poor and cholera is highly endemic . In addition , MSF and the MoH planned to organize an OCV vaccination campaign in 2013 as additional control measures , to cover the gap during the time that the main component of the intervention was implemented . MSF requested a loan to the MSF requested a loan to the International Coordinating Group ( ICG ) , that manages the global stockpile of oral cholera vaccines , and 252 , 000 doses with a close expiration date ( 3 months for 42% of the doses provided and 9 months for the rest ) were approved for shipment to Kalemie . The vaccination campaign was planned to cover the four health areas with the highest historical attack rates , covering a population around 120 , 000 people . Both the MoH and the MSF feared that this highly targeted approach could generate tensions within the city , especially in neighborhoods not receiving the vaccine . Thus , the selected vaccination strategy was door to door , to avoid mass gathering , and the social mobilization campaign was light with message passed only within the targeted communities through community health workers without using mass media . The vaccination campaign started in November 25; however , the three days later , the team was evacuated for a complex security incident unrelated to the campaign . One month later ( end of December 2013 ) the vaccination team was allowed to come back to Kalemie , however 41 . 6% of the initial doses achieved the expiration date and were discarded . The MSF team re-discussed with the MoH the vaccination strategy and the target population was reduced to two health areas ( Undugu and Kataki ) , with the highest historical incidence rates , covering a population around 50 , 000 people . The initial vaccination strategy was maintained and the campaign started on July 1 , 2014 . Two doses , 14 days apart , were provided to all individuals one year and above excluding women who orally reported to be pregnant due to the concerns from local authorities and the lack of safety data for this group at the time . Here we describe the vaccine coverage achieved with this highly targeted vaccination campaign and the acceptability of the campaign in Kalemie .
The city of Kalémie is located in Katanga , in Eastern Health District of Tanganyika in the DRC , near the lake of the same name . The Lukuga River separates the city into southern districts ( eight health areas ) and northern districts ( three health areas ) ( Fig 1 ) . The population of Kalémie is estimated at 262 , 963 people . Cholera cases are reported throughout the year in Kalémie with two seasonal peaks , one in September-October and another in January . All cases of cholera are treated in the diarrheal diseases treatment center of the General Referral Hospital of Kalémie . The health areas reporting the highest attack rates are located along the Lukuga River and Lake Tanganyika ( Kituku Undugu in Nyemba; HGR and Kataki ) ( Fig 1 ) . We carried out a cross-sectional survey in the areas included as part of the target population of the 2014 OCV campaign ( strata 1 ) in Kalemie including Undugu , Kataki and three neighborhoods of Kituku ( Tanganika , Quartier Industriel and Central-Singa ) . We also assess the vaccination coverage of the three days of vaccine distribution in November , 2013 in HGR and the neighborhoods of Kituku not included in 2014 ( strata 2 ) . All individuals residing in the health areas targeted for vaccination were eligible for inclusion in the survey ( Fig 1 ) . Residents were defined as persons living ( sleeping and eating ) in the area since the starting date of the vaccination campaign . Households were selected using spatial random sampling [6] . A household was defined as a group of people sleeping under the same roof and sharing meals every day for at least the previous two weeks to the starting date of the vaccination campaign . The sample size was calculated to obtain a representative estimate of the proportion of residents who received two doses of OCV by age group ( 1–4 , 5–14 , 15 years and older ) in the vaccinated areas in 2014 ( strata 1 ) . Sample size was calculated to ensure a sufficiently precise estimate for children aged 1 to 4 years as this group was the smallest . We considered the following assumptions: 85% of children would receive two doses of vaccine , alpha error of 5% , and absolute precision of 5% . Taking into account a census conducted prior the campaign in October 2013 , we expected 0 . 9 children 1–4 year old per household ( average of 6 . 2 individuals per household and 14% of the population between 1 and 4 years ) . Assuming 10% of missing data , we planned to visit 202 households and include all household members . For the non-vaccinated areas in 2014 that were initially targeted in 2013 ( strata 2 ) it was difficult to forecast the coverage , however we expected a two dose coverage lower than 15% in this population , thus the same sample size was considered in this second strata . Teams conducted face-to-face directed interviews after consent . A standardized pre-piloted questionnaire was used to collect the following information: demographic data ( age , sex , and household size ) , vaccination status ( card-confirmed and orally reported ) , reasons for non-vaccination ( open question ) , and acceptability data ( adverse events , taste and beliefs about the vaccine ) . Questions concerning acceptability were only collected in participants older than 15 years . Interviews were conducted in the local language , Swahili . Survey teams asked for the help of neighbors to trace absentees and re-visit empty ( but not abandoned ) households later in the day . If during the second visit the occupants could not be found or if they refused to participate , that household was skipped . Our main outcome was the OCV coverage ( full course and at least one dose ) in the target population . Vaccine coverage was calculated dividing the number of individuals reporting being vaccinated by the survey population and expressed as a percentage . Vaccination coverage estimates include both card-confirmed and oral reporting . Secondary outcomes included vaccine coverage by age group , sex , and reasons for non-vaccination . We also estimated the geographical distribution of the vaccine coverage using spline functions [7] . Crude vaccination coverage estimates and 95% confidence intervals ( 95% CI ) were obtained considering the survey design . The design effect ( deff ) was calculated to estimate the loss of precision due to the selection of all member of the household . The geographical distribution of the vaccine coverage was estimated using a binomial regression in a general additive model framework . The dependent variable was the vaccination status of each individual , and the location of the household was included as a smoothing spline term . The smoothing parameter of the spline term was chosen by cross validation . We plot the vaccine coverage and the standard deviation of the term as an indicator of the uncertainty in the estimates . The reasons for non-vaccination were entered in French by data entry clerks speaking fluently French and Swahili . Key words associated with different categories of answer were identified . An automatic detection of those key words allowed to categorize the answers . Data entry was performed using EpiData 3 . 1 ( EpiData Association , Denmark ) and data analysis was performed using Stata 13 . 0 ( College Station , USA ) and R Statistical Software ( The R Foundation for Statistical Computing ) . The Ethical Review Board of the University of Lubumbashi approved the study protocol ( study protocol ethical number: UNILU/CEM/028/2013 ) . Oral informed consent was obtained from participants in all instances . All children had consent given from a parent/guardian and all adult participants provided their own consent . Oral informed consent was requested since the study did not present any risk of harm to subjects and did not involve procedures for which written consent is normally required outside the research context . The request of consent was registered in a log-book . Privacy and confidentiality of the data collected from participants were ensured both during and after the conduct of the surveys .
Overall , 398 households were visited between August 21 and 29 , 2014 and 2 , 465 individuals were included as part of the survey . Among these individuals 48 were excluded in the analysis; 8 were children less than 1 year of age and 40 were pregnant women . Finally , 2 , 417 were included; 1 , 160 in the first strata ( areas targeted in the 2014 campaign ) and 1 , 257 in second strata ( initially included in the 2013 campaign but not targeted in 2014 ) . The average household size was 6 . 4 ( standard deviation: 3 . 2 ) . The median age was 15 year old ( inter quantile range ( IQR ) : 7–28 years ) . The male:female ratio was 0 . 92 ( 95%CI: 0 . 84–0 . 99 ) . Vaccination coverage varied within the areas targeted in these campaigns . In the three areas covered by the campaign in 2014 ( Undugu , Katataki and the three neighborhoods of Kituku ) we observed areas with coverage of two doses close to 100% and areas with coverage as low as 30% ( Fig 3 ) . The uncertainty about the estimation was relatively low as shown in Fig 3 . The coverage in strata 2 was as well heterogeneous with a high percentage of individuals vaccinated with at least one dose in HGR ( Fig 3 ) . The main reason for non-vaccination was obtained for all 1301 and 358 people who had not received vaccine during the first and second round of vaccination in 2014 , respectively , and for 368 people who had received zero ( n = 204 ) or one dose ( n = 164 ) during the first two rounds in 2014 and had not received the dose during the mop-up campaign ( Table 1 ) . The main reason for non-vaccination was to be absent during the campaign , representing the main reason for more than 2/3 of the individuals . Between 5% and 7% of non-vaccinated refused the vaccine , either for religious reasons or tradition , or because they considered that the vaccine was unsafe . The proportion of persons who have not been vaccinated in 2014 for lack of information was very low ( Table 1 ) . During the vaccine coverage survey , a part of the interview focused on the adverse event following immunization ( first and second round of 2014 ) . Among those who received the vaccine , 4 . 3% ( n = 39 ) reported having been ill after the first round of vaccination and 1 . 4% ( n = 13 ) after the second round . The main signs or symptoms reported over the two rounds were diarrhea ( 22 . 4% ) , fever ( 31 . 0% ) , stomachache ( 15 . 5% ) nausea and vomiting ( 12 . 0% ) . Among those who said they experienced side effects , 36 . 4% ( n = 12/33 ) reported consulting after the first round , 30 . 8% ( n = 4/13 ) after the second round . Among the 367 participants aged at least 15 years included in the survey in the strata 1 ( targeted areas in the 2014 campaign ) , 334 reported being aware of the vaccination campaign in 2014 ( 91 . 0% , 95% CI: 87 . 0 to 93 . 9 ) . Around half of them thought that the vaccine had a good effect on their health status ( 51 . 0% , 95% CI: 43 . 4 to 58 . 5 ) , while a quarter ( 26 . 4% , 95% CI: 20 . 6 to 33 . 2 ) felt that the vaccine had no effect and 8 . 2% ( 95% CI 4 . 9 to 13 . 3 ) a bad effect . In addition , 15 . 5% ( 95% CI: 10 . 2 to 23 . 0 ) thought that the vaccine could cause illness . Most of those participants vaccinated reported that the vaccine had a bad taste ( 85 . 1% , 95% CI: 79 . 7 to 89 . 3 ) . However , the vast majority of respondents said they would go to get vaccinated again if a new vaccination campaign was organized in the future ( 93 . 4% , 95% CI: 88 . 9 to 96 . 2 ) .
This study shows that highly targeted mass vaccination campaigns in urban settings are feasible and that high vaccine coverage can be reached within the areas selected for vaccination . The campaign was organizing without major logistic difficulties and any civil unrest . Most of the doses were distributed in the predefined target areas for vaccination; however , some spillover of vaccine in the surrounding areas was documented . This issue has been already reported even in rural settings , where it is easier to target specific villages [8] . This is one of the first times that OCV has been used in urban setting outside a vaccine trial . The only other example of highly targeted vaccination in an urban African setting was a demonstration project carried out in Beira , Mozambique , where a lower coverage was achieved [9] . In that case , the strategy proved as well to be feasible and well accepted by the local community [9] . Feasibility and acceptability of this highly targeted approach in urban settings is similar to what has been described for rural areas or refugee camp settings [10 , 11] . Interestingly , the data from the coverage survey did not correlate well with the administrative coverage . The administrative coverage was close to 100% since all the vaccines were use and the number of vaccine doses available was close to the target population for the 2014 campaign . This means that some of the vaccines that should have been delivered in the strata 1 were provided to people not living in that area or who were not initially part of the estimated target population . This could be explained as a combination of two different issues: ( i ) individuals living in bordering areas came to be vaccinated into the strata 1 , ( ii ) there was an increase in the target population between the census in October 2013 and the implementation of the camping in July 2014 . We have indications that probably the two processes occurred in Kalemie . The coverage survey revealed that some people were vaccinated in strata 2 during the 2014 campaign even in they were not part of the target population . In addition the administrative coverage was over 100% in children under five which could be the result of children from bordering areas coming to get vaccine . This could be partially also the result of a high fertility rate in those communities that end up increasing the denominator among children 1 to 5 years old . This hypothesis is supported by a slight reduction in the age average in the survey sample compared with the census . Additionally , the survey showed a slightly higher average household size than the census , which supports as well the increased size of the target population linked with the high fertility rate . All the doses available were distributed in 2014 but the data shows that especially male adults were vaccinated at lower rates that the rest of the population . A small mop-up activity was needed to complete this distribution of the doses , but was clearly insufficient to reach all the unvaccinated individuals . One of the reasons that might explain the lower coverage among adults is the vaccination strategy itself . The door to door distribution benefited the intake among those who tend to spend more time at home . In addition , the social mobilization activities was implemented with a low profile because of the fear of running out of vaccine doses if the people from surrounding areas would have come massively to get vaccination . These two issues highlight the need to find innovative strategies to reach population , like the male adults , that systematically show lower vaccine coverage in cholera vaccination campaigns [8 , 12–14] . The lesser coverage of adult men could be due to lower availability because of work during daytime or because of the perception that the vaccine is mainly targeting children as it has already been suggested in other vaccination campaigns [8 , 15] . A possible solution might be to change the unit for vaccine distribution from the individual to the household , having a senior adult responsible of delivering the vaccine to the entire family such as the self-administration strategy adopted to distribute the second dose to fishermen in Malawi [15] . This could reduce the cost and increase the efficiency of the campaigns . Alternative vaccination strategies should be design taking into account three characteristics of the vaccine that can increase the field effectiveness: the efficacy of the first dose [16] , the high thermal-stability of the lipopolysaccharide ( main antigen of Shanchol ) [17] , and the high indirect protection provided when high vaccine coverage is achieved [18–20] . Related to this last point , an element to consider when designing future campaigns should be the geographical heterogeneity , which has been previously described in other campaigns [12] . A large part of the potential of cholera vaccination campaigns to reduce cases and deaths is expected to be obtained through to the indirect protection induced by this vaccine [21 , 22] . Therefore , the existence of pockets of non-vaccinated individuals represents a risk of cholera spread in cholera prone areas . Real-time information on the distribution of the vaccination coverage could help to reduce these geographical disparities , which represents one of the major risks than could reduce the impact of oral cholera vaccines . In addition , in absence of enough vaccine to cover the desire vaccine coverage , as it was the case in Kalemie , modelling work suggest that might be more beneficial in the short term to provide one dose to a larger population than two doses to a smaller number of individuals [22] . The major limitation in our evaluation was the high percentage of vaccinated individuals with missing vaccination cards despite the short interval between the vaccination camping and the coverage survey . We think that this short delay limits the risk of having a large information bias in the vaccination status; however , it highlights the difficulty to be certain about the vaccination status , which might be the major limitation to understand the effectiveness and impact of vaccination campaigns in the medium-long term . The sample size that we expected for each stratum was achieved , and this allowed having adequate precision in the estimates , despite the slightly lower vaccine coverage than the one assumed in the sample size calculation . In strata 2 , where we had a higher uncertainty about coverage , the two dose coverage was lower than the 15% assumed in the sample size calculation and thus we had a good overall precision in our estimates . An additional limitation is the use the vaccine coverage and other quantitative measure to evaluate the acceptability of the campaign . Qualitative assessment could enrich the knowledge that we have about the knowledge , perceptions and barriers that limits the performance of the vaccination activities . Qualitative measurements are needed to understand the best approach to vaccinate mobile communities like the fishermen in Kataki area at it has been shown in refugee camps settings like South Sudan [23] . Despite the robust sampling method , the potential lower availability of adult men during daytime , as reflected by our sex ratio slightly but significantly lower than 1 , could have resulted in a slight overestimate of the global vaccine coverage , since the coverage appeared significantly lower in this group . However a sex ratio lower than 1 could also simply reflect a higher mortality rate of adult male , as is frequent in African settings . In conclusion , the vaccination campaign in Kalemie was implemented using a highly targeted vaccination strategy without major logistical difficulties . The vaccination coverage obtained was relatively high , but lower than expected , especially among the male adults . The relative low coverage among male adults highlights the need to find better vaccination strategies and to consider alternative methods for vaccine distribution for specific subpopulations . These alternative vaccination protocols should take full advantage of the relative high vaccine efficacy of the first vaccine dose and the good thermal-stability of the vaccine . | The oral cholera vaccine , Shanchol , has already been shown as an effective tool in controlling a cholera outbreak . The limited amount of doses , concurrently with the logistic constraints associated with a targeted vaccination campaign are serious difficulties to tackle in order to organize a vaccination campaign in an urban setting . Although the World Health Organization recommends its use for cholera control in endemic countries , the fact remains that the use of the oral cholera vaccine in endemic setting has scarcely been described , especially in an urban setting , until now . Médecins Sans Frontières and the Ministry of Health from Democratic Republic of Congo organized a vaccination campaign of a limited part of the urbanized and highly endemic city of Kalemie , in the Tanganyika Province using a door to door strategy . The vaccine coverage in the targeted zones was high and demonstrated the feasibility of cholera vaccination campaign in this setting but also the need for creative strategies in order to reach population remaining hard to vaccine . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"and",
"health",
"sciences",
"clinical",
"research",
"design",
"population",
"dynamics",
"immunology",
"tropical",
"diseases",
"census",
"vaccines",
"preventive",
"medicine",
"bacterial",
"diseases",
"research",
"design",
"global",
"health",
"neglected",
"tropical",
"diseases",
"population",
"biology",
"infectious",
"disease",
"control",
"vaccination",
"and",
"immunization",
"research",
"and",
"analysis",
"methods",
"public",
"and",
"occupational",
"health",
"infectious",
"diseases",
"cholera",
"cholera",
"vaccines",
"adverse",
"events",
"survey",
"research",
"biology",
"and",
"life",
"sciences",
"geographic",
"distribution"
] | 2018 | Highly targeted cholera vaccination campaigns in urban setting are feasible: The experience in Kalemie, Democratic Republic of Congo |
Candida albicans is the most common cause of hematogenously disseminated and oropharyngeal candidiasis . Both of these diseases are characterized by fungal invasion of host cells . Previously , we have found that C . albicans hyphae invade endothelial cells and oral epithelial cells in vitro by inducing their own endocytosis . Therefore , we set out to identify the fungal surface protein and host cell receptors that mediate this process . We found that the C . albicans Als3 is required for the organism to be endocytosed by human umbilical vein endothelial cells and two different human oral epithelial lines . Affinity purification experiments with wild-type and an als3Δ/als3Δ mutant strain of C . albicans demonstrated that Als3 was required for C . albicans to bind to multiple host cell surface proteins , including N-cadherin on endothelial cells and E-cadherin on oral epithelial cells . Furthermore , latex beads coated with the recombinant N-terminal portion of Als3 were endocytosed by Chinese hamster ovary cells expressing human N-cadherin or E-cadherin , whereas control beads coated with bovine serum albumin were not . Molecular modeling of the interactions of the N-terminal region of Als3 with the ectodomains of N-cadherin and E-cadherin indicated that the binding parameters of Als3 to either cadherin are similar to those of cadherin–cadherin binding . Therefore , Als3 is a fungal invasin that mimics host cell cadherins and induces endocytosis by binding to N-cadherin on endothelial cells and E-cadherin on oral epithelial cells . These results uncover the first known fungal invasin and provide evidence that C . albicans Als3 is a molecular mimic of human cadherins .
The fungus Candida albicans causes both hematogenously disseminated and oropharyngeal disease . During the initiation of hematogenously disseminated candidiasis , blood-borne organisms invade the endothelial cell lining of the vasculature to infect deeper tissues . Fungal invasion of the superficial oral epithelial cells is also characteristic of oropharyngeal candidiasis [1–3] . The invasion of either endothelial or oral epithelial cells by C . albicans is necessary for the organism to damage these cell types in vitro [4 , 5] . Host cell invasion and damage are likely critical virulence attributes of C . albicans because mutants with defects in these processes in vitro are highly likely to have attenuated virulence in experimental animal models of hematogenously disseminated or oropharyngeal candidiasis [5–8] . Because of the importance of host cell invasion in the pathogenesis of candidiasis , we have been investigating the mechanism by which this process occurs . Previously , we have found that C . albicans invades both endothelial cells and oral epithelial cells in vitro by inducing its own endocytosis [4 , 5 , 7] . Endothelial cell endocytosis of C . albicans is induced when the organism binds to N-cadherin and other endothelial cell surface proteins . This binding induces microfilament rearrangement , which results in the formation of pseudopods that engulf the organism and draw it into the cell [9 , 10] . Endothelial cell endocytosis of C . albicans is dependent on extracellular calcium and is governed at least in part by the tyrosine phosphorylation of endothelial cell proteins [9 , 11] . C . albicans is a dimorphic fungus that can grow as either ovoid-shaped blastospores or as filamentous hyphae . The ability to reversibly change between blastospores and hyphae is a key virulence factor for this fungus , and mutants that are unable to form hyphae have greatly attenuated virulence [12 , 13] . We have found previously that C . albicans mutants with defects in the signal transduction pathways that govern hyphal formation have a significantly reduced capacity to invade both endothelial and oral epithelial cells in vitro [5–7] . Furthermore , studies with killed organisms indicate that induction of endocytosis is passive on the part of the organism because killed hyphae are endocytosed as avidly as are live hyphae [5 , 7] . An important , yet-unanswered question from these studies has been the identity of the C . albicans surface structure ( s ) that induces host cell endocytosis . Numerous other medically important fungi such as Aspergillus fumigatus , Cryptococcus neoformans , and Rhizopus oryzae invade normally nonphagocytic host cells [14–18] . Remarkably , no fungal surface protein that induces fungal uptake by these host cells has been identified . Discovering such an invasin-like protein is important not only for its biological significance , but also because it could serve as a therapeutic or vaccine target . In bacterial pathogens , proteins that mediate host cell invasion frequently also function as adhesins [19] . In C . albicans , a major group of adhesins consists of the proteins encoded by the ALS gene family . This gene family encodes eight GPI-linked cell surface proteins that mediate binding to diverse host substrates [20–25] . Each Als protein has three domains . The N-terminal domain contains the substrate-binding region [23 , 26] . The central portion of the protein consists of a variable number of 36–amino acid tandem repeats . The C-terminal domain is rich in serine and threonine , and contains a GPI anchorage sequence that is predicted to be cleaved as the protein is exported to the cell surface [20] . Computer-assisted modeling of the N-termini of Als1 , Als3 , and Als5 predicts the presence of antiparallel β sheets that define the immunoglobulin superfamily [23] . This prediction has been substantiated by circular dichroism and Fourier transform infrared spectroscopic analysis of the N-terminus of Als1 and Als5 [23 , 27] . Interestingly , the predicted 3-D structures of the N-termini of many Als proteins are remarkably similar to the Yersinia pseudotuberculosis invasin , a protein that induces the endocytosis of Y . pseudotuberculosis by mammalian epithelial cells [28 , 29] . This relationship led us to investigate whether any of the Als proteins could induce host cell endocytosis . We found that Saccharomyces cerevisiae expressing Als3 were avidly endocytosed by endothelial cells . Similar experiments suggested that Als1 and Als5 could also induce endocytosis , although with considerably less efficiency than Als3 [23] . In the current study , we investigated the role of Als1 and Als3 in inducing the endocytosis of C . albicans by both endothelial and oral epithelial cells . We discovered that Als3 , and to a lesser extent Als1 , induce endocytosis of C . albicans by binding to N-cadherin on endothelial cells and E-cadherin on oral epithelial cells . Furthermore , computer-assisted molecular modeling of the binding of Als3 with N- and E-cadherin suggests that Als3 is a structural and functional mimic of these cadherins .
The interactions of als1Δ/als1Δ and als3Δ/als3Δ mutant strains of C . albicans with human umbilical vein endothelial cells and the FaDu oral epithelial cell line were investigated . Using our standard differential fluorescence assay , we determined the capacity of each strain of C . albicans to adhere to these host cells and induce its own endocytosis . Hyphae of the als1Δ/als1Δ null mutant interacted with both host cell types similarly to the wild-type strain ( Figure 1A and 1B ) . In contrast , there was a 90% reduction in the number of hyphae of the als3Δ/als3Δ mutant that were endocytosed by either endothelial or oral epithelial cells compared to the wild-type control strain . Also , there was a significant reduction in the number of hyphae of the als3Δ/als3Δ mutant that were cell-associated with endothelial cells , but not oral epithelial cells . These results suggest that Als3 contributes to adherence to endothelial cells . We verified that the reduced endocytosis and adherence of the als3Δ/als3Δ mutant was due to the absence of ALS3 , because complementing this mutant with a wild-type copy of ALS3 restored endocytosis and adherence to wild-type levels . Endocytosis of C . albicans by endothelial cells is induced when the organism binds to N-cadherin on the endothelial cell surface [9] . We used an affinity purification approach to determine whether Als1 or Als3 was required for C . albicans hyphae to bind to N-cadherin in endothelial cell membrane extracts . As predicted by the results of the endocytosis assay , the als1Δ/als1Δ mutant bound to the same endothelial cell surface proteins , including N-cadherin , as did the wild-type strain ( Figure 1C ) . In contrast , the als3Δ/als3Δ null mutant did not bind to N-cadherin , and it bound very weakly to the other endothelial cell surface proteins . This defect in binding to N-cadherin was reversed when the als3Δ/als3Δ null mutant was complemented with a wild-type copy of ALS3 . Therefore , Als3 is required for C . albicans to bind to N-cadherin and other proteins on the endothelial cell surface . The surface proteins on oral epithelial cells that are bound by C . albicans hyphae have not been identified previously . We found that the wild-type strain of C . albicans bound to at least four different proteins on the surface of FaDu oral epithelial cells ( Figure 1D ) . Many of these proteins appeared to be different from the endothelial cell proteins that were bound by wild-type C . albicans . FaDu oral epithelial cells express very low levels of N-cadherin ( unpublished data ) , but high levels of E-cadherin . Therefore , we investigated whether E-cadherin was one of the epithelial cell proteins that was bound by C . albicans . Using an anti-E-cadherin monoclonal antibody , we detected a significant amount of this protein in immunoblots of epithelial cell membrane proteins that were bound by hyphae of wild-type C . albicans ( Figure 1D ) . Furthermore , although the als1Δ/als1Δ mutant still bound to E-cadherin , the als3Δ/als3Δ null mutant did not . To verify that these results were not specific to the FaDu oral epithelial cell line , we repeated these experiments using the OKF6/TERT-2 cell line . This cell line was developed by the forced expression of the human telomerase catalytic subunit in oral mucosal keratinocytes obtained from a healthy adult [30] . The als3Δ/als3Δ null mutant was endocytosed poorly by OKF6/TERT-2 cells and bound poorly to the surface proteins , including E-cadherin , of this cell line ( Figure S1 ) . In contrast , wild-type C . albicans bound to similar surface proteins of both OKF6/TERT-2 and FaDu epithelial cells . Collectively , these results indicate that C . albicans hyphae bind to epithelial cell E-cadherin in an Als3-dependent manner . When C . albicans interacts with the oral epithelium in vivo , the organisms are coated with saliva , which has been reported to either inhibit or enhance the adherence of C . albicans to oral epithelial cells [31 , 32] . Also , it is possible that salivary proteins could potentially act as bridging molecules between the hyphae and oral epithelial cells and thereby facilitate endocytosis of C . albicans . To investigate these possibilities , we incubated C . albicans hyphae in the presence and absence of 20% saliva and then measured their interactions with both intact FaDu oral epithelial cells and oral epithelial cell surface proteins . Killed organisms were used in these experiments to obviate potential effects of saliva on the growth of the hyphae . We have previously shown that killed hyphae adhere to and are endocytosed by oral epithelial cells similarly to live hyphae [5] . Incubating hyphae in saliva prior to adding them to the epithelial cells had no effect on the number of endocytosed or cell-associated organisms ( Figure S2A ) . As predicted by the endocytosis assay , saliva also had no detectable effect on the binding of epithelial cell surface proteins , including E-cadherin , to C . albicans hyphae ( Figure S2B and S2C ) . Therefore , salivary components do not act as bridging molecules between the organisms and epithelial cells . Next , we confirmed that oral isolates of C . albicans other than those derived from the oral strain SC5314 also bound to E-cadherin and were endocytosed by FaDu oral epithelial cells . Two additional strains that had been obtained from patients with AIDS who had oropharyngeal candidiasis were tested in our assays . Hyphae of these strains were endocytosed similarly by the oral epithelial cells , although strain 7392 had slightly higher adherence to these cells ( Figure 2A ) . Also , all strains bound to the same oral epithelial cell surface proteins , including E-cadherin ( Figure 2B and 2C ) . To verify that that these strains expressed Als3 on their surface , we used flow cytometry with an Als3-specific polyclonal antiserum . As expected , all wild-type strains had detectable Als3 on their surface , whereas the als3Δ/als3Δ null mutant did not ( Figure 2D ) . Next , we verified that binding of C . albicans hyphae to N-cadherin and E-cadherin on the surface of intact host cells required Als3 and was associated with induction of endocytosis . Monolayers of endothelial cells or FaDu oral epithelial cells were infected with the various strains of C . albicans , after which N-cadherin and E-cadherin were detected by indirect immunofluorescence . To determine if the organisms were in the process of being endocytosed , the host cells were also stained for f-actin , which accumulates around such organisms [4 , 5] . We observed that endothelial cell N-cadherin colocalized with hyphae of the wild-type strain and the als1Δ/als1Δ null mutant that were being endocytosed ( Figure 3 ) . In contrast , N-cadherin did not colocalize with hyphae of the als3Δ/als3Δ null mutant , and almost none of these hyphae were endocytosed . As expected , N-cadherin accumulated around hyphae of the als3Δ/als3Δ::ALS3 complemented strain similarly to the wild-type control strain . A comparable pattern was observed with E-cadherin and oral epithelial cells ( Figure 4 ) . E-cadherin colocalized with the endocytosed hyphae of all strains except for the als3Δ/als3Δ null mutant . Also , very few hyphae of the als3Δ/als3Δ mutant were endocytosed . The anti-N-cadherin and anti-E-cadherin monoclonal antibodies did not bind to C . albicans hyphae in the absence of endothelial or epithelial cells ( Figure S3 ) , indicating that these antibodies did not recognize any cross-reacting C . albicans antigens . Collectively , these findings support a model in which C . albicans Als3 binds to endothelial cell N-cadherin and oral epithelial cell E-cadherin , thereby stimulating actin-mediated endocytosis of the organism . Infection of endothelial cells or oral epithelial cells with wild-type C . albicans hyphae in vitro causes significant damage to these cells , and endocytosis of the organisms is required to induce host cell damage [4 , 5 , 10] . Therefore , we investigated the extent of damage to endothelial cells and oral epithelial cells caused by the als1Δ/als1Δ and als3Δ/als3Δ null mutants . Consistent with the results of the endocytosis assay , the als1Δ/als1Δ null mutant caused the same amount of damage to both types of host cell as the wild-type strain ( Figure 5 ) . In contrast , the als3Δ/als3Δ null mutant caused essentially no damage to either cell type . Complementing the als3Δ/als3Δ null mutant with a wild-type copy of ALS3 restored its capacity to damage these host cells . These data indicate that the inability of the als3Δ/als3Δ null mutant to invade endothelial cells and oral epithelial cells is associated with a significantly reduced capacity to damage these cells . To determine if either Als1 or Als3 alone was able to induce endocytosis , we coated latex beads with the purified , recombinant N-terminal portion of Als1 ( rAls1-N ) or Als3 ( rAls3-N ) , which consisted of amino acids 17 to 432 of either protein . Control beads were coated with biotinylated bovine serum albumin ( BSA ) . We then assessed the number of beads that were endocytosed by and associated with endothelial cells and FaDu oral epithelial cells . Although rAls1-N–coated beads were endocytosed similarly to the BSA-coated beads by endothelial cells , significantly more of these beads were endocytosed by FaDu epithelial cells ( Figure 6A and 6B ) . The number of rAls1-N–coated beads that were associated with both cell types was similar to that of the BSA-coated beads , indicating that the N-terminal portion of Als1 by itself is insufficient to mediate adherence to either cell type . Beads coated with rAls3-N were endocytosed avidly by both endothelial and oral epithelial cells . Also , a greater number of beads coated with Als3 were associated with the endothelial cells compared with the control beads . Next , we examined the interactions of the beads with Chinese hamster ovary ( CHO ) cells expressing N-cadherin or E-cadherin to investigate whether binding of Als1 or Als3 to these cadherins alone was sufficient to induce endocytosis . The CHO cell assay was more sensitive than the assays using endothelial cells or oral epithelial cells because the background adherence and endocytosis of the BSA-coated beads was lower when CHO cells were used ( unpublished data ) . Beads coated with rAls1-N were endocytosed more efficiently than the control beads by CHO cells expressing either N-cadherin or E-cadherin ( Figure 6C and 6D ) . Also , slightly more rAls1-N–coated beads were associated with the CHO cells expressing N-cadherin , but not the ones expressing E-cadherin . However , the beads coated with rAls3-N were endocytosed significantly more efficiently by the CHO cells expressing N- or E-cadherin than were the beads coated with either BSA or rAls1-N . In addition , a slightly greater number of rAls3-N–coated beads were associated with cells expressing either N-cadherin or E-cadherin than were the control beads . The enhanced endocytosis mediated by rAls3-N required the presence of either N- or E-cadherin on the surface of the CHO cells because beads coated with rAls3-N were endocytosed similarly to beads coated with BSA by control CHO cells that did not express either N- or E-cadherin ( Figure 6E ) . These results demonstrate that Als3 , and to a lesser extent Als1 , bind directly to N-cadherin and E-cadherin , and that this binding is sufficient to induce endocytosis . These data also indicate that the cadherin-binding domains of Als1 and Als3 are located in the N-termini of these proteins . Although the N-terminal regions of Als1 and Als3 are 84% identical at the amino acid level , they interacted differently with N- and E-cadherin . Therefore , we used computer-assisted modeling to identify differences in the secondary structure of the N-terminal regions of Als1 and Als3 that might account for their functional differences . Our initial models of the N-terminal regions of these proteins indicated that they share general structural features , including a modular β sheet configuration interposed by extended regions , and an overall negative surface potential [23] . In the current study , we generated higher resolution models of the N-terminal regions of Als1 and Als3 . These models predicted that the N-termini of both proteins contain recurrent β barrel ( “beads-on-a-string” ) domains ( Figure 7A ) . Interestingly , the N-terminal regions of Als1 and Als3 bear a striking resemblance to the N-terminal ectodomains of N- and E-cadherin . The cadherin ectodomains have five β barrel domains , the first two of which are shown in Figure 7A . Similarly , the N-terminal regions of Als1 and Als3 contain at least four β barrel domains ( the first two of these domains are shown in Figure 7A , and all four domains are shown in Figure 8 ) . Notably , the first such domain ( N1 domain ) of each of the four proteins contains their most hydrophobic sections ( Figure 7B ) . However , Als1 and Als3 differ from the cadherins in that the fungal proteins bear less negative charge on their N1 than their N2 domains , whereas the cadherins bear the most negative charge on their N1 domains ( Figure 7C ) . The present computational models of the Als1 and Als3 N-terminal regions also indicated potentially significant differences between these two proteins . The most conspicuous differences are the predicted size and shape of these regions . Both Monte Carlo and molecular dynamics simulations indicate that the N-terminal domains of Als1 favor an extended conformation , while those of Als3 tend toward a more compact , β propeller–like arrangement ( Figure 8 ) . Also , the spatial distribution of surface hydrophobicity and electrostatic potential within the N1 and N2 domains differ between Als1 and Als3 ( Figure 7B and 7C ) . Next , we modeled potential sites of interaction between the N-terminal regions of Als1 , Als3 , N-cadherin , and E-cadherin . The N-terminal domains of Als1 and Als3 were modeled in common , semi-extended conformations ( Figure 8 ) . This consensus conformation allowed comparable access of any potential Als-docking domains to the cadherin ligands . Significant differences between Als1 and Als3 docking interactions with N- and E-cadherin were found . The binding of Als1 to N-cadherin was predicted to have greater contact area and electrostatic energy compared to the binding of Als3 to N-cadherin ( Table 1 ) . The hydrophobicity and hex energy of the binding of either Als1 or Als3 to N-cadherin were similar . However , the overall favorability of binding to N-cadherin was much better for Als3 than for Als1 . Interestingly , the predicted parameters of Als3 binding to N-cadherin were very similar to those of one molecule of N-cadherin binding to another ( N-cadherin self-association ) . The docking studies also predicted that different regions of Als1 and Als3 bind to N-cadherin . The N1 and N3 domains of Als1 bind to domains N1 and N2 of N-cadherin , whereas the N2 and N3 domains of Als3 bind to domains N1 and N2 ( Table 1 and Figure 8A ) . The N-terminal regions of Als1 and Als3 are also predicted to interact differently with E-cadherin . The energetics of Als1 binding to E-cadherin are considerably weaker than either Als3 binding to E-cadherin or E-cadherin binding to itself ( Table 1 ) . Also , Als1 N1 and N3 domains likely interact only with the N1 domain of E-cadherin , and they do not appear to have energetically favorable interactions with the N2 domain of this cadherin ( Table 1 and Figure 8 ) . Als3 is predicted to interact with E-cadherin differently than Als1 . Moreover , the Als3-E-cadherin interaction is somewhat different from the Als3–N-cadherin interaction . Domains N1 , N2 , and N3 of Als3 appear to form a cleft that binds E-cadherin domains N1 and N2 ( Figure 8 ) . More important , the geometry of this interaction assumes a parallel orientation , such as that which occurs when E-cadherin binds to itself ( Figure 8A ) .
Previous studies have focused on the adherence properties of the Als proteins [23–25 , 33–38] . The different Als proteins mediate adherence to broad variety of host substrates , and this adherence is likely critical for C . albicans to infect host surfaces . In this study , we report that an additional function of Als3 is that it can act as an invasin protein and induce endocytosis by normally nonphagocytic host cells . Furthermore , we show that this endocytosis occurs as a result of the N-terminal region of Als3 binding to either N-cadherin or E-cadherin . These conclusions are supported by the following results . First , an als3Δ/als3Δ null mutant was endocytosed poorly by endothelial cells and two different oral epithelial cell lines . Second , latex beads coated with rAls3-N were efficiently and specifically endocytosed by both endothelial cells and oral epithelial cells . Third , the rAls3-N–coated beads were avidly endocytosed by CHO cells expressing either N-cadherin or E-cadherin , but not by CHO cells that did not express these cadherins . These current results are also consistent with our previous data that S . cerevisiae expressing Als3 are endocytosed by endothelial cells [23] . In our prior studies , we have determined that C . albicans must be endocytosed by either endothelial cells or oral epithelial cells to cause damage to these cells [4 , 5 , 39] . We found that the als3Δ/als3Δ null mutant caused virtually no damage to either endothelial cells or FaDu oral epithelial cells . Similarly , Zhao et al . [24] reported that an als3Δ/als3Δ null mutant had markedly reduced capacity to damage reconstituted human epithelium . Our results suggest that the reduced capacity of the als3Δ/als3Δ null mutant to damage endothelial cells and oral epithelial cells in vitro is due to its defect in invading these cells . The invasion and damage defects of the als3Δ/als3Δ null mutant suggest that Als3 is important for virulence during hematogenously disseminated and oropharyngeal candidiasis . Investigations to determine the role of Als3 in C . albicans virulence are currently ongoing . However , we have found that vaccination of mice with rAls3 provides significant protection from lethal hematogenously disseminated candidiasis and ameliorates the course of oropharyngeal and vaginal candidiasis [40] . Thus , Als3 is a useful therapeutic target . In our assays , the als3Δ/als3Δ null mutant had significantly reduced adherence to endothelial cells , but normal adherence to oral epithelial cells . It has been reported by others that an als3Δ/als3Δ null mutant had reduced adherence to endothelial cells , buccal epithelial cells , and FaDu epithelial cells [24 , 35] . Also , we found that S . cerevisiae expressing Als3 bound to both endothelial cells and FaDu oral epithelial cells [23] . The discrepancy between the previous and current results is likely due to differences in the methodology of the assays . Specifically , the longer incubation time and the use of a 24-well tissue culture plate ( rather than a 6-well plate ) in the current endocytosis assay make it less sensitive to differences in adherence among strains . We also found that the als1Δ/als1Δ mutant was endocytosed normally by oral epithelial cells , whereas latex beads coated with rAls1-N induced some endocytosis by these cells . The probable explanation for these apparently conflicting results is that the als1Δ/als1Δ mutant still expressed Als3 on its surface . Thus , the presence of Als3 masked the effects of the absence of Als1 when the endocytosis assay was performed using whole organisms . Studies performed with latex beads coated with rAls1-N indicated that Als1 is capable of inducing epithelial cell endocytosis by itself . Collectively , these results demonstrate the utility of using both mutant analysis and recombinant proteins to assess the contribution of an individual Als protein to host cell interactions . Previously , we determined that the N-terminal region of Als proteins contains the substrate-binding site [23 , 26] . This conclusion is supported by our current finding that latex beads coated with just the N-terminus of Als3 were able to induce endocytosis . However , it was notable that coating the latex beads with rAls3-N resulted in a much greater increase in endocytosis than adherence . Also , coating beads with rAls1-N resulted in little to no increase in adherence , even though previous studies have clearly demonstrated that Als1 is an adhesin [36 , 37] . The likely explanation for the relative lack of adherence of the beads is that they were coated with fragments of Als1 or Als3 that lacked the tandem repeats present in the central portion of the full-length proteins . Previously , we have found that the adhesive function of Als1 is dramatically reduced when some or all of the tandem repeats are eliminated [26] . Similarly , Oh et al . [35] found that versions of Als3 with fewer tandem repeats mediated less adherence than did versions of Als3 with more tandem repeats . In addition , Rauceo et al . [38] reported that the absence of tandem repeats reduced the binding affinity of Als5 . Collectively , these results support a model in which the tandem repeats influence the conformation and/or the substrate accessibility of the N-terminal region of Als3 and thereby enhance its binding affinity for host substrates . However , even the weak binding of the N-terminal portion of Als3 to N-cadherin or E-cadherin is sufficient to induce endocytosis . The results with the als3Δ/als3Δ null mutant and the latex beads coated with rAls3-N demonstrate that endothelial cell N-cadherin and epithelial cell E-cadherin are two host cell ligands for Als3 . Furthermore , binding of the N-terminus of Als3 to either of these cadherins is sufficient to induce endocytosis . Previously , we have reported that N-cadherin is one of the endothelial cell surface proteins that mediate endocytosis of C . albicans [9] . E-cadherin is known to mediate the endocytosis of Listeria monocytogenes by epithelial cells [41] , but it was not previously known to mediate the endocytosis of any fungi . Our results with C . albicans raise the interesting possibility that perhaps L . monocytogenes can utilize N-cadherin to invade endothelial cells . Although the N-termini of Als1 and Als3 share considerable homology at the amino acid level , Als3 induced endocytosis much more efficiently than did Als1 . The results of computer-assisted modeling provide a possible structural explanation for these functional differences . A striking finding was that the 3-D structures of the N-terminal regions of Als1 and Als3 are predicted to have recurrent β barrel domains and a global negative surface potential similar to the N-terminal regions of N-cadherin and E-cadherin . Even though N-terminal regions of Als1 and Als3 share general features , they are predicted to differ both in their 3-D structure and their interactions with N- and E-cadherin . While Als1 appears to favor an extended conformation , Als3 tends toward a more folded structure that forms a more distinct molecular cleft . Our molecular modeling results also suggest that Als1 binds more weakly to N- and E-cadherin than does Als3 . These results are consistent with our finding that beads coated with rAls3-N were endocytosed more avidly by CHO cells expressing N- or E-cadherin than were beads coated with rAls1-N . Computer-assisted modeling also predicted that Als1 and Als3 interact with N- and E-cadherin via different domains , and the geometry of Als3 binding to E-cadherin is similar to that of E-cadherin binding to itself . Collectively , these findings support the hypothesis that Als3 functions as a molecular mimic of mammalian cadherins , thereby facilitating C . albicans invasion of endothelial and oral epithelial cells . The N-terminal regions of Als1 and Als3 are also predicted to share structural homology to bacterial adhesins and invasins . The conformation of Als1 is similar to that of the pfam02368 family of bacterial immunoglobulin-like domain surface protein/adhesins , which includes the invasin of Y . pseudotuberculosis [42 , 43] . In contrast , the cleft motif of Als3 is more similar to that of the COG4886 family of leucine-rich repeat invasins/intimins [42] . Interestingly , this family of proteins includes internalin A of L . monocytogenes , which mediates epithelial cell invasion by binding to E-cadherin [41 , 44] . The models presented here do not purport to specify exact interactions between cadherins and Als proteins . Defining such interactions at the atomic level requires physicochemical methods such as ultracentrifugation and x-ray crystallography . However , our findings are strengthened by the fact that the results of our modeling of cadherin–cadherin interactions were very similar to what has been found by x-ray crystallography [45] . Previously , we have reported that S . cerevisiae expressing Als1 or Als3 binds to a variety of host constituents , including endothelial cells , epithelial cells , laminin , and fibronectin [23] . In the affinity purification experiments , hyphae of the als3Δ/als3Δ null mutant failed to bind to multiple host proteins , including N-cadherin and E-cadherin . Thus , these results are consistent with the model that Als3 binds to a broad range of host substrates . However , our data also indicate that there is some specificity in the binding of Als proteins to host constituents . For example , beads coated with rAls3-N were endocytosed much more efficiently than were beads coated with rAls1-N . These results , combined with those of the molecular modeling studies , suggest that Als3 binds with greater affinity to N- and E-cadherin than does Als1 . Nevertheless , it is highly likely that additional host cell ligands , other than N- or E-cadherin , also contribute to the endocytosis of C . albicans , and we are currently working to identify them . For many microbial pathogens , invasion of host cells is critical for the initiation and maintenance of infection , and many of these organisms have more than one mechanism for inducing their own uptake by host cells [46] . Thus , it is logical to speculate that C . albicans has also evolved at least one Als3-independent mechanism to invade host cells . The Als family of proteins contains six members in addition to Als1 and Als3 [20–25] . Most of these proteins are known to have adhesive function . When expressed in S . cerevisiae , Als5 induces weak endothelial cell endocytosis similar to that induced by Als1 [23] . Als6 , Als7 , and Als9 do not appear to mediate significant endothelial cell endocytosis , and Als2 and Als4 have not yet been tested in this assay . While it is possible that Als3 is the only member of its family that has significant invasive function , it is also conceivable that one or more of the other Als proteins mediates invasion of other types of host cells . Thus , a detailed analysis of these proteins will likely provide further insight into the pathogenesis of hematogenously disseminated and mucosal candidiasis , and perhaps additional therapeutic targets .
For use in the experiments , all C . albicans strains were grown as blastospores in liquid YPD ( 1% yeast extract [Difco , http://www . bdbiosciences . com] , 2% bacto-peptone [Difco] , and 2% D-glucose ) medium in a shaking incubator at 30 °C overnight . The blastospores were harvested by centrifugation , washed twice in phosphate-buffered saline without calcium or magnesium ( PBS ) , and enumerated using a hemacytometer . Auxotrophic strains were grown on YPD agar supplemented with histidine , arginine , or uridine as needed . To produce hyphal phase organisms , blastospores were incubated in RPMI 1640 medium ( Irvine Scientific , http://www . irvinesci . com ) in a shaking incubator at 37 °C for 90–120 min as described [9] . The strains of C . albicans used in this study are listed in Table 2 . Strains 7371 and 7392 were isolated from patients with AIDS who had oropharyngeal candidiasis , and were a generous gift from Thomas Patterson ( University of Texas Health Sciences Center , San Antonio , Texas , United States ) . rAls1-N and rAls3-N were produced in S . cerevisiae as previously described [36 , 40] . The secreted proteins were concentrated by passing the conditioned culture supernatant through a 10-kDa filter , after which the Als proteins were affinity purified by passing the concentrate over a Ni-NTA column ( Qiagen , http://www . qiagen . com ) . The purity of the recombinant proteins was verified by SDS-PAGE and was at least 90% . Latex beads were coated with rAls1-N , rAls3-N , or biotinylated BSA by a minor modification of the method of Dersch and Isberg [47] . Briefly , 5 × 107 fluorescent yellow-green amine–modified polystyrene latex beads ( 2 . 0 μm diameter; Sigma-Aldrich , http://www . sigmaaldrich . com ) were washed first with PBS and then with coupling buffer ( 0 . 2 M Na2HCO3 [pH 8 . 5] and 0 . 5 M NaCl ) . The beads were then incubated with Als1 , Als3 ( 0 . 5 mg/ml ) , or coupling buffer at 37 °C for 30 min . Next , 1% rabbit serum was added to the beads that had been coated with Als proteins , and biotinylated BSA ( 1 mg/ml ) was added to the control beads . After the beads were incubated at 37 °C for an additional 30 min , they were sonicated briefly and then blocked by adding unlabeled BSA ( to 10 mg/ml ) in coupling buffer and incubating them for 1 h . Finally , the beads were then washed twice with PBS containing 10 mg BSA/ml , and suspended in PBS containing 2 mg BSA/ml . Binding of rAls1-N to the beads was verified by indirect immunofluorescence with a monoclonal antibody directed against rAls1-N [36] . Binding of rAls3-N to the beads was confirmed similarly using an affinity-purified rabbit polyclonal antiserum that been raised against rAls3-N ( PickCell Laboratories , http://www . pickcell-b2b . com ) and adsorbed on hyphae of the als3Δ/als3Δ mutant . Binding of biotinylated BSA to the beads was verified using Alexa 568–conjugated strepavidin ( Invitrogen , http://www . invitrogen . com ) . Endothelial cells were harvested from human umbilical cord veins by the method of Jaffe et al . [48] . They were grown in M-199 medium ( Invitrogen ) containing 10% fetal bovine serum ( FBS; Gemini Bio-Products , Inc . , http://www . gembio . com ) , 10% defined bovine calf serum ( Gemini Bio-Products , Inc . ) , and 2 mM L-glutamine with penicillin and streptomycin ( Irvine Scientific ) as described previously [4] . They were used at the second or third passage . The FaDu oral epithelial cell line , originally derived from a human pharyngeal carcinoma , was obtained from the American Type Culture Collection ( http://www . atcc . org ) . The cells were maintained in MEM Earl's salts ( Irvine Scientific ) containing 10% FBS , 1 mM pyruvic acid , 2 mM L-glutamine , 0 . 1 mM nonessential amino acids , 100 IU/ml penicillin , and 100 IU/ml streptomycin . The OKF6/TERT-2 cell line , which had been constructed by transforming oral mucosal keratinocytes from a healthy adult with a constitutively active telomerase catalytic subunit , was kindly provided by J . Rheinwald ( Harvard University , Cambridge , Massachusetts , United States ) [30] . These cells were maintained in subconfluent cultures as described elsewhere [30] . To grow confluent cells , the medium was changed to a mixture containing 50% keratinocyte serum-free medium , 25% Ham's F-12 , and 25% calcium- and glutamine-free DMEM ( all from Invitrogen ) . This mixture was supplemented with 1 . 5 mM L-glutamine , 25 μg bovine pituitary extract/ml , and 0 . 2 ng epidermal growth factor/ml . CHO K-1 cells were also obtained from the American Type Culture Collection . They were grown in Ham's F-12K medium ( American Type Culture Collection ) supplemented with 10% FBS , penicillin , and streptomycin . All cell types were grown at 37 °C in 5% CO2 . CHO cells expressing human N-cadherin and control CHO cells transfected with the empty vector pcDNA3 . 1 ( which expressed no human cadherins ) were produced as described previously [9] . CHO cells expressing human E-cadherin were generated following a similar procedure . Briefly , CHO cells were transfected with plasmid hEcad/pcDNA3 , a generous gift from C . Gottardi ( Northwestern University , Evanston , Illinois , United States ) [49] , using Lipofectamine Plus ( Life Technologies , http://www . invitrogen . com ) following the manufacturer's protocol . The transfectants were selected with G418 ( AG Scientific , Inc . , http://www . agscientific . com ) . Both the N-cadherin and E-cadherin transfectants were enriched by fluorescence-activated cell sorting for cells with high-level expression of these proteins [9 , 50] . N-cadherin– and E-cadherin–expressing cells were detected using mouse monoclonal antibodies that recognized the extracellular domains of N-cadherin ( clone GC34; Sigma-Aldrich ) and E-cadherin ( clone HECD-1; Abcam Inc . , http://www . abcam . com ) . The cells were used in the experiments when at least 75% of the transfectants had high level expression of either N- or E-cadherin . The number of organisms or latex beads that were endocytosed or cell-associated with the various host cells was determined using our standard differential fluorescence assay [5 , 7 , 9 , 51] . The host cells were grown to 95% confluency onto 12-mm diameter coverslips coated with fibronectin in 24-well tissue-culture plates . They were incubated with 105 C . albicans hyphae in RPMI 1640 medium . After 45 min , the cells were rinsed twice with Hank's balanced salt solution ( HBSS; Irvine Scientific ) in a standardized manner and then fixed with 3% paraformaldehyde . In experiments performed with C . albicans , the adherent but nonendocytosed organisms were labeled with rabbit polyclonal anti–C . albicans antibodies ( Biodesign International , http://www . biodesign . com ) that had been conjugated with Alexa 568 ( Invitrogen ) , which fluoresces red . Next , the cells were permeablized with 0 . 5% Triton X-100 ( Sigma-Aldrich ) in PBS , and then the cell-associated organisms ( the endocytosed plus nonendocytosed organisms ) were labeled with the anti–C . albicans antibodies conjugated with Alexa 488 ( Invitrogen ) , which fluoresces green . The coverslips were viewed using an epifluorescent microscope , and the number of endocytosed organisms was determined by subtracting the number of nonendocytosed organisms ( which fluoresced red ) from the number of cell-associated organisms ( which fluoresced green ) . At least 100 organisms were examined on each coverslip , and the results were expressed as the number of endocytosed or cell-associated organisms per high-powered field . Experiments investigating the endocytosis and adherence of the yellow-green fluorescing latex beads were performed similarly , except that 3 × 105 beads were added to each well . The adherent , nonendocytosed control beads coated with biotinylated BSA were labeled with strepavidin Alexa 568 . The adherent beads coated with either rAls1-N or rAls3-N were detected by indirect immunofluorescence using rabbit polyclonal anti-Als1 antibodies followed by Alexa 568–conjugated goat anti-rabbit antibodies . Membrane proteins from host cells were labeled with biotin and then isolated as described previously [9] . Briefly , endothelial or oral epithelial cells were incubated with EZ link Sulfo NHS- LC biotin ( 0 . 5 mg/ml; Pierce , http://www . piercenet . com ) in PBS with calcium and magnesium ( PBS++ ) for 12 min at 37 °C in 5% CO2 . After extensive rinsing with cold PBS++ , the cells were scraped from the tissue-culture dishes and collected by centrifugation at 500g . Next , the cells were lysed by incubating them for 20 min on ice in PBS++ containing 5 . 8% octyl-glucopyranoside ( wt/vol; Sigma-Aldrich ) and protease inhibitors ( 1 mM PMSF , 1μg/ml pepstatin A , 1μg/ml leupeptin , and 1μg/ml aprotinin ) . The lysate was clarified by centrifugation at 5000g for 5 min at 4 °C , after which the supernatant was centrifuged at 100 , 000g for 1 h at 4 °C . The resultant supernatant was stored in aliquots at −80 °C . The protein concentration was determined by the Bradford assay ( Bio-Rad Laboratories , http://www . bio-rad . com ) . The host cell membrane proteins that bound to C . albicans hyphae were isolated by a minor modification of our previously described affinity purification procedure [9] . In brief , 250 μg of the biotinylated membrane proteins in PBS++ containing 1 . 5% octyl-glucopyranoside and protease inhibitors was incubated with 2 × 108 C . albicans hyphae on ice for 1 h . The unbound proteins were removed by extensive rinsing in the same buffer . Next , the proteins that had bound to the hyphae were eluted with 6 M urea . The eluted proteins were separated by SDS-PAGE on an 8% gel and detected by immunoblotting with an anti-biotin murine monoclonal antibody ( clone BN-32; Sigma-Aldrich ) , an anti-N-cadherin murine monoclonal antibody ( clone 32; Transduction Laboratories , http://www . bdbiosciences . com ) , or the anti-E-cadherin murine monoclonal antibody . The membranes were then incubated with an appropriate horseradish peroxidase–conjugated secondary antibody , and the bands were visualized using enhanced chemiluminescence ( Pierce ) . Indirect immunofluorescence was used to determine whether endothelial cell N-cadherin and oral epithelial cell E-cadherin colocalized with hyphae of the various strain of C . albicans as previously described [9] . Endothelial or FaDu oral epithelial cells were grown onto 12-mm diameter glass coverslips as in the endocytosis experiments . The cells were infected with 2 . 5 × 105 blastospores of C . albicans for 90 min . During this time , the organism germinated and formed hyphae . Next , the cells were fixed and permeabilized with 3% paraformaldehyde containing 0 . 5% triton X-100 in PBS++ . They were rinsed once with 1% BSA in PBS and blocked with 5% goat serum in PBS . To visualize F-actin , the cells were incubated with Alexa 568 phalloidin . After extensive rinsing , the cells were incubated with either the anti-E-cadherin murine monoclonal antibody or an anti-N-cadherin monoclonal antibody ( clone 32 ) , rinsed and then incubated with polyclonal goat anti-mouse antibodies conjugated with Alexa 488 . The C . albicans cells were then labeled by incubation with the rabbit anti–C . albicans antiserum conjugated with Alexa 633 ( Invitrogen ) . The coverslips were mounted inverted , and imaged by confocal scanning laser microscopy . Images of the organisms and host cells were acquired , and three to five optical sections were stacked along the z-axis to produce the final images . To determine the effects of saliva on the interactions of C . albicans with FaDu oral epithelial cells , the organisms were germinated in RPMI 1640 medium to form hyphae as described above . Next , they were killed by incubation in 0 . 02% thimerosal ( 2-[ethylmercuriomercapto] benzoic acid sodium salt ) in HBSS at 37 °C for 1 h [5] and then rinsed extensively . This method of killing was used to minimize changes in the cell wall [52] . Uninduced saliva was collected from normal volunteers . It was centrifuged at 500g and then sterilized by passage through a 0 . 22-μm filter . The killed organisms were incubated in 20% saliva in RPMI 1640 medium at 37 °C for 1 h , after which they were rinsed twice with HBSS . Control organisms were incubated in RPMI 1640 medium without saliva in parallel . The ability of the saliva-coated organisms to bind to oral epithelial cell proteins and induce endocytosis was determined as described above . Flow cytometry was used to analyze the surface expression Als3 on hyphae of the various strains using a minor modification of our previously described method [23] . Briefly , hyphae of the different strains of C . albicans were fixed in 3% paraformaldehyde and blocked with 1% goat serum . The hyphae were then incubated with either a rabbit polyclonal antiserum raised against rAls3-N or purified rabbit IgG . After extensive rinsing , the cells were incubated with a goat anti-rabbit secondary antibody conjugated with Alexa 488 . The fluorescent intensity of the hyphae was measured using a FACSCaliber flow cytometer ( Becton Dickinson , http://www . bd . com ) . Fluorescence data for 10 , 000 cells of each strain were collected . The structures of rAls1 and Als3 were modeled using a combination of knowledge- and energy-based methods to avoid limits of any single method in assessing folding of multiple Als domains . Homolog searches were conducted on 100–200 residue blocks of each Als sequence . The domains identified were joined , and the most likely domain configuration was determined by Monte Carlo searches of ϕ and ψ angles for connecting regions . The domains presented were the consensus of folds identified by Composer [53] , MatchMaker [54] , GeneFold [55] , Robetta [56] , Fugue [57] , and Swiss-Model [58] . The AMBER99 force fields were used throughout these studies to optimize models . AMBER8 [59] and SYBYL 7 . 0–7 . 2 ( Tripos Inc . , http://www . tripos . com ) were implemented on multiple graphics workstations ( Sun Microsystems , Inc . , http://www . sun . com; and SGI , Inc . , http://www . sgi . com ) using LINUX and IRIX operating systems . In addition to the semi-automated procedures above , backbone trajectories were obtained using FASTA searches of the Research Collaboratory for Structural Bioinformatics ( RCSB ) protein databank ( http://www . rcsb . org ) . Side chains of the target rAls1 and rAls3 models were refined by molecular dynamics and minimization of strain energies , while retaining backbone trajectory by fixing positions of the peptide backbone atoms . The torsion angles of all peptide bonds were adjusted to 180° ± 15° , with minimal constraints . Where appropriate , selected model constraints applied a 0 . 4-kJ penalty to the Ramachandran ϕ and ψ angles to enforce regions of predicted α and β structures . Final global energy minimizations were performed for each model after removal of all constraints . Finally , the physicochemical properties of the peptides were visualized by MOLCAD [60] as implemented in SYBYL 7 . 0–7 . 2 and HINT [61] . In these utilities , the physical properties of the peptides are superimposed upon the backbone trajectory , or projected onto the water-accessible surface of the molecules . Models of the N-terminal regions of Als1 and Als3 generated as above were then examined for predicted interactions with E-cadherin or N-cadherin [62] in silico using the program Hex 4 . 5 [63 , 64] . The structures of these cadherins have been determined to a resolution of 2 Å by x-ray crystallography [45] . To enhance the predictive accuracy of modeling and evaluate all probable interactions , multiple docking simulations were performed from different starting positions . For example , in a systematic approach , docking was initiated with the target cadherin located next to each of the major Als N-terminal domains . The search range was set to 60 Å using full rotation mode and a 0 . 75 Å step size with two substeps . Initial searches were based on shape and electrostatics; post-processing was based on 3-D shape and surface topography . Post-processing with molecular mechanics or molecular mechanics minimization ( 0 . 75 Å step size , two substeps ) gave equivalent results . As a comparator to Als-cadherin interactions , the self-association of two N-terminal domains of E-cadherin and N-cadherin was also assessed . To assess the validity of the hex algorithms applied , we separated the two E-cadherin molecules , the structures of which are contained in the Protein Data Bank ( http://www . pdb . org ) [45] . The two peptides were then redocked using the approach described above , initiated from multiple orientations of the peptides . In every case , the best dock was clearly discernable by the energy score . Visually , E-cadherin redocking was virtually indistinguishable from that of the reported structure determined by crystallography [45] . Quantitatively , redocked E-cadherins had a root mean square deviation of < 1 Å deviation from the crystal structure . No attempt was made to optimize the output because the crystal packing forces may account in part for any differences between published structures [45] and hex algorithm model output . Nonetheless , docking parameters ( Table 1 ) varied less than 10% between the experimental and redocked E-cadherins ( unpublished data ) . The data were evaluated using analysis of variance , and p-values ≤ 0 . 05 were considered significant .
The Protein Data Bank ( http://www . pdb . org ) accession numbers for the structures mentioned in this article are E-cadherin ( 1EDH ) and N-cadherin ( 1NCJ ) ; the Candida Genome Database ( http://www . candidagenome . org ) accession numbers for the genes mentioned in this article are ALS1 ( orf19 . 5741 ) and ALS3 ( orf19 . 1816 ) . | The fungus Candida albicans is usually a harmless colonizer of human mucosal surfaces . In the mouth , it can cause oropharyngeal candidiasis , also called thrush . In hospitalized and immunocompromised patients , C . albicans can enter the blood stream and be carried throughout the body to cause a disseminated infection , which is associated with a mortality rate of up to 40% . The organism invades the epithelial cell lining of the mouth during oropharyngeal candidiasis and invades the endothelial cell lining of the blood vessels during disseminated candidiasis . We discovered that Als3 , a protein expressed on the surface of C . albicans , is required for this invasion process . Cadherins on the surface of human cells normally bind other cadherins for adhesion and signaling; however , we found that Als3 also binds to cadherins on endothelial cells and oral epithelial cells , and this binding induces these host cells to take up the fungus . The structure of Als3 is predicted to be quite similar to that of the two cadherins studied , and the parameters of the binding of Als3 to either cadherin are similar to those of cadherin–cadherin binding . These results suggest that Als3 is a functional and structural mimic of human cadherins , and provide new insights into how C . albicans invades host cells . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] | [
"infectious",
"diseases",
"cell",
"biology",
"yeast",
"and",
"fungi",
"microbiology"
] | 2007 | Als3 Is a Candida albicans Invasin That Binds to Cadherins and Induces Endocytosis by Host Cells |
Hepatitis C virus ( HCV ) is present in the host with multiple variants generated by its error prone RNA-dependent RNA polymerase . Little is known about the initial viral diversification and the viral life cycle processes that influence diversity . We studied the diversification of HCV during acute infection in 17 plasma donors , with frequent sampling early in infection . To analyze these data , we developed a new stochastic model of the HCV life cycle . We found that the accumulation of mutations is surprisingly slow: at 30 days , the viral population on average is still 46% identical to its transmitted viral genome . Fitting the model to the sequence data , we estimate the median in vivo viral mutation rate is 2 . 5×10−5 mutations per nucleotide per genome replication ( range 1 . 6–6 . 2×10−5 ) , about 5-fold lower than previous estimates . To confirm these results we analyzed the frequency of stop codons ( N = 10 ) among all possible non-sense mutation targets ( M = 898 , 335 ) , and found a mutation rate of 2 . 8–3 . 2×10−5 , consistent with the estimate from the dynamical model . The slow accumulation of mutations is consistent with slow turnover of infected cells and replication complexes within infected cells . This slow turnover is also inferred from the viral load kinetics . Our estimated mutation rate , which is similar to that of other RNA viruses ( e . g . , HIV and influenza ) , is also compatible with the accumulation of substitutions seen in HCV at the population level . Our model identifies the relevant processes ( long-lived cells and slow turnover of replication complexes ) and parameters involved in determining the rate of HCV diversification .
Hepatitis C virus ( HCV ) is a member of the hepacivirus genus within the flaviviridae family of virus , and it has a single positive stranded RNA molecule ( ∼9500 nucleotides ) as its genome [1]–[3] . After entering a cell this RNA is translated into a single large polyprotein , which is cleaved to produce the viral structural and non-structural ( NS ) proteins [1]–[3] . The NS5B protein is a viral-specific polymerase , which is involved in replicating the HCV RNA genome [1] , [4] . During genome replication the virion's positive strand RNA is copied into a complementary negative strand , which then must be copied back to produce a new positive strand . In the simplest replication model , this negative strand or a complex of the original positive strand and the newly created negative strand form an intermediate that acts as the template for producing new positive strands . This template plus various non-structural proteins form a structure called a replication complex [5] . If all new positive strands , and hence virions , are created from the same replication complex , we say that replication occurs by a “stamping machine" mechanism [6]–[9] . However , HCV infected cells often have more than one replication complex; indeed in vitro and in situ studies suggest there are about 40 such complexes in one infected cell [4] , [10] . The HCV polymerase is an RNA-dependent RNA polymerase ( RdRp ) and hence does not possess error correcting mechanisms . Thus HCV replication , like that of other RNA viruses , is highly error prone [1]–[3] . Measuring the actual mutation rate , which derives both from the ( + ) RNA to ( − ) RNA and the ( − ) RNA to ( + ) RNA steps of replication , has been difficult [6] , [11] , [12] . A recent study determined the intrinsic error rate of the HCV polymerase in vitro using enzyme kinetic measurements [12] . They found high error rates , of ∼10−3 per site , for transitions and about 100-fold lower rates for transversions . Still , the in vivo mutation rate is likely different . Mutation is difficult to estimate in vivo due to selection , multiple rounds of replication and incomplete sampling [6] , [11] . One proposed way to determine the in vivo mutation rate is to estimate it based on the frequency of lethal mutants in the viral population at any given time [13] . In fact , classical genetics shows that the frequency of a lethal mutation in a haploid population in mutation-selection balance is μ , the mutation rate . A recent study used this method to estimate an upper limit for the in vivo mutation rate of HCV as ( 1 . 15±0 . 29 ) ×10−4 per nucleotide per replication round [13] , which is within the range of other RNA viruses [6] . This high mutation rate is consistent with the high degree of HCV diversity found across the population of infected individuals [14] , [15] . Indeed , HCV is highly variable , with multiple subtypes , and a global diversity that is higher than that of HIV-1 [15] . Clearly , this population level diversity , which reflects the HCV evolution rate , is in part prescribed by the mutation rate of the virus in vivo [16] . Moreover , in chronically infected individuals the HCV viral population is also diverse [17] . This diversity allows fast evolution and escape from immune [18] or antiviral drug pressure [19] , and may contribute to HCV pathogenesis [18] , [20] . An important question is how HCV diversity is generated . While it clearly depends on the mutation rate , we shall show using a model of HCV replication that it also depends on other parameters of the HCV life cycle [7]–[9] , such as the long-lived nature of infected cells , as compared to HIV infected cells [21] , [22] , the existence of multiple replication complexes within an infected cell [4] , [10] , and the turnover rate of these replication complexes . In order to validate this model and obtain quantitative estimates of the in vivo HCV mutation rate , we shall exploit our observations in an accompanying report [23] and those of others [24] , [25] that during the initial stages of primary infection the viral population is comprised of discrete low diversity lineages of viral sequences emanating from the transmitted/founder viral genomes [23] . Further , early on , diversity increases with time since infection . We shall show that the rate of diversification is not constant but rather slows as infection is established . Our model provides a quantitative explanation for this phenomenon . Analyses of HIV evolution in acute infection have been used to estimate the time since infection [26] , [27] . Here , we know with reasonable accuracy the time of infection , but use the same ideas to estimate the in vivo mutation rate of HCV .
The early dynamics of viral increase in HCV infection is different from that seen in other chronic infections , such as HIV [28] and HBV [29] . The HCV viral load in the subjects in this study increases roughly exponentially until it reaches a plateau ( Figure 1A ) . This has also been observed in a prior study of acute HCV infection [30] and observed in chimpanzees experimentally infected with HCV [31] . Quantitative characteristics of this early increase are given in Table 1 . The median time between the last negative sample and the first HCV positive sample in our dataset was 5 days , which is consistent with a viral dynamics analysis of larger numbers of plasma donors [30] . Because of this short interval , we assumed that the virus started expanding at the last negative sample . If the virus started expanding after this , our estimated expansion rate would be an underestimate . The median HCV RNA exponential growth rate was 2 . 2/day , corresponding to a doubling time of 0 . 31 days ( or 7 . 4 hours ) . The median peak viral load observed was 3×106 HCV RNA IU/ml and it took a median of 21 days to reach this level . The virus then stayed at approximately this high viral load level for a median of at least 26 days . In two subjects , we did not have enough follow-up to conclusively affirm whether a plateau exists or not . These estimates are in agreement with a previous study of 77 plasma donors with longer follow-ups , which reported an estimate of ∼6 days of viral expansion before the first positive measurement ( compared to a median of 5 days in our dataset ) and a mean plateau duration of ∼56 days [30] . The observation of the viral load plateau suggests that the number of infected cells reaches a steady state level a couple of weeks post infection . It is possible that this is a dynamic steady state , with removal of infected cells in equilibrium with generation of new infected cells . However , HCV is likely non-cytolytic [32] , consistent with the normal levels of alanine aminotransferase ( ALT<40 IU/L is upper limit of normal [33] , [34] ) in these individuals early in infection ( Figure 1B ) . In addition , prior work has suggested that the cytolytic immune response takes weeks to months to emerge [31] , [35] , [36] ( consistent with an increase in ALT to 10× to 20× the normal level late in acute infection [37] ) . Thus , it is likely that the rate of infected cell death during this early period is comparable to that of uninfected cells . The lifespan of uninfected hepatocytes has been estimated as being on the scale of months to years [38] , [39] , and thus infected cell death is probably negligible at these early times . In this case , the plateau in viral load suggests an equilibrium where all cells that can be infected are infected and producing virus . Assuming that there are 1011 hepatocytes in the liver [40] , we estimate that a median of 6% ( with range 1 . 7%–22% ) of these are infected across our subjects ( Table 1 ) , consistent with experimental measurements in chronic infection [41] , including recent estimates by two-photon microscopy of frozen sections of liver tissue [42] . Thus , primary HCV infection is characterized by fast growth of viral load to a plateau where only a minority of hepatocytes is infected . To evaluate how HCV diversity changes during primary infection , we performed single genome amplification ( SGA ) followed by direct amplicon sequencing [23] , [26] , otherwise known as single genome sequencing [43] , at multiple time points in the subjects shown in Figure 1 . SGA is achieved through serial dilution of the cDNA obtained by reverse transcription of HCV RNA from plasma ( see Methods and [23] for details ) . We amplified 5′ half-genome sequences , on average 4879 nucleotides , covering core , E1 , E2 , p7 , NS2 and most of the NS3 proteins of HCV . For early samples , with low viral loads , we amplified the same region , but in two separate assays of one quarter genome each to enhance sensitivity of amplification . In this way , we obtained 84 sets of sequences for the 9 subjects at multiple ( between 3 and 5 ) time points . On average , we had 44 sequences per time point . All of the sequences were deposited in Genbank; see Li et al . [23] for further details and accession numbers . We then aligned separately the set of sequences for each time point and for each sequence region and used a sequence visualization tool ( Highlighter – www . HIV . lanl . gov ) , to analyze the sequence diversity based on individual nucleotides . This tool allowed us to identify low diversity monophyletic lineages corresponding to the putative transmitted/founder ( T/F ) viruses – the consensus at the earliest time point from SGA data [23] . We next confirmed that these lineages were maintained across the times sampled , to guarantee that we were analyzing the diversification of the same lineage over time . In cases where there were two or more putative T/F viruses , we analyze only the dominant lineage , as SGA sequence data was too limited to study the minor lineages . From these 84 sequence alignments , we were able to study the evolution of the virus and the emergence of new mutations from very early in infection ( mean: 7 days , range 2 to 15 days since the last negative sample across the 9 patients ) until late in the plateau phase of viral load ( mean: 33 days , range 21 to 42 days ) . We found that HCV sequence diversity increases quickly early on , but then stabilizes in 7 patients , starting at about day 14; in subject 10051 there was not enough follow up to assess this issue , and in subject 10029 a clear stabilization of diversity was not observed . The plateau of diversity occurred when an average of 46% of the sequences were still identical to the inferred T/F viral genomes . In three subjects ( 10029 , 10062 , 9055 ) there was an increase in diversity at late times , ∼35 days . Note that for 10062 , this is coincident with an increase in ALT levels suggesting turnover of infected hepatocytes ( Figure 1B ) . We also found that in the vast majority of cases , HCV diversity at each time point was consistent with a star-like phylogeny , i . e . the viruses' sequences coalesce at a single genome founder [27] , [44] . The only exception was the 5′-half of 9055 at the last sampling time point , day 38 , when there was evidence for the onset of immune selection [23] . The mutations detected in the sequence sets also conformed to a Poisson distribution in the inter-sequence pairwise Hamming distances [27] . The exceptions were the 5′-half of 10029 at day 13 , the second 5′ quarter of 10029 at day 34 , the second 5′ quarter of 10051 at day 7 , and the first quarter and 5′-half of 10051 at day 21 . Due to the specifics of the HCV replication life-cycle , one predicts occasional violations in star-like diversification and in the fit to the Poisson distribution , because there is a non-negligible probability of shared stochastic mutations between HCV sequences . That is , shared mutations may occur even in the absence of selective forces . See the accompanying report [23] for a more detailed discussion of these issues . We next developed a model of HCV replication to study the time course of accumulation of mutations and to estimate the in vivo mutation rate of HCV needed to describe the observations above . This stochastic model of HCV replication allowed us to study the time course of viral load changes and the accumulation of mutations in the study subjects ( see Methods ) . In the model , we assume cells are infected by a single virion , i . e . , that superinfection does not occur [45] , [46] . We further assume that in every infected cell , on average , only a fraction k of newly synthesized viral ( + ) strand RNA ( vRNA ) is exported in new virions , and the rest , 1-k , forms new replication complexes ( RC ) . We assume that vRNA degradation can be neglected , i . e . , that the newly synthesized vRNA is either rapidly complexed with proteins and converted into stable RC , or rapidly encapsidated and exported . ( Note that this is very different from analyses of HCV treatment , when production of vRNA and/or virion assembly/release may be blocked , and vRNA degradation becomes an important parameter in the clearance of infection [47] ) . These processes are assumed to continue until the cell generates a maximum number of replication complexes ( RCM ) . Note that if we set k = 1 , so that all synthesized vRNAs are exported , we recover the “stamping machine" mode of replication [7]–[9] , where all virions result from the same replication complex , i . e . , the same negative strand of RNA . The existence of multiple replication complexes within one cell corresponds to “geometric growth" . In our model , after a virus is exported , a fraction 1-θ of the released virions is assumed to be cleared from circulation [5] , and the remaining fraction , θ , is assumed to infect new cells . We also assume infected cells are long lived , and thus , we initially neglect death of infected cells during the first few weeks of infection . This assumption is consistent with the viral load profiles seen in the infected subjects , where viral load increases rapidly to a maximum level and plateaus at this level for weeks . We used our model to reproduce the viral load data ( Figure 1A ) . For each subject , the only free parameter available to determine the trajectory of virus over time is the fraction of vRNA exported , k , since all other parameters are fixed a priori or are calculated as a function of k ( see Methods ) . We found that the model could describe the viral load data well with just this single adjustable parameter . The values estimated for k indicate that most of the synthesized vRNA is exported as virions ( median k = 0 . 77 , range 0 . 42–0 . 89 ) . Moreover , the estimated values of k are quite similar among the different individuals , with the exception of 10025 , who has a lower estimated k ( = 0 . 42 ) . However , this subject has only one viral load measurement during the up-slope of the virus , which strongly influences the value estimated for k . Indeed , for this individual , choosing higher values for k lead to only slightly lower quality fits ( not shown ) . Next , we used our model to analyze the diversification profiles of HCV in these patients . As the viral RNA is copied , errors in the incorporation of nucleotides are possible , i . e . , mutations occur . If we let μ denote the probability that a base in the newly produced virion differs from that in the infecting virion , then for the stamping machine model the mutation rate , μ , is simply twice the rate at which bases are miscopied by the HCV RdRp , to account for the cycle of ( + ) RNA strand→ ( − ) RNA strand→ ( + ) RNA strand copying . With multiple replication complexes in a cell , opportunities exist for additional copying errors to be made since a newly synthesized ( + ) RNA strand needs to be copied again to make a replication complex . Every time a RNA strand incorporating a mutation is made , there is a probability that this mutation is lethal , and the virus or replication complex made from such RNA is non-functional . Prior experimental studies indicate that the fraction of random mutations that are lethal is about 40% in RNA viruses [48] . We incorporated mutation in our model to analyze the viral diversification data and estimate the mutation rate needed to match the observed accumulation of mutations . We assume that at time zero the putative T/F virus starts replicating and mutating . We then compute the decrease over time in the fraction of sequences identical to the T/F virus ( i . e . , “the fraction of unmutated viruses" ) . We compare this model prediction to the identical measurement in our subjects and varied the mutation rate to obtain the best agreement between model and sequence data obtained from plasma HCV RNA , which corresponds to ( + ) RNA strands . The best description of the data was obtained for a median mutation rate ( for the half-genomes ) of μ = 2 . 5×10−5 per nucleotide per replication ( Figure 2A–C ) . Moreover , this estimate was consistent across subjects and across regions of the genome ( range: 1 . 6×10−5–6 . 2×10−5 per nucleotide per replication , Table 1 ) . Our model exhibits a fast decrease in sequence identity early in infection , as the viral load increases exponentially and more and more cells are infected , followed by a stable viral diversity level as the virus reaches and stays at its plateau . This stasis in viral diversification is compatible with the assumption that the plateau in viral load corresponds to a stable pool of infected cells . This indeed seems to be the case for 5 of the patients ( Figure 2A–C ) ; for 1 case there is not enough data . If the plateau in viral load corresponded to a dynamic steady state in which infected cells were dying and being rapidly replaced , our model would predict a continuous increase in diversification resulting from the continuous replacement of replication complexes . In a few cases , we did see an increase in diversity at times later than 30 days , and in three patients ( 10029 , 10062 , and 9055 ) the observed long term behavior ( later than about day 35 ) deviates from that predicted by our simulations . This difference between model and data could be due to sampling error , for example the 95% CI for theory and data at day 42 overlap for patient 10062 . Alternatively , some processes not accounted for in the model may be operational at these later time points , leading to increased diversity . For example , for subject 9055 anti-HCV antibodies are detectable at this late time point and there is strong evidence of CTL selection ( escape or reversion ) [23]; and for 10062 there is a late increase in ALT ( Figure 1B ) , which suggests the initiation of a CTL response consistent with renewed cycles of infection . Our model also makes predictions about the distribution of mutations across the population . Interestingly , our model not only matches the fraction of unmutated viruses , but also the fraction of viruses with 1 , 2 , 3 , … mutations , even though this detailed data was not used to parameterize the model ( Figure 3A–C ) . We obtained excellent agreement with the data , except when we observed a late increase in diversity in the three patients discussed above ( 10029 , 10062 , 9055 ) . We tested this agreement for the 5 h genomes by a Monte Carlo test [49] , since the number of expected mutations is low ( <5 ) in several cases . The null hypothesis is that the data follows the theoretical expected values , and with the exception of those three patients , there was good agreement between observed and predicted mutation counts ( p>0 . 05 ) . Moreover , if we consider the distribution of mutations at the previous time for which we have SGA data , this agreement was also seen in 10029 and 10062 ( p>0 . 05 , and we cannot reject the null hypothesis ) . We next tested whether our results were dependent on the particular values assigned to the parameters that we fixed in the simulation ( see Methods ) . We found that both the viral load time course and the viral diversification were not sensitive to particular values of these parameters ( Figure S1 in Text S1 ) . For example , we assumed a maximum of RCM = 40 replication complexes per infected cell , as seen in vitro [4] and in situ [10] . Clearly this number could be different in vivo . However , our results were essentially the same , when we varied RCM from 10 to 80 ( Figure S1 in Text S1 ) . To further confirm the robustness of our results , we next used the method suggested by Cuevas et al . [13] for estimating the mutation rate of HCV by analyzing the frequency of lethal mutations . Classical genetics shows that the frequency of lethal mutations is equal to the mutation rate , since all such mutations should be produced directly by mutation in the last replication round . As in Cuevas et al . [13] , we used non-sense ( stop codon ) mutations as a proxy for lethal mutations . The concept is to count all stop codons in the data set and to divide this by the number of mutation targets ( non-sense mutation targets – NSMT ) , i . e . codons that by a single mutation could generate a stop codon ( see Text S1 for details ) . For these analyses , we were able to use all 17 patients in our cohort , thus expanding our data set . In total we had 898 , 335 NSMTs and 13 stop codons in the over 1×107 bases sequenced [23] ( Tables S1 and S2 in Text S1 ) . Surprisingly , 4 of the stop codons were identical and at the same position in 10051 at two different time points ( see Table S1 in Text S1 ) . This strongly indicates that this stop codon appeared only once in this patient , and that stop codons may not be lethal in HCV but instead complemented by intact genomes within the same cell . Thus , we counted this stop codon only once , for a total of 10 mutations leading to stop codons . A calculation identical to that proposed in [13] then shows that μ = 3 . 2×10−5 per nucleotide per replication , which is fully consistent with our estimate above . We also propose an improved way to calculate this rate from the same data ( see Text S1 ) , and with this method obtain μ = 2 . 8×10−5 ( binomial 95% CI: 1 . 4–5 . 2×10−5 ) . Altogether , these data and analyses indicate that HCV sequences diversify early in infection , during the exponential increase of viral load , which is then followed by a plateau in diversity for up to a few weeks . The mutation rate needed to explain these observations ( μ≈2 . 5–3 . 2×10−5 per nucleotide per replication , Figure 3D ) is 5 and 100 times smaller than previously reported for HCV [13] and its purified RdRp [12] , respectively . We next investigated in detail why HCV diversification appears to stop after a few weeks of infection , and what processes could break this plateau in diversity , since in chronic HCV infection the virus is much more diverse [23] . In particular , we analyzed the effect of turnover of replication complexes and the emergence of the cytolytic immune response . In the baseline simulations of the model , we neglected the turnover of replication complexes ( RC ) . However , RC may degrade . In this case , to sustain viral replication , the RC would need to be continuously produced to balance their degradation . Thus , we next analyzed the impact on our model predictions of including RC degradation . For fast RC turnover ( e . g . , half-life 1 . 5 d ) , most ( median of 59% ) of the simulated infections die out , and those that lead to sustained infection show a slow growth of the virus that is not compatible with the data ( Figure 4A , left panel ) . It is possible to recover fast viral growth rates , if one postulates that a larger fraction of newly synthesized RNA is used to form new replication complexes ( i . e . , if k is smaller ) . When the turnover of replication complexes is not negligible ( t½<5 days ) , on the time scale of our simulations , the accumulation of mutations is faster at later times as replacement of replication complexes occurs ( Figure 4A , right panel ) . In this case , to describe the data a smaller mutation rate would be needed , at least in some patients . Importantly , turnover of replication complexes also implies a continued increase in diversity throughout the observation period , since more ( − ) strand RNA needs to be made and hence there is more opportunity for mutations to occur . However , such a continued increase in diversity is not seen for subjects 10012 , 10017 , 10021 and 10025 . On the other hand , this process could help explain the marked increase in diversity seen at late time points in subjects 10029 , 10062 and 9055 . Note however that even these subjects seem to have a stabilization of diversity prior to this marked increase , which is not compatible with fast turnover of replication complexes . If the turnover of replication complexes is much slower ( eg . , ∼15-day half-life ) then the profiles do not differ from our baseline case where there is no turnover over the 50 day period studied . Here we studied RC turnover inside the cell , but it is also possible that cells die due to the immune response against HCV , thus forcing re-generation of RC . Thus , we next considered the effects of cell turnover on the results of our model . An effect of immune processes is removal of infected cells . Because there may be some limit to the number of infected cells in the liver [42] , the death of infected cells may allow new cells to be infected , which in turn generates new RC and the opportunity for mutation accumulation . For all subjects for whom there is enough data , we find a stabilization of diversity , which in a few cases is then followed by a “sudden" marked increase at a later time point ( 10029 , 10062 , 9055 ) . It could be that this latter pattern is an artifact of sampling . For example for 10062 , the observed fraction of unmutated sequences at the three time points sampled have confidence intervals that overlap , and those fractions are not significantly different , p = 0 . 07 ( Figure 2C , overlap of vertical bars ) . In our model this stabilization in diversity accumulation occurs because a steady-state is attained for the numbers of replication complexes and infected cells , without continued turnover . Rather than a sampling issue , it is possible that the observed increase in diversity is due to an immune response emerging at late time points , which leads to an increase of the infected cell death rate ( δ ) . Indeed , this is indicated in studies of experimental infection of chimpanzees , where the immune response is delayed several weeks [31] , [36] . In this context , an alternative explanation for the increase in diversity in 10062 is the appearance of such an immune response as suggested by the increase in ALT in this subject ( Figure 1B ) . To study the effect of a late immune response that kills infected cells , we allowed for this process starting at 30 days post infection ( Figure 4B ) . As expected , the emergence of an immune response lowers the viral load , possibly leading to a new lower viral load steady state , as is observed in some experimentally infected chimpanzees [31] . With the loss of infected cells , new cycles of infection occur along with creation of new replication complexes , and the model predicts a renewed increase in the accumulation of diversity , which mimics the data in some subjects ( eg . , 10062 , 9055 ) . However , we do not have enough data to precisely estimate the timing and magnitude of this immune response .
We analyzed the viral dynamics and viral diversification of HCV very early in acute infection . The early diversity of HCV is very low , and the inter-sequence Hamming distances follow a Poisson distribution , as would be expected when the mutations occur approximately at the same rate at all positions and the sequences are not selected for diversity [27] , [44] . Given this observation , the number of mutations at early times should depend on the time since infection , the mutation rate and the biology of viral replication . This idea has been used before in the context of primary HIV infection to estimate the time of infection , assuming a given mutation rate [26] , [27] . In the present study , the time of infection is known to within a short time window , with the first HCV positive sample within 5 days of the last negative sample . With this information , we could use our data to estimate the in vivo HCV mutation rate . By developing a model of HCV replication that takes into account the details of the viral lifecycle , we found the estimated mutation rate varied among subjects between 1 . 6×10−5–6 . 2×10−5 mutations per nucleotide per replication cycle , with a median of 2 . 5×10−5 ( Table 1 , 5 h genome ) . This estimate was very robust to different assumptions about model parameter values ( see Text S1 ) . Moreover , we systematically made conservative assumptions for the less well known parameter values leading to higher estimates for the mutation rate . To further confirm our results , we estimated the mutation rate by a completely different approach based on the frequency of stop codons ( non-sense mutations ) , corrected by the number of non-sense mutation targets , as proposed by Cuevas et al . [13] . With this calculation we obtained a mutation rate of 2 . 8×10−5 or 3 . 2×10−5 mutations per nucleotide per replication cycle depending on the calculation method ( see Text S1 ) , which is consistent with the estimate from our more complex dynamical model and substantially less than the rate ( ∼10−4 ) estimated by Cuevas et al . [13] . A likely explanation for the difference between the findings of our nonsense mutation analysis and that of Cuevas et al . is that in our study Taq polymerase errors are eliminated from the finished sequences by the SGA-direct amplicon sequencing method and thus do not enter in the error rate calculations; this was not the case for the previous analyses [6] , [13] . We further note that estimates of the HCV mutation rate based on nonsense mutations are likely to be overestimates since we found that stop codons were not always lethal ( see Text S1 ) . One explanation for this observation is that there are multiple HCV RNAs in an infected cell and another RNA may complement nonsense mutations . Indeed , we also found a case of a chronically infected patient who has a strain with a large deletion replicating in plasma at multiple time points [23] . Moreover , for dengue virus ( in the same Flaviviridae family of HCV ) there is a report of a viral strain with a stop codon that spread and attained a high frequency in the population , implying replication in both humans and mosquitoes [50] . In addition , our analysis does not account for mutational errors resulting from the cDNA synthesis step of the sequencing process , which again may lead to an overestimation of the mutation rate . However , we used Superscript IIITM Reverse Transcriptase ( Cat . No . 18080-093 , 2000 units , Invitrogen Life Technologies , Carlsbad , CA ) that has been reported to have an error rate of ∼2×10−6 mutations/nucleotide/replication [23] , [51] , which is at least 10-fold lower than our HCV mutation rate estimates , and hence should not significantly influence our estimates . Our estimates of the mutation rate for the HCV RdRp of ∼2 . 5×10−5 are notable because previous reports have suggested that the in vivo mutation rate of HCV is of the order of 10−4 mutations per nucleotide per replication [13]; and that the in vitro rate of the isolated RdRp could be as high as 10−3 [12] . One possible explanation for the latter discrepancy is that the mutation rates observed with purified RdRp enzymes are generally larger than those seen in vivo , because in vitro analyses cannot recapitulate the intracellular milieu of the replication or polymerase complex . For example , in the case of HIV reverse transcriptase , the errors measured with purified enzyme were found to be up to 20-fold higher than those measured in infected cells [52] . Another possibility is that we may have missed some low prevalence strains . However , a detailed power calculation shows that with the number of sequences obtained per patient , we would only miss strains that are present at very low levels , below 2% [23] , which is much better than was possible before [25] , [53] ( see Li et al . [23] for a detailed discussion ) . Moreover , for the dynamical model we follow time courses and analyzed the fraction of virus identical to the T/F virus; and for the stop codon analyses , we corrected for the mutational targets . Both of these lower the impact of missing strains . Given the low level of diversity observed in early infection and the relatively low mutation rate , the enormous diversity of HCV [14] , [15] , [18] and its high substitution rate ( i . e . , substitutions/site/year ) have to be understood in light of HCV's replication mechanism [16] . Relatively long-lived infected cells , with multiple replication complexes allow for the accumulation of diversity in the virions produced . At the same time , the turnover of both replication complexes and infected cells , which must surely ensue as the immune response develops , allows for renewed generation of diversity throughout the course of infection ( compare 10062 in Figure 1B and Figure 2C ) . Indeed , it could be that these details of the life cycle are responsible for the large diversity of HCV . We note that HIV and influenza , which are thought to have similar mutation rates to the one estimated here [6] , [52] , also have high substitution rates [54] . In this context , we see that accumulation of diversity is not only dependent on mutation rate , but also to a great extent on the particular processes of the viral life cycle [7] , [8] , [16] . Clearly , the pressure of the immune response , once established , will be important in determining relative fitness of many of the mutations and in determining the spectrum of mutations observed . That we see only scarce evidence of positive selection in our dataset indicates that there is a window of several weeks before the effects of the immune response can be detected . Another important parameter that we estimated was the fraction of infected cells during the early plateau in viral load , which ranged between 1 . 7% and 22% of hepatocytes . This fraction is in reasonable agreement with other studies of HCV [41] , [42] . In our model , this fraction depends on the value assumed for the maximum number of replication complexes ( RCM ) . The larger the number of replication complexes in an infected cell , the more viruses this cell can produce per unit of time , and thus the fewer the number of infected cells needed to maintain a given steady state viral load . However , increasing RCM has little effect on our estimate of the mutation rate ( see Text S1 ) . In this study , we constructed a simple model of HCV replication that tried to capture the most salient features of the viral life cycle . Moreover , we were careful to choose parameters consistent with the literature a priori , so that only 2 parameters had to be adjusted to fit the data on viral growth and diversity increase . We tested variation in the model assumptions and found that the results were quite robust . Still , it is clear that many complexities could be added to the model . For example , instead of having a fixed RCM , we could allow it to vary from cell to cell and possibly even from time to time; or we could allow for a distribution of generation times for RNA synthesis . These and other processes are easy to include in the model , however we opted to keep to the essential aspects of the replication process , so that we did not have to make further assumptions , which would complicate the interpretation of the results . In essence , this is akin to choosing a simple experimental system that is amenable to easy manipulation and interpretation of results , even if it does not represent fully all the details of in vivo system . Altogether , the unique dataset presented here , including HCV viral kinetics and genomic diversification very early in infection , revealed that the initial exponential expansion of HCV RNA is followed by a plateau in viral load that lasts up to a few weeks [30] . The initial viral expansion is accompanied by a fast early increase in sequence diversity , whereas during the viral plateau viral diversity remains approximately constant . During the plateau viral production continues but is simply balanced by the rate of viral clearance . In order to understand why viral diversity did not continue to increase during this period , we develop a novel stochastic model of HCV infection . The basic idea behind the model is that during the early exponential expansion of the virus , new cells are being infected and generating multiple replication complexes in each infected cell . This involves multiple copying events of ( + ) RNA to ( − ) RNA to ( + ) RNA , etc , with errors potentially being generated at each stage . We postulate that once the viral plateau is reached a stable population of long-lived infected cells has been generated which then produce the plateau virus without any need for new RC generation . If no new replication templates are made then there is little opportunity for mutations to accumulate , though each virus can still mutate in relation to its parent RC due to the ( − ) RNA to ( + ) RNA copying event . We found that our model , based on this idea , agreed with both the viral load kinetic data and the sequence diversity data if we assumed that the in vivo mutation rate of HCV is ∼2 . 5×10−5 per nucleotide per replication cycle . This is about 5-fold lower than previously reported , but still high enough that coupled with the long-lasting nature of HCV infection and the very high turnover of virus in chronic infection leads to substantial HCV diversity in an individual and in the population .
Plasma samples were obtained from seventeen regular source plasma donors , who became HCV infected during periods of twice-weekly plasma donations . The donors were untreated and asymptomatic throughout the collection period . All subjects gave written , informed consent and the study protocols were approved by institutional review boards at the University of Pennsylvania , the University of Alabama at Birmingham and Duke University . HCV RNA and antibodies were analyzed as described elsewhere [23] . Single genome amplification ( SGA ) followed by direct amplicon sequencing was performed on sequential plasma vRNA samples ( i . e . , ( + ) RNA strands ) , as described in detail elsewhere [23] . For our dynamical analyses , we selected subjects who had at least two time points sampled with single genome amplification assays [23] . Thus , three subjects were not included – 6213 , 6222 , 10004 . Six subjects ( 10002 , 10003 , 10016 , 10020 , 10029 , 106889 ) had more than 7 putative T/F viruses , which makes a diversification analysis impractical , both due to the complexity of the viral species in the subjects and the small number of sequences representing each lineage [23] . The exception was 10029 , who had a dominant lineage with more than 38 sequences for each time point , and we included this subject in our analyses . Thus , there were 9 subjects who were sampled at multiple time points and who had a clearly dominant putative T/F virus lineage [23] . Here we only analyzed these dominant lineages , for which we have the most data ( SGA sequences ) . Sequence alignments were initially made with ClustalW and then checked individually using JalView 2 . 6 . 1 ( www . jalview . org ) . We used ConsensusMaker ( www . HIV . lanl . gov ) to calculate the consensus of the first set of sequences sampled by SGA , which is the putative T/F virus [23] . The set of sequences from each SGA sample with the corresponding consensus was analyzed by PoissonFitter ( www . HIV . lanl . gov ) to calculate for each sequence the number of mutations away ( i . e . , Hamming distance ) from the T/F , and to test whether sequence diversification conforms to a star-phylogeny and if the set of inter-sequence Hamming distances follow a Poisson distribution [44] . Altogether we analyzed time courses of thousands of sequences with over 11 . 9 million base pairs and 1887 mutations [23] . To analyze the process of replication of HCV and how it affects the generation of diversity in primary infection , we developed an agent based model of HCV infection and replication . We assumed cells are infected by a single virion , and that in every infected cell , on average a fraction k of newly synthesized viral RNA ( vRNA ) is exported in new virions , and the rest , 1-k , form new replication complexes in the cell . These processes continue until the cell contains a maximum number of replication complexes ( RCM ) . We assume this maximum value is set by the availability of host factors . After a virus is exported , a fraction 1-θ of released virions are cleared from circulation , and the rest , θ , infect new cells . As the vRNA is copied , errors in the incorporation of nucleotides are possible . Every time a mutation occurs , there is a probability that this mutation is lethal , implying a virus or replication complex made using such mutant vRNA is non-viable . Sanjuan [48] estimates that the fraction of random mutations that are lethal is about 40% for RNA viruses . We assumed HCV is noncytolytic [32] . Thus , infected cells can produce virus for long periods of time – until the infected cell dies , either from natural death or immune attack . Early in acute infection there is little evidence of cytotoxic T cell activity and CD8+ T cells do not appear to enter the liver until many weeks after infection [36] . In addition , normal hepatocytes live for months [39] to a year or more [38] , thus , we either totally neglect death of infected cells or allow death after the first few weeks of infection . The assumption of no early death is consistent with the normal levels of alanine aminotransferase ( ALT ) measured in these individuals ( Figure 1B ) and the viral load profiles , where viral load increases rapidly to a maximum level and then stays at that level for some time . ( This is in stark contrast for example with HIV , a cytolytic virus , where a clear peak in viral load is seen during primary infection followed by a decrease in viral load [28] . ) Replication of the RNA and formation of a new virion or replication complexes is not instantaneous , as it takes a certain amount of time for synthesis of the different molecular components and their assembly . Although this time is most likely variable from replication cycle to replication cycle and from cell to cell , we assume that it is similar for all replication events in our model , fixing it at an average time to complete all the replication steps . This time we call the “generation time" . Most likely it will take longer to produce the first copied RNA upon cell infection than later ones , as various molecular events need to occur before virus production begins ( eg . , uncoating , polyprotein synthesis and cleavage , assembly of the replication complex , etc ) [5] . In the simulation , based on the assumptions described above , we follow the number , age ( in the sense of generations ) and mutational burden of each virion and each replication complex inside infected cells . The simulation was implemented in the R language ( www . r-project . org ) . Because these are stochastic simulations , there is variability from one run to the next , even when all parameters remain the same . Thus , for each patient and each set of parameters ( in Figures 1–4 ) we present results from 100 runs . Including more runs ( we tested some cases with 200 runs ) does not significantly alter the results presented . The parameters of the stochastic model are as follows: | Hepatitis C virus ( HCV ) is a RNA virus that infects over 170 million people across the world . It leads to a chronic infection in the majority of people who are infected ( >70% ) . Most people only discover that they are infected long after initial infection . Thus , it is difficult to study the very early events in infection . Here we study 17 individuals during the earliest possible stages of infection , from before the virus is detectable in the plasma to around 35 days post-infection . We focus on understanding the viral kinetics and the diversification of HCV during this acute phase of infection . During chronic infection HCV is present in the host as a swarm of multiple variants generated by its error prone copying . We studied the early diversification of HCV during acute infection using a new mathematical model of HCV replication . We found that after a phase of fast increase in viral load , accompanied by viral diversification , there is a stabilization of viral load and diversity levels . Using our model , we were able to estimate for the first time the HCV mutation rate during acute infection . We estimated the median in vivo viral mutation rate is 2 . 5×10−5 mutations per nucleotide per genome replication ( range 1 . 6–6 . 2×10−5 ) , about 5-fold lower than previous estimates . We also used a different approach , based on results of classical genetics , to calculate HCV's mutation rate and obtained consistent results ( 2 . 8–3 . 2×10−5 ) . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"hepatitis",
"c",
"medicine",
"infectious",
"diseases",
"infectious",
"hepatitis",
"hepatitis",
"gastroenterology",
"and",
"hepatology",
"viral",
"diseases",
"infectious",
"disease",
"modeling",
"liver",
"diseases"
] | 2012 | Quantifying the Diversification of Hepatitis C Virus (HCV) during Primary Infection: Estimates of the In Vivo Mutation Rate |
Why gene order is conserved over long evolutionary timespans remains elusive . A common interpretation is that gene order conservation might reflect the existence of functional constraints that are important for organismal performance . Alteration of the integrity of genomic regions , and therefore of those constraints , would result in detrimental effects . This notion seems especially plausible in those genomes that can easily accommodate gene reshuffling via chromosomal inversions since genomic regions free of constraints are likely to have been disrupted in one or more lineages . Nevertheless , no empirical test has been performed to this notion . Here , we disrupt one of the largest conserved genomic regions of the Drosophila genome by chromosome engineering and examine the phenotypic consequences derived from such disruption . The targeted region exhibits multiple patterns of functional enrichment suggestive of the presence of constraints . The carriers of the disrupted collinear block show no defects in their viability , fertility , and parameters of general homeostasis , although their odorant perception is altered . This change in odorant perception does not correlate with modifications of the level of expression and sex bias of the genes within the genomic region disrupted . Our results indicate that even in highly rearranged genomes , like those of Diptera , unusually high levels of gene order conservation cannot be systematically attributed to functional constraints , which raises the possibility that other mechanisms can be in place and therefore the underpinnings of the maintenance of gene organization might be more diverse than previously thought .
Collinearity conservation , i . e . conservation of local gene order , across distantly related phyla is often viewed as the result of functional constraints that prevent the occurrence of breaks of chromosomal rearrangements during evolution . The nature of these constraints is still poorly understood . They may merely reflect the presence of yet-to-be annotated protein and nonprotein-coding genes in intergenic regions [1]–[3] . A second type of constraints may be linked to the existence of regulatory domains , i . e . genomic regions associated with complex regulatory inputs . These regulatory domains can adopt at least two forms . One common form of complex regulatory inputs corresponds to genes that are coordinately expressed or repressed [4]–[7] . Alternatively , regulatory constraints can adopt the form of long-range regulatory interactions , which often involve the interdigitation of cis-regulatory sequences with genes that are not their targets [8]–[10] . These regulatory domains are enriched for noncoding DNA ( highly conserved noncoding elements or HCNEs ) with putative regulatory potential [11]–[16] . These HCNEs tend to be found in the vicinity of protein-coding genes that participate in key processes during development , such as regulation of gene expression and signal transduction [10] , [12] , [17] . Disruption of genomic regions under constraints can be accompanied by alteration of gene activity , as illustrated by chromosomal rearrangements that modify gene expression as a result of the separation of a gene from its regulatory sequences [18]–[20] . These alterations in gene activity may have a detrimental effect , which would lead to conservation of gene organization [19] , [21] . Highly rearranged genomes , such as those of the Diptera , are especially suitable for analyzing the presence of regulatory-based constraints that preserve collinearity since regions free of them are likely to have been disrupted in one or more lineages . Gene order comparisons have helped to delineate collinear blocks [22] across nine Drosophila species that represent ∼380 million years ( myr ) of total divergence time [23] , [24] . A minimum of ∼6 , 100 chromosomal breakpoints have been estimated to have occurred in 2 , 688 intergenic regions [22] scattered across the main chromosomal elements ( the so-called Muller's elements A–E ) that constitute the Drosophila genome [25] , [26] . The analysis of the expected patterns of evolution of gene organization under different evolutionary scenarios indicate that fragile regions , i . e . those that accumulate most chromosomal breaks during evolution [27] , are the main factor that explains the patterns of gene arrangement across Drosophila species [22] . Constraints nevertheless may be responsible for ∼15% of gene order conservation and their presence is positively correlated with the size of the collinear block [22] . The top 1% largest collinear blocks , or ultraconserved regions [22] , are enriched for genes associated with particular expression profiles , but the functional signature most prominently found in ultraconserved regions are stretches of DNA with multiple HCNEs ( 14 . 5% of the 145 HCNE peaks mapped as compared to 6 . 7% expected ) . To our knowledge , only two empirical tests for the presence of functional constraints have been performed in eukaryotic genomes [20] , [28] . In both cases , engineered chromosomal inversions were induced to disrupt clusters of genes with shared expression attributes and , subsequently , the phenotypic consequences of such disruptions monitored . For example , the disruption of the mouse Hoxd cluster , which is conserved across vertebrates , helped determine the presence of two functional subdomains and two long-range enhancers on either side of the cluster . This functional organization of the Hoxd cluster was proposed as the underlying cause that kept the cluster intact during the evolution of vertebrate lineages . In Diptera , three gene neighborhoods expressed in testes and one gene neighborhood expressed during embryogenesis of D . melanogaster have been disrupted [28] , but no modification of the expression of the genes included in the neighborhoods examined was detected . These neighborhoods are conserved within the D . melanogaster species subgroup but disrupted in some other Drosophila lineages ( Table S1 ) , which does not clarify the phylogenetic scope of the putative constraints tested . The phenotypic consequences of disrupting a collinear region conserved throughout the genus Drosophila , which is presumably maintained by constraints , are presently unknown . Here , we used chromosome engineering to disrupt an ultraconserved region located on Muller's element C of the Drosophila genome ( Figure 1A ) . This ultraconserved region is delimited by the genes CG15121 and CG16894 and is ∼701 kb long in D . melanogaster [22] . Importantly , this ultraconserved region is , based on current information , the one with the largest number of functional signatures , suggestive of the presence of regulatory-based constraints [22] . After disrupting the ultraconserved region CG15121–CG16894 , we examined the resulting phenotypic consequences both by performing a variety of genetic and competition experiments that tested for differences in viability , fertility , and relevant parameters of general homeostasis and by characterizing levels of mRNA abundance in both sexes . Our results indicate that , contrary to the prevalent view , the extraordinary conservation of some of the largest collinear blocks in eukaryotic genomes might not necessarily derive only from functional constraints that result in severe detrimental effects .
The ultraconserved region CG15121–CG16894 ranked first in length and eleventh in the number of genes encompassed among 2 , 683 regions of conserved collinearity across nine Drosophila species [22] . Data from another comparative analysis on gene organization in the genus Drosophila [29] are consistent with the overall maintenance of the collinearity in this genomic region . In addition , this region shows statistically significant enrichment for genes encoding proteins involved in chemosensory perception and for genes preferentially expressed in males , many of them showing this same trend across multiple Drosophila species ( Figure 1B and Table S2 ) . This higher-than-expected local density of genes with coherent patterns of expression supports the presence of a male-biased gene expression neighborhood , which is intertwined with a smaller chemosensory perception gene neighborhood . This ultraconserved region is spanned by four HCNE peaks [12] ( Figure 1C ) , more than any other collinear block . Genes responsive to HCNEs have been postulated to be preferentially associated with a particular kind of core promoters . Specifically , using promoter predictions for 42% of the protein-coding genes of D . melanogaster , a significant overrepresentation of genes with some kind of Inr-motif ( Inr only , Inr/DPE , or Inr/TATA ) was found in HCNE-dense regions [12] . We screened 500 nt upstream of each protein-coding gene in the region under study using McPromoter and obtained reliable predictions for eight genes . Six of these genes were predicted to have a core promoter responsive to HCNEs . These genes are found scattered along the region ( Figure 1B and Figure S1 ) and encode proteins involved in the regulation of gene expression , from mRNA binding proteins ( sm ) , to signal transducers ( Toll-7 , 18-w ) , to ubiquitins ( Isopeptidase-T-3 ) , and also to others that we did not predict a priori ( Obp56a and CG8654 ) . Given the large number of genes for which a prediction was not provided by McPromoter , we checked for the presence of genes that are expressed during key developmental stages using the expression profile characterization generated by modENCODE [2] . Fifteen out of 36 protein-coding genes show a preferential pattern of expression during embryogenesis; among them , there is a cluster of eight genes mostly displaying high levels of expression during the first 16 h of development and moderate expression during the larva-pupa transition ( Figure S1 ) . Six of these eight genes are predicted to have core promoters of the Inr-motif type and four are associated with lethal phenotypes , the latter underscoring their functional relevance prior to imago emergence ( Figure S1 ) [30] , [31] . Together , these functional features associated with some of the genes in the region are suggestive of regulation by HCNEs . Examination of the organization of the ultraconserved region CG15121–CG16894 in Anopheles gambiae revealed the presence of orthologues in six different locations ( Figure S2A and Table S3 ) . This degree of dispersion is not surprising given the fast differentiation of the Drosophila and Anopheles genomes via the accumulation of chromosomal rearrangements [32]–[34] and the lack of common constraints reflected in the pronounced differences in development , morphology , and ecology of these two Diptera [33] , [35] , [36] . Nevertheless , we detected a conserved gene arrangement including the gene Toll-7 and the Obp genes , an association that is also present in A . aegypti ( Figure S2B ) . Phylogenetic analyses [37] unambiguously support the close phylogenetic relationship among several Obp genes in the ultraconserved region under study ( Obp56a , Obp56d , and Obp56e ) and some that are adjacent to Toll-7 in A . gambiae ( OBP23 , OBP25 , OBP26 , OBP28 ) , which in turn are closely related to those that cluster nearby Toll-7 in A . aegypti ( [38] and this work; Figure S2C ) . Interestingly , the genes Obp56a , Obp56d and Toll-7 are found within the same expression cluster in D . melanogaster and two of them have core promoter types presumably responsive to HCNE-mediated regulation ( Figure S1 ) . To sum up , the existence of two intertwined gene neighborhoods associated with very marked expression profiles , the enrichment for HCNE peaks and presence of their putative targets , and the detection of a region conserved across Diptera reinforces the possibility that one or more regulatory-based constraints might exist in the ultraconserved region CG15121–CG16894 . To assess the importance of the integrity of the ultraconserved region CG15121–CG16894 , we aimed to disrupt it and characterize the resulting phenotypic effects . We examined the existence of stocks carrying isolated naturally occurring inversions disrupting the ultraconserved region and none was found . Thus , we generated a disruption of the ultraconserved region CG15121–CG16894 by inducing the inversion In ( 2R ) 51F11-56E2 in D . melanogaster , a species in which nonallelic homologous recombination ( NAHR ) events can be mediated between FLP recombination target-bearing transposable elements ( FRT-bearing TEs hereafter ) via activation of a heat-inducible flippase-recombinase [28] , [39]–[42] ( Figure S3 , S4 , S5 ) . For that , we used two TEs bearing FRT sites in opposite orientation: P{RS5}5-HA-1995 , which is inside the ultraconserved region CG15121–CG16894; and P{RS3}CB-0236-3 , which is located 4 . 35 Mb upstream ( Figure 1A ) . The outer element is virtually terminal within a collinear block of 15 genes and is inserted into a naturally occurring copy of the TE 1360{}835 [43] . We adopted several measures to avoid confounding effects that could overlay those of the intended disruption of the ultraconserved region CG15121–CG16894 . First , TEs were selected to avoid disrupting any known regulatory sequences of flanking genes and those presumably embedded in HCNE peaks , thus preventing the generation of artifactual position effects . The comparison of the size and sex ratio of the progeny of flies homozygous for each of the TEs confirmed the absence of any obvious detrimental effect associated with particular TE insertions ( Figure S6 ) . Second , in addition to strains carrying the inversion In ( 2R ) 51F11-56E2 ( INV1 and INV2 ) , we generated several control strains to account for further mutations that could have been incidentally generated by our approach [44] . Specifically , three kinds of control strains were constructed: strains carrying two FRT-bearing TEs in cis ( REC ) , i . e . just before inducing the NAHR event that mediates the inversion; strains carrying the standard arrangement as a result of failed induced NAHR events but that were exposed to the same experimental conditions as the INV strains -SIMultaneous controls- ( SIM1 , SIM2 , and SIM3 ) ; and strains in which the inverted segment is reverted back to its original orientation -REVertant controls- ( REV1 and REV2 ) ( Figure 2A , Figure S3 and S4 ) . The main molecular changes at the inversion breakpoints of all relevant strains , plus those carrying the original FRT-bearing TEs , are depicted in Figure S5 . In the absence of any secondary effect of our procedure , SIM , REV , and REC should perform likewise as measured by the size and sex ratio in the progeny of low-density crosses with homozygous flies ( Figure S7 ) . INV strains for which no SIM and/or REV control lines could be generated were discarded . For the remaining strains , their putative karyotype was verified at the cytological level and the expected molecular organization at their breakpoint regions confirmed by PCR and Sanger sequencing ( Figure 2B–2C and 8 ) . Further expression profiling confirmed the absence of local position effects at the inversion breakpoints ( see below ) . Lastly , engineered chromosomes were maintained in homozygosis thus preventing the otherwise unavoidable accumulation of detrimental mutations if kept in heterozygosis over a balancer chromosome . Most of the strains that were not discarded were included in one or more downstream analyses . The disruption of the ultraconserved region CG15121–CG16894 did not lead to viability impairment in the progeny of the carriers ( INV1 , INV2 ) as compared to that of non-carriers based on two proxies examined ( Figure S7 ) . Cursory examination of embryos , larva , and pupa did not detect any obvious morphological defect either . We then explored the possibility of a detrimental effect in heterozygous condition before reaching adult eclosion , either because of the disruption of the collinear block , meiotic distortion due to the presence of a chromosomal inverted rearrangement , or both . The comparison of different chromosome combinations revealed no departures from the expected Mendelian ratios ( 1∶2∶1 ) and absence of sex-specific effects ( Figure S9 ) . Further , we assessed differential viability during early stages ( i . e . prior to imago emergence ) , when competition among individuals is specially intense [45] , [46] , and when most of the genes putatively targeted by HCNEs exhibit high levels of expression . Differences among carriers of different 2R chromosomes were detected in frequency-dependent competition experiments involving different pairwise combinations of embryos , but those differences were not consistently shown by the carriers of the disrupted ultraconserved region ( INV1 , INV2 ) compared to those that carry it in its intact form ( REC ) across different genetic backgrounds ( Figure S10 ) . Together , these results do not point out to any obvious detrimental effect on viability , and therefore on stages of the life cycle that encompass key developmental transitions , as a result of disrupting the ultraconserved region CG15121–CG16894 . Next , we examined the effect of disrupting the ultraconserved region CG15121–CG16894 after imago emergence since many of the protein-coding genes included in the region under study are expressed during adulthood , often in a sex-dependent fashion ( Figure S1 ) . Specifically , the region under study is populated with male-biased genes in expression , 11 of them preferentially expressed in testes [22] and four , all of them Obp genes , present in the seminal fluid ( Figure 1; Table S2 ) . We tested for differences in several parameters of male fertility: progeny size , sperm performance , and mating ability . The comparison of males from the strains INV1 , INV2 , and REC , which are all red-eyed so that differences due to differential pigmentation can be factored out , revealed that although there are differences in progeny size ( Figure S11A ) , strains INV1 and INV2 are more different from each other than either of them is to strain REC . Double-mating experiments did not reveal any substantial difference in sperm performance ( Figure S11B ) and mating ability ( Figure S12 ) . Further , given the presence of nine odorant-binding and one odorant-receptor protein-coding genes in the ultraconserved region CG15121–CG16894 , we analyzed the odorant abilities of the different strains . We examined the response to three volatile compounds ( ethanol , acetone , and benzaldehyde ) and found statistically significant differences in five out of 12 sex-by-strain combinations . Importantly , for females exposed to ethanol at a concentration of 10−3 and males exposed to acetone at a concentration of 10−4 . 5 , strains INV1 and INV2 exhibited a coherent pattern of differentiation from strain REC ( Figure 3; Table S14 ) , involving in all cases an attenuated attraction to the chemical in question . This attenuated response does not result from an overall impairment of the odorant abilities of the flies as shown by the response to the repellent compound benzaldehyde . Analyses of different proxies for the general adult homeostasis ( negative gravitaxis , heat-shock resistance , desiccation resistance , and starvation resistance ) did not uncover any other difference between the strains with and without the disrupted ultraconserved region CG15121–CG16894 ( Figure S13 ) . Lastly , we tested whether the disruption of the ultraconserved region CG15121–CG16894 resulted in a perturbation of gene expression . We performed a microarray-based characterization of the transcriptome of six lines ( REC , INV1 , INV2 , SIM1 , REV1 , and REV2 ) during adulthood , the stage in which we found evidence of phenotypic differences associated with the disruption generated . At FDR 0 . 01 , we found a very limited number of differentially expressed transcripts both in males and females ( 0 . 07% -11/16 , 637- and 6 . 2% -1 , 033/16 , 637- , respectively; Dataset S1 and Table S16 ) . Further analyses confirmed the similarity of the expression profiles between equivalent strains generated by our procedure ( first three planned contrasts in Table S17; Datasets S1 and S2 ) . Likewise , these analyses indicated that the most statistically significant differences in mRNA abundance found are associated with differences in pigmentation ( last three planned contrasts in Table S17; Dataset S2 ) , in good agreement with the clustering of expression profiles among strains ( Figure 4 ) . The inspection of the chromosomal distribution of the differences in gene expression showed that a few of them were related to genes in the ultraconserved region . However , these alterations in mRNA levels are in fact the result of pigmentation differences , given that these alterations were found invariably between red-eyed and white-eyed strains , regardless of whether the former carry the disrupted ( INV1 and INV2 ) or the intact form ( REC ) of the ultraconserved region ( Figures S14 and S15 ) . Searches for biologically coherent patterns among differentially expressed genes indicated that , for example for females , statistically significant enrichment was found for functional classes related to perception of visual stimuli ( Table S18 ) . Further , patterns of sex bias in gene expression were not affected either ( Figure S16; Dataset S3 ) . Hence , no discernible effect on the levels of mRNA of genes both inside and outside of the ultraconserved region CG15121–CG16894 was detected as a result of its disruption .
The interplay among organization , function , and evolution of eukaryotic chromosomes is still poorly understood . Collinearity conservation is an important genomic feature that can conflate all three aspects , especially in genomes characterized by their ease in accommodating structural variation . Up to date , no empirical evaluation has been performed on the effects of disrupting the integrity of one of the largest genomic regions whose overall gene organization has been preserved over a large time scale . Importantly , this conservation entails species with very different behaviors and ecologies and therefore subject to very different selective pressures [46] , [47] . Our results show an absence of detrimental effects on the carriers of the disrupted ultraconserved region for a variety of traits associated with viability and fertility . Although the absence of detrimental phenotypic effects could result from a limited ability to detect differences as statistically significant , the comparison of actual and estimated ideal samples sizes shows that this potential limitation could be the explanation in only a few cases ( Table S19 ) . Therefore , at least in Diptera genomes , our results show that an unusually high degree of collinearity conservation coupled with enrichment for functional coherent patterns is not necessarily associated with severe detrimental effects upon perturbation [20] , [48] . The ultraconserved region CG15121–CG16894 harbors at least three sets of protein-coding genes that may be associated with regulatory-based constraints . The genes fall into the following broad categories: detection of chemical stimuli; sperm manufacturing and performance; and developmental processes . The only phenotype detected in association with our disruption was the more attenuated odorant response to attractant volatile compounds . Although variation in olfactory response occurs both within and between species [49] , [50] , it has been shown that the altered function of Obp genes , including some in the region studied , can affect fitness components in D . melanogaster [51] . Transcriptome characterization of adult flies did not uncover any obvious consequences in the activity of the Obp genes due to our disruption as measured by mRNA abundance , and magnitude and direction of sex bias . This lack of evidence for misregulation cannot rule out though that some other expression attributes , such as the spatial distribution of transcripts , had not been altered since they would have gone unnoticed by our approach . The evolutionary relevance of the maintenance of the clustering of Obp genes throughout the genus Drosophila is reinforced by the presence of an orthologous arrangement in different mosquito species to an extent not explained by chance [37] . In fact , the maintenance of the arrangement that includes a core cluster of Obp genes and the gene Toll-7 for ∼970 myr was unexpected due to the extent of rearrangement undergone by the genomes of the Diptera involved [22] , [33] , [34] ( Figure S2B ) . Regardless , the conservation of the cluster of Obp genes and Toll-7 would not explain the collinearity conservation observed elsewhere in the ultraconserved region under study . In relation to the two other biological signatures suggestive of constraints , we detected no evidence that our perturbation resulted in a detectable phenotype . The genes included in the two intertwined gene expression neighborhoods of the ultraconserved region CG15121–CG16894 do not show evidence of altered mRNA levels or malfunction that could result in impaired male fertility . Whether the constituent genes are under a tight coordinated regulation [52] that could affect the long-term stability of the region does not seem to be upheld by our results . This conclusion would be reinforced by the genus-wide lability shown by clusters of male-biased genes in expression [22] . The absence of misregulation , as measured by mRNA abundance , is consistent with the results obtained when D . melanogaster male-specific gene neighborhoods spanning over hundred of kb were disrupted [28] . Further , protein-coding genes presumably responsive to long-range regulation mediated by HCNEs during development are found scattered across the ultraconserved region CG15121–CG16894 at both sides of the disruption . The lack of a detrimental phenotype could reflect that the region is a composite of several autonomous genomic regulatory domains , each of them under the control of a particular HCNE peak . Under this scenario , our disruption would have separated different genomic regulatory domains without affecting any long-range interaction between HCNEs and their targets . Nevertheless , marked autonomy among HCNE peaks should result in certain regionalization of the expression profiles , i . e . physically closer HCNE targets should exhibit more similar expression profiles . In fact , we find putative targets of HCNEs located at both sides of the disruption , such as CG9854 and CG8896 , showing very similar expression profiles ( Figure 1 and Figure S1 ) . If this similarity denotes the existence of long-range interactions , the latter are not associated with detrimental effects that explain the collinearity conservation . Alternatively , some degree of regulatory redundancy could exist at both sides of the disruption , which would explain the lack of phenotypic effect but not the collinearity conservation . A different view of collinearity conservation entails factors other than functional constraints . DNA stretches with sufficient sequence identity can mediate NAHR events that give rise to chromosomal rearrangements [53] . Large collinear regions could be depleted of this type of sequences thus explaining why this ultraconserved region has maintained its integrity . A search for well-annotated sequences with the potential to mediate NAHR events [54] in the studied region across Drosophila species revealed that these sequences are present ( Table S20 ) . These sequences include ncRNA genes such as tRNAs [55] , [56] , rRNAs genes [57] , [58] , and snoRNAs [59] . Equivalent comparative analysis focusing on TEs , once properly annotated , will enable to test whether the region under study is particularly depleted for these sequences , which would decrease its propensity of being rearranged . Further , recent findings indicate that orthologous landmarks harboring genes that bind to the nuclear periphery are significantly larger than those that do not harbor any suggesting that particular intranuclear localizations might provide molecular environments associated with higher levels of genome stability [60] . The ultraconserved region CG15121–CG16894 is known to establish some contact with the nuclear periphery [61] . Specifically , it contains at least five protein-coding genes ( CG16716 , CG13872 , CG10822 , CG8654 , CG16898 ) that exhibit statistically significant association with the B-type Lamin protein , a key component of the inner nuclear membrane [60] , [62] . This pattern leaves open the possibility that some regulatory-based constraints had evolved under the enhanced evolutionary stability enjoyed by genomic regions associated with the nuclear periphery , and therefore functional constraints would not be the only mechanism contributing to their collinearity conservation . Regardless how extensive the characterization of individuals carrying engineered genomic regions could be , our ability to detect phenotypic effects will always be contingent to the experimental setting used and the timescale in which the assays are performed . Nevertheless , our results raise the possibility that , at least in Diptera genomes , the mechanistic basis of collinearity conservation might be much more subtle and diverse than previously thought and that regulatory-based interactions might not suffice to account for the patterns of extensive conservation seen in some genomic regions [22] . Only further empirical tests for this and other ultraconserved regions can shed light on the scope of our observations .
Table S4 describes the strains used . Fly cultures were grown and maintained on dextrose-cornmeal-yeast medium at room temperature . Unless otherwise stated , all phenotypic assays were performed at 25°C in a temperature-controlled chamber and fly manipulation , sorting , and scoring were carried out under CO2 anesthesia . Strains generated are available upon request . Strains carrying the original FRT-bearing TE insertions ( 5-HA-1996 and CB-0236-3 ) were selected from the DrosDel collection and examined at different levels prior to their use in the generation of the chromosomal inversion In ( 2R ) 51F11-56E2 . Sanger sequencing confirmed the insertion point of the FRT-bearing TEs used . Low-density crosses using homozygous flies were also performed to test for potential detrimental effects associated with the TE insertions . Briefly , six sexually mature individuals ( three four-day-old females and three two-day-old males ) were used per cross and strain; each cross was replicated five times . Males were discarded after 24 hours while females were transferred twice every three days until discarded on the ninth day . To generate the inversion In ( 2R ) 51F11-56E2 , we followed essentially [42] with slight modifications ( John Roote , pers . comm . ; Figure S3A–S3B ) . Three types of strains are generated under this procedure ( REC , INV , and SIM ) ; a fourth type of strain ( REV; Figure S3C ) was generated by reverting the inversion In ( 2R ) 51F11-56E2 [40] . The strains generated were scrutinized for evidence of side effect associated with our procedure ( see below ) . Individuals from strains deemed as valid were crossed with those from the strain w1118 , which possesses the standard arrangement for all chromosomes , and the third-instar salivary gland polytene chromosomes of the progeny examined . Chromosome squashes were stained with orcein and inspected with a Zeiss AX10 Imager M1 microscope . Cytological analysis was performed using the photographic polytene maps of D . melanogaster as a reference [63] . Diagnostic DNA stretches at the breakpoint regions were PCR amplified and their identity verified by Sanger sequencing; sequences were deposited at GenBank ( http://www . ncbi . nlm . nih . gov/genbank/; JN805541–JN805602 ) . Amplicons C , D , E , and F confirmed the presence of FRT-bearing TEs and their derivatives at the breakpoint regions . Amplicon G , or G′ after excision of one of the exons of the reporter gene , and H correlate with the absence and presence of the inversion , respectively ( Tables S5 and S6 for further details on primers used and amplicons ) . Genomic DNA used in PCR genotyping was extracted from 50 individuals of each strain as described [39] and quantified using a NanoDrop 8000 Spectrophotometer . Takara Taq and Takara Ex Taq , depending on the size of the DNA fragment to be amplified , were used according to manufacturer conditions . PCR products were resolved on 1% agarose gels and visualized in an AlphaImager HP system . Amplicon sequences were examined for point mutations and indels that could have been generated incidentally during the course of our procedure . Low-density crosses using homozygous flies were set up for each strain , as described above for the evaluation of the TE insertions , to confirm absence of differences among control strains and to test for differences among the latter and the strains carrying the disrupted ultraconserved region . Additional crosses , five per strain , evaluated the effects of the disruption in heterozygous condition with different chromosomes carrying the standard arrangement ( SIM1 , REV1 , REV2 ) ; 10 sexually mature individuals of each sex were used per cross . Further , adapted frequency-dependent experiments [64] evaluated differences in viability of mixtures of embryos for two different genotypes . Two types of mixtures were prepared . In both , w+ individuals ( INV1 , INV2 , or REC; the tested chromosome ) compete with w− individuals carrying the ultraconserved region in its intact form . The only difference between the mixtures was the condition of the tested chromosome ( heterozygosis , e . g . INV1/SIM1 versus SIM1/SIM1; homozygosis , e . g . INV1/INV1 versus SIM1/SIM1 ) . For both types of mixtures , three sets of experiments were done varying the standard arrangement of the competing embryo ( SIM1 , REV1 , REV2; the tester embryo ) , which was always in homozygosis . Embryos were collected from grape juice-sucrose-agar plates supplemented with yeast paste using FlyStuff Small Embryo Collection Cages . Using dissecting needles , 100 embryos were deposited on small grape juice-sucrose-agar cube according to one of three starting proportions for the two competing embryos ( 1∶3 , 1∶1 , 3∶1 ) . The cube was subsequently introduced into a vial of fresh medium; each assay was replicated 10 times . Separate previous pilot experiments for the strains REC , SIM1 , and INV1 , using 100 embryos , indicated that the rate of survival was approximately 40% and therefore enough to detect differences between the strains to be compared; each assay was replicated 10 times . In total , 42 different competition settings were set up ( 7 tester by tested combinations×2 genotype conditions for the tested chromosome×3 starting proportions ) . Given the number of embryos used across competition experiments ( 42 , 000 ) , only those involving the same tester w− embryo ( e . g . SIM1/SIM1 ) were performed simultaneously , which determined how the contrasts were done ( three for each genotype condition of the tested chromosome and particular starting proportion , i . e . 18 in total ) . No bias in sex ratio was assumed in all cases . Progenies from each cross were scored after 15 days to ensure the emergence of all surviving imagoes and the relative viability between the two competing genotypes estimated as ( n1′×n2 ) / ( n1×n2′ ) , where n1 and n2 , and n1′ and n2′ , are the number of embryos and imagoes , respectively , of the two competing genotypes . Values of relative viability were log2-transformed . The effects of the induced disruption on male fertility were assayed by examining progeny size , sperm performance , and mating ability . For the first test , we exposed single four-day-old virgin females carrying the ultraconserved region CG15121–CG16894 in its intact form ( SIM1 , REV1 , REV2 ) to single two-day-old males with ( INV1 , INV2 ) or without ( REC ) the disrupted ultraconserved region . Males were discarded after 24 hours while females were transferred daily to vials with fresh food over a 10-day period . The number of replicates ranged from seven to nine . To evaluate sperm performance in the carriers of the disrupted ultraconserved region , we followed a similar experimental design to that for monitoring progeny size with the exception that the females mated in the first day were exposed to single males of their own strain in subsequent days . The progeny sired by the first and second males was scored based on eye color . For those days during which progeny from both parents were detected , i . e . those oviposited roughly the same day , we estimated the fraction of red-eyed individuals ( necessarily sired by INV1 , INV2 , or REC males ) in relation to the total . To account for the effect of the order of the males used , we performed identical experiments but this time the first male possessed the same genotype as the female while the second male was from the strains INV1 , INV2 , or REC . The number of replicates for each combination of genotypes ranged from four to nine . As for the comparison of mating abilities , we exposed single two-day-old males to 10 four-day-old virgin females for different time periods ( 1 hr , 3 hr , 6 hr ) . Afterwards , the females were transferred individually into vials with fresh food . After 15 days , 10 vials were examined for the presence of progeny , which indicates that at least one successful fertilization event occurred , and the number of females successfully fertilized recorded . Ten replicates were done per strain and time period combination . The “dipstick” method was used [65] . Briefly , virgin individuals from INV and REC strains were separated by sex and transferred by aspiration in groups of five to marked empty plastic vials ( O . D . ×H: 25×95 mm ) 24 hr after emergence . The vials were marked at 3 and 6 cm from the bottom . Fisherbrand Q-tips dipped into the odorant dilutions to be tested were introduced into the vials up to the 6 cm mark , and secured with a cotton plug to avoid contact with the walls of the vial . After a 15-second recovery period , the number of flies in the bottom compartment was recorded 10 times every five seconds and the avoidance scored estimated as the average over those 10 measurements . A score of 2 . 5 indicates indifference to the odorant tested while values >2 . 5 and <2 . 5 indicate repulsion and attraction , respectively . All tests were performed between 2 and 6 pm after starving the flies for no less than four hours , and the vials were always placed sideways to prevent interfering with the geotactic response . For each strain , 10 groups of five individuals for each of the sexes were analyzed . We tested three odorants , two of them usually considered to elicit an attracting response ( ethanol , Gold Shield; acetone , Fisher Chemical ) while the other is considered to be a repellant ( benzaldehyde; Sigma Aldrich B1334 Benzaldehyde-ReagentPlus ) . Since odorant response can be concentration dependent , we tested two concentrations deferring by several orders of magnitude . These concentrations ( vol/vol ) were: ethanol , 10−3 and 10−0 . 5; acetone , 10−4 . 5 and 10−1 . 5; and benzaldehyde , 10−3 and 10−0 . 5 . Only fresh and thoroughly mixed dilutions were used to prevent oxidation , which is particularly relevant in the case of benzaldehyde . As a control , we used distilled water , which is known to attract starved flies [65] . Four proxies , including gravity response and survival to three stressors ( heat-shock , desiccation , and starvation ) , were assayed . In all cases , males and females were separated after emergence and left for 24 hours in vials with fresh food . Subsequently , all individuals were transferred to new vials by aspiration for performing the pertinent tests . Negative gravitaxis was measured essentially as reported [66] . Twenty flies per sex and strain were transferred individually into a 250 ml glass volumetric cylinder . Flies were knocked down by tapping the cylinder ten times on a pad , and the height reached by each fly in 20 s recorded using the volumetric scale , which was divided by the maximum height . In our experience , flies performed more consistently after some training , reason why only measurements from a third trial were recorded . For the heat-shock resistance test [67] , five groups of 20 flies per sex and strain were transferred in pairs into 3 ml Pyrex vials and then incubated in water baths at 35°C for 30 m ( as a pre-conditioning step ) and then at 39°C for another 30 m . The mobility of the flies was restricted to the submerged portion of the vial with cotton plugs . After the heat-shock treatment , flies were collected in vials with food that had been incubated at 25°C overnight . Next morning , the fraction of flies alive within each group was recorded . For the desiccation resistance test [68] , groups of five flies , from a total of 20 per sex and strain , were transferred into empty vials . The mobility of the flies was restricted to the lower third of the vial using foam plugs , over which 3 g of Drierite desiccant were added . Next , the vials were sealed with Parafilm to maintain low humidity . Flies were checked every hour and the elapsed time-to-death recorded . As for the starvation resistance test , 20 flies per sex and strain were transferred into glass vials containing 2 ml of 1% agar dissolved in water to ensure normal humidity conditions . Flies were examined every 12 hr and the elapsed time-to death recorded . For each phenotype , normality of the data was visually inspected using normal quantile plots and precisely determined with the Shapiro-Wilk test . Homogeneity of variances was estimated using the Levene's test . Parametric tests were used if the departure from the assumptions of normality and homoscedasticity was absent or negligible . The Welch ANOVA was used if only heterodasticity was detected . Different transformations ( log2 or arcsine square root , depending on the test ) of the measurements were calculated for some phenotypes to improve fit to normality although this had a very little effect . Alternatively , non-parametric tests ( e . g . Kruskal-Wallis ) were used . When multiple post-hoc contrasts were necessary , appropriate tests that account for multiple comparisons were used ( e . g . Steel-Dwass ) . In the case of the departure from expected Mendelian ratios in frequency-dependent competition experiments , the G-test for goodness of fit , upon applying the William's correction , was used . Statistical contrasts were performed with JMP 4 . 1 ( SAS Institute ) ; the evaluation of the sensitivity to detect statistically significant differences was done with GPower 3 . 1 . 3 when needed [69] . Three low-density crosses were set up for the strains REC , INV1 , INV2 , SIM1 , REV1 , and REV2 . The resulting virgin progeny was collected , separated by sex , and allowed to age for 5–7 days . Fifty individuals for each sex were snap frozen in liquid nitrogen at the same time of the day within a 2 hours window and subsequently stored at −80°C . Total RNA from biological samples was extracted using TRIzol Reagent ( Invitrogen ) and purified using the RNeasy Mini Kit ( Qiagen ) . Concentration , quality and integrity of the RNA samples were estimated using a NanoDrop 8000 Spectrophotometer and the RNA 6000 Nano Chip Kit ( Agilent Technologies ) with an Agilent 2100 Bioanalyzer . Ten µg of total RNA were reverse transcribed into cDNA using the SuperScript Double-Stranded cDNA Labeling Kit ( Invitrogen ) . Probe labeling , hybridization , array scanning , and data extraction were performed by Roche NimbleGen Service Group in Iceland . We used the oligonucleotide NimbleGen 12×135k D . melanogaster arrays , which contain 135 , 000 probes including different types of controls and 16 , 637 transcripts of protein-coding genes as annotated in release 5 . 7 . The experiment consisted of 36 one-color hybridizations ( 6 samples×2 sexes×3 biological replicates ) . Raw fluorescence intensity values of probe pairs were summarized for each transcript using the median value after log transformation . Subsequent data analysis was performed using the tools implemented in the online pipeline WebArrayDB [70] . Data were normalized between arrays using the scale method [71] implemented in the LIMMA package [72] . Data for males and females were analyzed separately . Statistically significant differences across strains were assessed using a fixed-effect model ANOVA and multiple testing was performed with the Benjamini-Hochberg correction [73] . Similarity in expression profiles across genes and strains was assessed by hierarchical clustering using Ward's minimum variances as a distance metric . The first principal component was used to assist in the sorting . Six biologically meaningful planned contrasts defined a priori , all of them orthogonal , were done likewise by pooling the appropriately expression data of different strains as necessary ( Table S17 ) . Scrutiny of the differences in post-hoc comparisons among strains was done using the “multcomp” R package [74] . Functional information for relevant genes was obtained from FlyBase [30] and enrichment for particular Gene Ontology term categories ( biological processes , molecular functions , and cellular components ) , KEGG pathways , and InterPro protein domains was evaluated using DAVID [75] . Benjamini-Hochberg correction [73] was applied to account for multiple tests . We proceeded likewise , but for each strain separately , to evaluate statistically significant differences between the sexes . Raw microarray data were deposited at the Gene Expression Omnibus database ( GSE31120 ) . In relation to the characterization of the expression profiles of the genes in the region under study during the life cycle of D . melanogaster , RPKM expression values across 30 timepoints and conditions [2] were extracted from FlyBase [30] , log transformed , and compared using hierarchical clustering as above . Core promoter predictions for all the transcripts of protein-coding genes examined were done with McPromoter using the most stringent parameter values [76] . We inspected 500 nt upstream of the 5′ UTR start of each transcripts as annotated in FlyBase [30] . In the absence of an annotated 5′UTR , a stretch of DNA of equal length upstream of the first nucleotide annotated was examined . Categorization of genes as responsive to HCNEs was based on the prediction of having an Inr core promoter type . We retrieved mapping coordinates of protein-coding genes included in the ultraconserved region under study in A . gambiae ( AgamP3 assembly ) using Biomart [77] . Any orthologous mapping information that did not conform to a one-to-one relationship between species was discarded . The global gene organization nearby the ortholog of Toll-7 in A . gambiae and A . aegypti was examined through VectorBase [78] . Phylogenetic relationships among Obp-related amino acid sequences encoded by genes in the same region that harbors the gene Toll-7 from the two mosquito species were conducted in MEGA 5 . 0 [79] . Amino acid sequences were downloaded from VectorBase [78] and aligned with CLUSTALW [80] . The best evolutionary model of amino acid evolution was found to be WAG and the consensus tree was built using the Maximum Likelihood method [81] . A discrete Gamma distribution was used to model evolutionary rate differences among sites ( +G; 4 categories ) . Bootstrapping was performed to determine the confidence of the branches ( 1 , 000 replicates ) . | Eukaryotic genomes have been reshaped by chromosomal rearrangements during evolution . However , the comparison of distantly related species has uncovered unusually large genomic regions with conserved gene organization . A widely accepted explanation is that , in those regions , there exist genes with joint and/or intricate regulation , which , if altered , might affect the performance of the carriers . We used a system that allowed us to precisely disrupt one of the largest genomic regions that has been conserved in the fruit flies since ∼63 million years ago and performed a variety of assays devised to detect differences in organismal performance and anomalous gene expression . Despite the overrepresentation of genes with expression profiles related to male fertility and detection of chemical stimuli in this genomic region , as well as the presence of genes expressed during development and putatively regulated by long distantly located sequences , we do not find evidence of diminished viability and fertility or of anomalous levels of expression . These results lead us to propose that regulatory constraints might not suffice to explain the maintenance of the integrity of some of the best candidate regions in one of the most dynamic eukaryotic genomes . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genomics",
"genetics",
"molecular",
"genetics",
"biology",
"evolutionary",
"biology",
"genetics",
"and",
"genomics"
] | 2012 | Evaluation of the Role of Functional Constraints on the Integrity of an Ultraconserved Region in the Genus Drosophila |
Dengue virus ( DENV ) is an extraordinary health burden on global scale , but still lacks effective vaccine . The Philippines is endemic for dengue fever , but massive employment of insecticides favored the development of resistance mutations in its major vector , Aedes aegypti . Alternative vector control strategies consist in releasing artificially modified mosquitos in the wild , but knowledge on their dispersal ability is necessary for a successful implementation . Despite being documented that Ae . aegypti can be passively transported for long distances , no study to date has been aimed at understanding whether human marine transportation can substantially shape the migration patterns of this mosquito . With thousands of islands connected by a dense network of ships , the Philippines is an ideal environment to fill this knowledge gap . Larvae of Ae . aegypti from 15 seaports in seven major islands of central-western Philippines were collected and genotyped at seven microsatellite loci . Low genetic structure and considerable gene flow was found in the area . Univariate and multivariate regression analyses suggested that anthropic factors ( specifically the amount of processed cargo and human population density ) can explain the observed population structure , while geographical distance was not correlated . Interestingly , cargo shipments seem to be more efficient than passenger ships in transporting Ae . aegypti . Bayesian clustering confirmed that Ae . aegypti from busy ports are more genetically similar , while populations from idle ports are relatively structured , regardless of the geographical distance that separates them . The results confirmed the pivotal role of marine human-mediated long-range dispersal in determining the population structure of Ae . aegypti . Hopefully corroborated by further research , the present findings could assist the design of more effective vector control strategies .
Aedes aegypti is the major vector of dengue virus ( DENV ) . An estimated 96 million dengue cases and 300 million asymptomatic infections were reported globally in 2010 [1] . Currently , there is no effective dengue vaccine available and the principal way to prevent the spread of the virus and the occurrence of outbreaks is to suppress the populations of vector mosquitos . Insecticide spraying has been the most common and effective method in the past decades , but massive and indiscriminate use generated a selective pressure that led to the development of insecticide-resistant strains of mosquitos . Alternative population control methods consist in the release of modified mosquitos which will interbreed with wild populations and suppress or replace it . Depending on the technique , the released individuals are sterilized , genetically engineered or carry symbiotic bacteria of the genus Wolbachia [2] . The dispersal ability of the vector should be taken into account when implementing any of the above-mentioned population suppression techniques . If a vector can travel long distances it can quickly and effectively spread insecticide resistance mutations , symbiotic bacteria or deleterious mutations . The flight range of Ae . aegypti is narrow ( usually less than 200m , rarely up to 500m ) [3–5] and its active dispersal in the wild should be limited . However , this mosquito is believed to occasionally travel long distances by taking advantage of human ( land , sea or air ) vehicles [6–8] . There are reports of Ae . aegypti collected from ships right after docking [9] and the trade of used tires is suspected to be one of the main contributions to the passive dispersal of Ae . aegypti immatures , both by land and sea [10] . Recently , larvae and pupae of Ae . aegypti have been for the first time collected at Narita Airport ( Tokyo , Japan ) . It is strongly suspected that they were transported via airplane from a foreign tropical country [11] . If Ae . aegypti's migration flows are mainly driven by active dispersal , the genetic differentiation between populations should increase with the geographical distance; this is called Isolation By Distance ( IBD ) [12] . Otherwise , if passive human-mediated dispersal is pivotal , IBD should be weak or absent . A remarkable amount of population genetics studies have been conducted on Ae . aegypti in many regions of the world . The genetic structure of this important vector mosquito has been analyzed at micro and macro geographical scales with different methods and purposes . Long-distance migration has been hypothesized and frequently related to human movements , but this association was not addressed in a quantitative way and only briefly brought up in the final discussion as a possible explanation of the results [13–15] . A few studies had their main focus on the effects of human transportation on Ae . aegypti's dispersal; yet , only the land transportation was taken into account [16–20] . One study undertaken in French Polynesia referred to the effect of human sea transportation; however , its main purpose was to determine if the circulation of DENV strains between Polynesian islands could be related to Ae . aegypti's migration [21] . Therefore , no study to date has been specifically aimed at quantitatively clarifying whether human-mediated long-distance marine dispersal is able to shape the genetic structure of Ae . aegypti at a macro-geographical scale . The Philippines is an archipelago located in Southeast Asia and is endemic for dengue fever . According to the World Health Organization , the annual reported cases of dengue fever increased around 17-fold in the decade 2000–2010 [22] . Unlike other southeast Asian countries ( mainly Thailand and Vietnam ) , very few ecological studies and no population genetics study have been conducted on Ae . aegypti in the Philippines [23–26] . A recent survey conducted by our group in the northern part of the country showed a widespread resistance to pyrethroid insecticides and a high prevalence of kdr mutations [27 , 28] . Furthermore , the nearly 7000 islands that make up the Philippine archipelago are connected by a dense network of cargo and passenger ships . The islands are remarkably different in size , demography and economic development; therefore , the seaports as well are strikingly different in size and commercial importance . Busy ports are more interconnected than idle ports , regardless of their geographical position . This offers a proper environment for the randomization of collection sites . For these reasons we decided to undertake a population genetics study aimed at understanding if and how human marine transportation affects the movements of Ae . aegypti among different islands of the Philippines . Specific objectives were: ( 1 ) to analyze the genetic structure of Ae . aegypti among different islands; ( 2 ) to quantitatively determine if natural ( geographical distance ) or anthropic ( port size and intensity of marine transportation ) factors affect the observed genetic structure .
In order to analyze the effect of human sea transportation on the migration of Ae . aegypti , seaport locations on several islands were sampled , as populations of mosquitos residing near the seaports are more likely to end up resting or laying eggs on the ships ( either cargo or passenger ) . Each site corresponded to one seaport and all individuals from a site were considered as one separate population , therefore the terms "site" , "population" and "port" will be used interchangeably , hereinafter . The survey was conducted between September and October 2013 in the central-western area of the Philippines ( Region IV-A and IV-B ) . Larvae were collected from 15 ports of different levels of connectivity and development , across seven islands of different sizes and degrees of urbanization ( Fig 1 and Table 1 ) . At each site , as many water-filled containers as possible ( artificial or natural ) were inspected in a 1 . 5 km-radius area; the radius was calculated from the center of the port . In case the container was located in a private area , the inspection was conducted only with consent of the owner . Collected larvae were conserved in absolute ethanol and identified by microscopic analysis [29 , 30] or by PCR amplification with species-specific primers [31] . Only larvae identified as Ae . aegypti were included in the genetic study . In order to reduce the proportion of sibling individuals ( which overestimate the genetic difference between populations ) , Ae . aegypti's flight range and its so-called "skip oviposition" behavior were taken into account [32] and a minimum distance of 200m was kept between the containers; then , one out of five larvae were randomly selected from each container and included in the genetic analysis . For port 6 one out of six larvae were selected . In case it was not possible to keep the minimum distance of 200m , larvae from multiple containers were grouped as if collected from the same . Genomic DNA was extracted from individual larvae with RedExtract-N-AmpTM Tissue PCR Kit ( SIGMA , USA ) following the manufacturer's protocol . Seven microsatellite loci were scored for each individual . Each PCR reaction ( 10μl ) contained 400nM of M13 fluorescent primer ( FAM-labeled ) [33] , 200nM of M13-tailed forward primer , 200nM of reverse primer , 1X of Buffer ( Mg2+ 2mM included ) , 200nM of each dNTP , 0 . 25U of Takara ExTaq ( Takara BIO , Inc . ) and 1μl of sample . Thermocycling conditions were: ( 5' at 94°C ) for one cycle , ( 30'' at 94°C , 30'' at Ta , 40'' at 72°C ) for 35 cycles , ( 5' at 72°C ) for 1 cycle and 4°C as holding T . T of annealing ( Ta ) was specifically determined for each pair of primers ( S1 Table ) . After PCR amplification , 1 . 5μl of product was added to 8 . 5μl of a GeneScan™ 500 ROX Size Standard:formamide solution ( 1:100 ) , heated at 95°C for 5' and run on an Applied Biosystems 3730xl DNA Genetic Analyzer . The allele scoring was conducted with the GeneMapper Software 5 ( Applied Biosystems ) . Seven microsatellite markers were selected out of 74 loci previously characterized in other studies [34–38] . After a search in the database , 33 were found to reside on different supercontigs . Preliminary PCR reactions with different Ta were conducted in order to find the most suitable for each primer pair ( S1 Table ) . Trial runs on capillary electrophoresis were performed to select the markers that reliably amplified in most of the samples , with sufficient allelic variability and that could be clearly and easily scored in the software GeneMapper . Many loci were discarded because the size of some of their allele peaks were misaligned with the sizes expected from their repetitive unit size; most of those loci turned out to be located near poly-A segments or to have deletions/insertions not related to the short tandem repeat of interest . The software GENEPOP v4 . 2 . 1 [39 , 40] was used to compute the basic population genetic parameters . The Inbreeding Coefficients ( Fis ) were calculated for all loci across populations according to Weir and Cockerham [41] , in order to detect any statistically significant deviation from Hardy-Weinberg Equilibrium ( HWE ) . Each pair of loci was also tested for Linkage Disequilibrium ( LD ) . Markow Chain parameters were set at 10 , 000 dememorizations , 20 batches , 5000 iterations for the former analysis and at 10 , 000 dememorizations , 100 batches , 5000 iterations for the latter one . GENEPOP was also used to calculate total and pairwise Fixation Indexes ( Fst ) ; P-values for pairwise Fst were computed with the software Arlequin v3 . 5 . 1 . 3 [42] . To test the hypothesis that human transportation influences mosquitos' dispersal , eight variables were elaborated; two were obtained from satellite images and six from official data published online by the Philippine Statistics Authority ( Table 2 and S2 Table ) . The variable named "Distance" accounts for the geographical location of the ports . The other seven variables account for port size and degree of ship connectivity , therefore we will refer to them as Anthropic Variables ( AVs ) hereinafter . Two AVs ( "Inhabitant" and "Density" ) are demographic data on the municipalities where ports are located . The variable named "Dock" expresses the physical size of a port , by measuring the perimeters of its concrete docks . The last four AVs ( "Vessel" , "Tonnage" , "Cargo" , "Passenger" ) measure the intensity of ship traffic at ports . The statistics for the last four AVs were incomplete and their correspondent values for ports 1 , 5 , 13 and 15 were missing . In order to interpolate them , four bivariate linear regressions with Inhabitant and Dock as independent variables were run . Inhabitant and Dock were chosen because of high correlation with Vessel , Tonnage , Cargo and Passenger ( Pearson's r≥0 . 47 , S3 Table ) . The command "lm" in the software R for MachIntosh ( v3 . 1 . 1 GUI 1 . 65 Snow Leopard build ) was used for the purpose . The command "princomp" in R was used to run a Principal Component Analysis ( PCA ) on the AVs . As the data showed considerable heteroscedasticity they were log-transformed before PCA [43] . As the AVs were expressed as a single value for each port ( 15 values per variable ) , while Fst values were pairwise ( 105 total port-by-port pairs , excluding same-port pairs ) , in order to quantitatively compare them , the values for the AVs were transformed to pairwise; the formula a+b was used , were a and b are the corresponding value for each two ports . Pearson correlation coefficients were computed in R with the command "rcorr" . To detect collinearity , the variance inflation factors ( VIF ) between eight predictors were computed ( command "vif" ) . The command "lm" was used to run the multivariate linear regression models . The best model was chosen with the command "step" , which performs a stepwise exclusion of each predictor and calculates the Akaike Information Criterion . Simple and partial Mantel tests were run using the online software "Isolation by distance web service" ( IBDWS ) [44] . One-tailed P-values were calculated with 10 , 000 bootstrap randomizations . The Bayesian clustering method implemented in the software STRUCTURE v . 2 . 3 . 4 [45] was used to assign individual Ae . aegypti to genetic clusters based on the microsatellite allelic frequencies . Admixture was selected as ancestry model with LOCPRIOR function . Alpha was allowed to vary and was separately calculated for each population . A correlated allele frequency model was chosen , with lambda set to one . Ten independent replicates were run with a burn-in step of 300 , 000 iterations followed by 500 , 000 replications . Average L ( K ) , ΔK [46] and proportions of membership ( i . e . the probability of a population to belong to a certain cluster , we will hereinafter refer to it as POM ) were computed in Excel . Pearson correlation coefficients between POM and principal component 1 ( PC1 ) were computed in R with the command "cor . test" .
An average of 211 Ae . aegypti larvae per site were collected ( total = 3166 ) . At each site between six and 15 containers were inspected and between four and eight were positive . One out of five larvae was randomly selected from each container and an average of 43 per population was genotyped ( total = 642 ) ( Table 1 ) . The test for deviation from HWE calculated 105 Fis values [41] , one for each locus-population pair ( S4 Table ) . Nine ( 8 . 6% ) showed significant departure from HWE ( after Bonferroni sequential correction ) , eight towards a heterozygosity deficit and one towards a heterozygosity excess . However , no locus showed significant values at more than three populations . The test for LD calculated a total of 315 values ( 21 locus-by-locus pairs for each population ) , of which 28 ( 8 . 9% ) remained significant after Bonferroni sequential correction . However , no locus-by-locus pair showed significance consistently across populations . Therefore this significance was unlikely due to linkage . None of the markers was excluded from the analysis . For most loci between four and six alleles were detected; only AG5 had 11 . The Fst for all loci and populations was 0 . 056 , while the pairwise Fst ranged from 0 . 005 and 0 . 147; after Bonferroni correction 79/105 ( 75 . 2% ) remained significant ( P<0 . 0005 ) ( Table 3 ) . Pearson correlation coefficient for Distance against Fst was close to zero and non-significant , while it was highly significant for all the AVs against Fst ( P<0 . 01 ) ( Table 4 ) . Tonnage and Passenger were excluded from the multivariate linear regression because of high collinearity ( square root of VIF>10 ) . Distance , Inhabitant , Dock and Vessel were dropped after stepwise analysis . The multiple R2 was 0 . 24 ( P<0 . 001 ) . The regression coefficients for Density and Cargo indicated a negative association with pairwise Fst ( P<0 . 001 ) ( Table 4 and Fig 2 ) . The square roots of VIFs are displayed in S5 Table . The PCA on the seven AVs returned a PC1 able to explain 74% of the total variance of the dataset; the directions of the vectors representing the AVs were almost opposite to the PC1 ( Fig 3 ) . It is therefore reasonable to assume that higher PC1 values indicate ports located in less developed and less connected municipalities and vice versa . STRUCTURE analysis could detect no genetic structure unless the option LOCPRIOR was chosen . This function uses information on sampling locations to assist the clustering in case the signal of structure is too weak [47] . A considerable degree of admixture was detected and most of the individuals were assigned to several clusters with almost equal probability . Nevertheless , with increasing K some populations had increasing probability to be assigned to a specific cluster ( Fig 4 ) . Following the method proposed by Evanno et al . [46] we found peaks of ΔK at K = 3 , K = 6 and K = 9 the last one being the highest ( S1 Fig ) . The peak at K = 3 was higher than K = 6 , but the bar plot of the latter was considerably more informative . The highest ΔK was at K = 9 , but L ( K ) decreases considerably between K = 6 and K = 9 ( S1 Fig ) and the additional structure explained by K = 9 is not remarkable ( Fig 4 ) . Therefore , we discuss clusterings at both K = 6 and K = 9 . Pearson correlation coefficients of POM regressed on PC1 scores were both high and significant ( for K = 6 , r = 0 . 76 , P<0 . 01; for K = 9 , r = 0 . 57 , P<0 . 05 ) .
Both univariate and multivariate regression models excluded IBD . In fact , Distance/Fst Pearson's r was almost zero ( 0 . 09; non-significant ) and Distance was dropped by multiple linear regression stepwise analysis ( Table 4 ) . On the other hand , Pearsons supported a negative correlation between Fst and the AVs ( all r values ≤-0 . 3; P<0 . 01 ) . The multivariate analysis confirmed the correlation only for Density and Cargo ( P<0 . 001 ) . The multiple linear regression coefficients were negative and very small in absolute value , therefore pairwise Fst decreases very slowly as Density and Cargo increase ( Table 4 ) ; this is also evident in the scatter plot ( Fig 2 ) . Our interpretation is that increasing cargo shipments increase the probability for Ae . aegypti to be accidentally transported , decreasing the genetic differentiation and consequently the Fst . Similarly , more densely populated municipalities have likely busier ports , with more congested ship traffic and offer more opportunities to Ae . aegypti to be transported . The results also suggest that cargo shipments transport Ae . aegypti more effectively than passenger ships . If we consider that cargo ships ( like barges ) are usually bigger in size than passenger ships , they may offer a higher number of breeding sites ( i . e . collections of water ) . Moreover , the cargo storage compartments are usually kept closed during the navigation , while passenger ships' decks are often not enclosed; this could cause more adult mosquitos to enter and being effectively transported by the cargo ships . Another hypothesis ( not excluding the previous ones ) is that cargo shipment routes are implemented earlier than passenger ships' routes during the development of the ports . In other words , ports of small and intermediate sizes could be connected by more cargo than passenger ships , while busiest ports operate a similar amount of the two ship types . This hypothesis is supported by the scatter plots between the four "ship connectivity" predictors ( i . e . Vessel , Tonnage , Cargo and Passenger ) ; all of them are highly correlated , but only Cargo shows a non-linear relation with the others . Values of Cargo increase earlier than values of the other three predictors ( S2 Fig ) . STRUCTURE analysis showed overall a considerable admixture and most of the genetic clusters were shared by different populations . Nevertheless , the populations showed variable degrees of genetic structure , from totally admixed ( port 3 ) to remarkably isolated ( port 1 ) ( Fig 4 ) . At both K = 6 and K = 9 , the level of admixture of the populations was strongly correlated to the development of the corresponding port ( POM regressed against PC1 ) . In other words , mosquitos from busier ports were more genetically admixed and mosquitos from smaller ports were more genetically isolated . Interestingly , the individuals had mixed ancestry while the populations as a whole were homogeneous ( this produced a pattern of "horizontal stripes" in the bar plot ) ( Fig 4 ) . This could be interpreted as the result of multiple connections between many of these ports for relatively long time . This pattern is different from other studies that analyzed Ae . aegypti's population structure at broader spatial scales ( several countries and continents ) ; in these studies the ancestry of individual mosquitos was more definite , but mosquitos from the same geographical populations were assigned to different genetic clusters ( this produced "vertical stripes" in the STRUCTURE bar plot ) [49 , 50] . A reasonable explanation is that mosquitos migrate more frequently and are more admixed within the same country than between countries and continents; as mosquitos are more distantly related at broader geographical scales , migrant individuals can be more clearly identified . Main limitation of this study could be the sample size ( Table 1 ) , as some of the populations were analyzed with few individuals ( especially ports 7 and 11 ) ; this could bias the estimation of Fst . Most locations yielded more than 100 individuals , but from few breeding sites . Therefore , we had to compromise between reducing sibling individuals ( which could also bias the analysis ) and have enough specimens to genotype . After Bonferroni correction 26 pairwise Fst values ( 24 . 8% of the total ) were not statistically significant and most of them corresponded to the ports with smaller sample size ( Table 3 ) . On the other hand , only four ( 3 . 8% of the total ) remained non-significant if Bonferroni correction was not applied ( S1 Dataset ) . Besides , simulation showed that between 25 and 30 individuals are enough to correctly estimate the allele frequencies with five-nine microsatellite markers [51] . The data from the Philippines Statistics Authority used to elaborate Vessel , Tonnage , Cargo and Passenger was incomplete and it was necessary to estimate the corresponding values for four ports . However , we believe that these limitations did not undermine the general meaning of our findings . Our overall interpretation is that the human ship transportation substantially shapes the dispersal patterns of Ae . aegypti in the central-western part of the Philippines . Regardless of their geographical location , large ports in highly-populated municipalities exchange individuals of Ae . aegypti at a faster rate than small-sized ports in less developed municipalities . Insecticide spraying , release of sterile males , release of Wolbachia symbionts: whichever vector control strategy will from now on be applied in the Philippines and other archipelagoes , these results should be taken into account . | Dengue fever threatens the health of millions in the tropics and its causative agent , dengue virus , is mainly transmitted by the mosquito Aedes aegypti . To control the spread of the virus , insecticides have been abundantly used but Ae . aegypti has developed a genetic resistance to them . Currently , alternative methods are being tested wherein artificially modified mosquitos are released in the wild to interfere with the mating of natural populations . It is important then to understand how the mosquito spreads in the environment . It is known that Ae . aegytpi can be passively transported for long distances by human vehicles , but it was not clear how common this event is , especially in case of marine transportation . In population genetics , a basic assumption says that if populations frequently exchange migrants , they become genetically more similar than relatively isolated populations . We estimated the genetic similarity between Ae . aegypti collected in the Philippines from 15 seaports of different sizes and ship connectivity . The mosquitos from busy ports , even distant ones , were genetically similar , while in the small ports , even close ones , Ae . aegypti were relatively differentiated . It was also suggested that Ae . aegypti's dispersal is affected by cargo shipments more than passenger ships . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | Human-Mediated Marine Dispersal Influences the Population Structure of Aedes aegypti in the Philippine Archipelago |
Indoor residual spraying ( IRS ) of DDT is used to control visceral leishmaniasis ( VL ) in India . However , the quality of spraying is severely compromised by a lack of affordable field assays to monitor target doses of insecticide . Our aim was to develop a simple DDT insecticide quantification kit ( IQK ) for monitoring DDT levels in an operational setting . DDT quantification was based on the stoichiometric release of chloride from DDT by alkaline hydrolysis and detection of the released ion using Quantab chloride detection strips . The assay was specific for insecticidal p , p`-DDT ( LoQ = 0 . 082 g/m2 ) . Bostik discs were effective in post spray wall sampling , extracting 25–70% of active ingredient depending on surface . Residual DDT was sampled from walls in Bihar state in India using Bostik adhesive discs and DDT concentrations ( g p , p`-DDT/m2 ) were determined using IQK and HPLC ( n = 1964 field samples ) . Analysis of 161 Bostik samples ( pooled sample pairs ) by IQK and HPLC produced excellent correlation ( R2 = 0 . 96; Bland-Altman bias = −0 . 0038 ) . IQK analysis of the remaining field samples matched HPLC data in identifying households that had been under sprayed , in range or over sprayed . A simple dipstick assay has been developed for monitoring DDT spraying that gives comparable results to HPLC . By making laboratory-based analysis of DDT dosing accessible to field operatives , routine monitoring of DDT levels can be promoted in low- and middle- income countries to maximise the effectiveness of IRS .
In India , DDT ( 2 , 2-bis ( p-chlorophenyl ) -1 , 1 , 1 , -trichloroethane ) is used extensively by the Indian National Vector Borne Disease Control Programme ( NVBCDP ) as part of efforts to eliminate VL [1] . VL is caused by Leishmania donovani and is transmitted by the female sand fly Phlebotomus argentipes [1] . An estimated 200 million people are at risk of this disease , of which 65 million live in India [2] , predominantly in Bihar State [3 , 4] . When DDT is sprayed at a target dose of 1 g/ m2 for VL control , it can be efficient in reducing the indoor abundance of sand -fly populations [5 , 6] . However , scale-up of this program for national implementation resulted in insecticide levels significantly below the target dose [1 , 7 , 8] , with up to 85% of walls being under sprayed in Bihar State [9] . Currently recommended approaches for quality assurance of IRS in vector control programmes falls into three categories: pre , during and post IRS quality assurance . Stock auditing , physically checking the state of spray equipment , servicing equipment at least once a year and rigorous training of IRS spray operators is done before IRS commences [10] . During operations supervision of spray operators by team leaders to ensure spray application technique is optimal and equipment malfunctioning is corrected , are advocated . In the case of DDT visual inspection of the powder delivered to the target surface is often used as an indicator of quality . Post IRS , WHO recommends cone bioassays to determine the quality of application of spray [10] . The imperative remains to develop practical field tools for close monitoring of DDT spray quality that allow reactive measures to poor spray quality . The efficacy of IRS is maximized , when its coverage is sufficiently extensive and the correct concentration of active ingredient used to kill the insect population targeted . Under spraying with sub-lethal doses of insecticide reduces impact on the disease vector and facilitates evolution of insecticide resistance . Routine monitoring of insecticide residues reaching the surface is critical to ensure correct dosing but is not done as the tools available for estimating insecticide amounts on surfaces are not practical for field use . These include cone bioassays requiring live insects , which are non-quantitative and high performance liquid chromatography ( HPLC ) , which is reliant on sophisticated and expensive systems , severely limiting their application in resource poor settings [11] . Most recently , colorimetric assays for cyanopyrethroids and carbamates have been developed for monitoring insecticide levels in bednets and sprayed structures [12–14] DDT is an organochlorine molecule , thus quantifiable by enzymatic [15] or chemical [16] release of chloride ions ( Cl- ) . Given the long development time and cost associated with producing enzyme based diagnostic assays and the fact that chloride detection kits such as Quantab Test Strips are available , we focused on developing a chemical platform for DDT detection with ‘off-the-shelf’ reagents that could be immediately deployed in the field . Here , we have developed a simple DDT dipstick assay and matched it against field samples taken in parallel with samples quantified by HPLC during a recent large-scale quality assurance survey of IRS operations in Bihar State , India in 2014 [9] . The results demonstrate that the IQK can provide accurate quantification of DDT , enabling the immediate implementation of robust quality assurance practices where IRS is used for vector control .
Insecticide analytical standards p , p`-DDT ( 1 . 1 . 1-Trichloro-2 . 2-bis ( p-chlorophenyl ) ethane , Dichlorodiphenyltrichloroethane ) 98 . 2% purity , o , p’-DDT 99 . 2% purity and DDE ( 1 , 1-dichloro-2 , 2-bis ( 4-chlorophenyl ) ethylene ) 99 . 5% were purchased from Chem Service ( West Chester , UK ) . Potassium Hydroxide BioXtra , ≥85% KOH basis and Dicyclohexyl phthalate 99% purity were purchased from Sigma Aldrich ( UK ) . HPLC grade solvents , heptane anhydrous 99% purity , 1-propanol anhydrous 99 . 7% purity , acetic acid 99 . 7% and HPLC grade water were purchased from Sigma Aldrich ( UK ) . Low range Chloride Quantab Test Strips , 30–600 mg/L , were purchased from Hach ( Hach Lange , Salford , UK ) . Step 1 , DDT extraction: DDT samples ( Bostik adhesive discs or filter ) ~10 cm2 in area were cut into small pieces into a 50 ml Falcon tube . 1mL heptane was added and vortexed for 2–3 minutes . Step 2 , Chloride release: 680 μl of the extract was transferred to a fresh tube , 100 μl of 2M KOH ( dissolved in 1-propanol ) was added , mixed vigorously for ~30 sec and incubated at room temperature ( RT ) for 1 hr . The reaction was neutralized with 100 μl of 2M acetic acid and the tube shaken vigorously for ~1 min then left to stand for ~1 min for phase separation . The bottom aqueous phase contains concentrated Cl- ions . Step 3 , DDT measurement: the bottom of Quantab strips were cut either side of the central wick to form a V shape and dropped into the tube with the wick immersed in the bottom aqueous phase . Strips were left for ~10 min for the Cl- solution to reach the yellow indicator strip . Once the indicator turned blue , indicating saturation , the strips were removed and Cl- readings in part per million ( ppm ) obtained by visual comparison of the chloride peak height against the manufacturer’s conversion chart . A p , p’-DDT quantitative chart in ( g/m2 ) was also prepared for direct readout of DDT content as detailed below . To examine the effects of contaminating salts or alkaline compounds on Quantab readings , 0 . 01–1 mg of NaCl or Ca ( OH ) 2 ( lime ) was added in Step1 . Standard stock solutions of p , p’-DDT in the range 62 . 5–3000 μg/ml were prepared in 1 ml heptane . 680 μl of each standard were mixed with 100 μl of 2M KOH ( 1-propanol ) and the IQK assay followed ( Steps 2–3 ) . Data was fitted to a third order polynomial ( cubic ) hyperbola curve using GraphPad Prism version 6 . 0 for Mac OS 10 . 10 ( GraphPad Software , La Jolla California USA ) and a DDT quantitative chart generated by interpolation of the manufacturers Quantab scale . The expected p , p`-DDT concentrations obtained from the curve were multiplied by a correction factor of 1 . 49 for heptane extraction efficiency ( 83 . 4% ) and a third order polynomial ( cubic ) model underestimation value ( 80% ) to give final readings of DDT in g/m2 resulted in DDT Quantitative chart to give a direct estimation of p , p`-DDT concentration in g/m2 vs . Quantab units on the scale . DDT breaks down into DDE and DDD . The linearity of DDE production and associated Cl- release in the IQK assay was measured at RT and 40°C using DDT concentrations in the range 62 . 5 to 3000 μg/ml . The concentration of DDE was measured from the heptane phase following KOH treatment ( IQK Step 2 ) by HPLC . 250μl of the heptane phase was evaporated to dryness under nitrogen and the residue dissolved in 1 ml methanol and centrifuged at 15000 rpm ( Microcentrifuge Eppendorf; Model 5424R , UK ) . 10μl was injected into a reverse-phase Hypersil GOLD C18 column ( 175 Å , 250 x 4 . 6 mm , 5 μm , Thermo Scientific , UK ) at 23–25ο C . A mobile phase of acetonitrile/water 93:7 was used at a flow rate of 1 ml/min . DDT and DDE peaks were detected at 232 nm with an Ultimate 3000 UV detector ( Dionex , Camberley , UK ) and analysed by Dionex Chromeleon software . DDE and DDT sample concentrations were calculated from standard curves produced from known concentrations of DDE and p , p’-DDT standards . The HPLC method was linear in the range 0 . 064–400 μg/ml , with a mean coefficient correlation r2 ≥ 0 . 99998 . The limits of detection and quantification were 0 . 008 and 0 . 032 μg/ml ( RSD ≤ 2% ) . Adhesive ( 2x5 ) cm2 Bostik discs ( Bostik , Leicester , UK ) and Sellotape ( 10 cm2 ) were used . For comparison of extraction efficiencies , rough sided tiles were evenly coated by pipette with Indian DDT wettable powder formulation ( DDT 50% WP ) at variable concentrations up to 2g/m2 . DDT recovered from the tile surface by adhesives were extracted by heptane and analysed by HPLC . Pearson’s correlation coefficient and Bland Altman analysis were used to assess systematic bias between the two analytical methods . To calculate sampling efficiencies of Bostik discs against different surfaces , ~0 . 25 m2 pieces of bamboo , thatch , brick un-plastered , brick plastered , brick lime washed and mud plaster surfaces were similarly coated with DDT formulation at 1 g/m2 of DDT 50% WP and left to dry for 1 day . IQK assays were performed in four replicates per surface , and the percentage p . p’-DDT recovered versus the quantity applied was calculated . Samples were collected from eight districts in Bihar State during the first DDT IRS round in 2014 ( Feb–July ) as described[8] . Briefly , two Bostik discs per HPLC sample ( 10 cm2 in total ) and two adjacent discs for IQK were placed onto the wall surface and rubbed firmly . Bostik discs were removed from the wall surface using tweezers and placed on Whatman no . 1 filter paper , labelled and stored in a polythene bag at 4°C until analysis . Samples were taken from each of the four walls in bedrooms , at a different position per wall: high ( 4–6 feet from ground level ) , medium ( 2–4 feet from ground level ) , low ( 0–2 feet from ground level ) . The concentration of DDT present on wall surfaces was determined using HPLC[8] and IQK assay . A 20% cut-off threshold was used to classify results whereby a concentration of less than 0 . 8g/m2 was considered an under spray , a range of 0 . 8–1 . 2g/m2 was considered within the target range , and a concentration of greater than 1 . 2g/m2 was considered an overspray Theoretical values for the limit of blank ( LoB ) , limit of detection ( LoD ) and the limit of quantification ( LoQ ) were determined according to EP17 guideline [17] . Statistical analyses were performed using GraphPad Prism version 6 . 0f for Mac OS 10 . 10 , GraphPad Software , La Jolla California USA , www . graphpad . com and Microsoft Excel 2010 . Correlation and Bland Altman statistics was used to compare and calculate the bias between analytical methods . Fisher’s exact test was used to evaluate statistical significance , with a cut-off P value of < 0 . 05 . Generalized mixed-effects modelling using R v3 . 1 . 0 was used to analyse variance within individual households .
Alkaline treatment of p , p’-DDT results in stoichiometric Cl- release ( DDT: Cl- , w/w , ~10:1 ) which is easily measured using Chloride Quantab Test Strips that contain silver ions that form a visible white precipitate with chloride in sample solutions . The limits of Cl- detection for the most sensitive strips are 30–600 ppm ( S1 Fig ) , well below the solubility limit of DDT ( 0 . 025 ppm ) . To address this , heptane was used to solubilize DDT and a four-step process developed for Cl- release , concentration and measurement ( Fig 1 ) . By dissolving DDT in heptane we were able to remove insoluble contaminant ions , which remain behind when DDT is transferred to a fresh tube . Similarly , by partitioning the Cl- released by KOH into a small volume of aqueous acetic acid ( stop solution ) , we were able to concentrate Cl- into the measurable range of the Quantab strips . Quantab strip readings matched the theoretical ( stoichiometric ) calculations of chloride ions released from DDT , allowing simple estimation of DDT concentration ( Fig 2 ) . The limits of blank ( unsprayed ) , detection and quantification were 0 . 009 , 0 . 07 and 0 . 08 g/m2 , respectively and the assay demonstrated high precision with RSD ≤ 4 . 4% at 1 g/m2 following IQK analysis of 24 sequential samples from rough tiles sprayed with 1 g DDT / m2 ( S2 Fig ) . Equivalent DDT measurements were obtained at 20°C and 40°C ( S3 Fig ) . DDT dehydrochlorination with concomitant release of DDE and Cl- was linear in the DDT concentration range 0–4 mM , equivalent to 0 – 3g DDT/m2 . There was 100% conversion of DDT to DDE with concomitant release of Cl- as measured by Quantab strips within 60 min ( S3 Fig ) . Overall , these data demonstrated that Quantab strips were able to measure DDT at the levels employed for IRS ( 1–2 g/m2 ) and under tropical temperatures ( 30–40°C ) The standard DDT WP formulation used for IRS contains a mixture of the active DDT ingredient p , p`-DDT ( ~80% of total DDT content ) , plus inactive DDT analogues including o , p`-DDT isomer ( ~15% ) and DDE ( ~5% ) . Thus , the specificity of the assay was tested against non-target DDT analogues and biologically inactive DDT breakdown products often present in DDT formulations . There was no cross-reactivity with DDE or the inactive o , p`-DDT isomer at concentrations up to 2 . 8 mM ( equivalent to 2 g/m2 ) . In India , where this assay was field-tested , lime wash ( Ca ( OH ) 2 ) is commonly used to paint house walls , which could potentially cause false positive readings along with other residual ions . Neither sodium chloride nor lime effected Quantab readings ( S4 Fig ) . DDT is applied to surfaces as a wettable powder , thus tractable to post spray sampling using sticky tape removal of the surface spray layer [13] , [15] . Two different sampling methods were examined: normal adhesive tape ( Sellotape ) and Bostik adhesive discs . The rough side of tiles was used as an initial laboratory reference surface for estimating extraction efficiency . DDT was extracted from sprayed tiles and analyzed using IQK and HPLC ( Fig 3 ) . Bostik discs produced highly correlated IQK and HPLC measurements ( R2 = 0 . 92; Bland Altman bias = 0 . 70 ) ( Fig 3A ) compared with Sellotape ( R2 = 0 . 70; Bland Altman bias = 0 . 15 ) ( Fig 3B ) , recommending Bostik discs for sampling . The average percentage recovery of DDT from rough tiles using Bostik adhesive discs was 53% +/- 5 . 6 . DDT recovery was further examined using six typical Indian household surfaces ( Fig 4 ) . Recovery rates ranged between 25–70% depending on surface material and roughness ( Bamboo 55 . 1% , Thatched 25 . 7% , Brick un-Plastered 40 . 5% , Brick Plastered 25% , Mud Plaster 60% and Lime Wash 69 . 9% ) , and these values were used for field estimations of DDT application . In order to test the IQK as a method for capturing QA performance data , the IQK assay was matched against HPLC using 1964 Bostik disc samples ( 491 households ) in Bihar state during the April–June 2014 DDT IRS season[8] . Double the numbers of samples were collected from each household in the post-IRS QA survey , and two types of analysis to test the IQK were performed . In the first , both pairs of Bostik discs underwent a common extraction process to obtain a uniform sample , which was then used for both HPLC and IQK ( the ‘pooled testing’ ) . In the second , each Bostik adhesive disc pair was analyzed independently ( ‘non-pooled testing’ ) . One hundred and sixty one samples were tested using pooled analysis ( Fig 5 ) . The two assays showed an excellent level of correlation , with an R2 value of 0 . 96 . The comparison of the IQK to HPLC using Bland-Altman analysis showed a very low bias of—0 . 0038 , which confirmed that testing Bostik discs using the IQK was comparable with using HPLC . The correlation of IQK versus HPLC for non-pooled sample analysis ( 1964 samples ) was lower ( R2 = 0 . 61 ) ( Fig 6A ) , reflecting the inherent variation in the DDT wettable powder formulation delivered to walls; however , the Bland-Altman analysis for the non-pooled samples still showed a low bias of 0 . 01 . Comparison of household averages increased correlation ( R2 = 0 . 81; Bland-Altman bias = 0 . 03566 ) ( Fig 6B ) between the two methods . No significant interaction was found between the height of sampling and wall surface type ( P = 0 . 7 ) , and no significant effect of time of DDT sampling up to 15 days’ post DDT spraying was found ( P = 0 . 5092 , interaction P = 0 . 7331 ) . The simple finger rubbing extraction procedure for recovering DDT from surfaces was consistent as measured by Levey-Jennings analysis of households sprayed within the correct target range ( 0 . 8–1 . 2g/m2 DDT ) ; 94% were within 2SD of the mean DDT concentration ( 0 . 93 g/m2 ) ( S5 Fig ) . The inclusion of data beyond 15 days of post spray sampling did show a significant effect on the concentration of DDT recovered from walls ( P = 0 . 0031 ) , although we believe this may be affected by the low number and uneven distribution of wall surface types in the post 15 days’ spray range . Graphs comparing IQK and HPLC data from all the replicates from each sampled house are presented in Fig 7 . For programmatic decision-making , cut-offs have been applied to indicate under spraying ( <0 . 8 g/m2 ) , correct spraying ( 0 . 8–1 . 2 g/m2 ) , and over spraying ( >1 . 2 g/m2 ) ) . There was no significant difference between the two methods ( P = 0 . 4 ) ( Fig 7B ) , and strong correlation in interpretation of the three dosage strata ( r = 1 . 0 , P = 0 . 005 ) ( Fig 7C ) . Overall , poor quality spraying was evident with only 8 . 6% ( n = 170 ) of samples at the target dose ( 0 . 8–1 . 2 g DDT/m2 ) . Of the remaining samples , 11 . 2% ( n = 220 ) were overdosed , and 80 . 1% ( n = 1573 ) under-dosed . Similar levels of poor spray quality were evident for all surfaces apart from thatch ( Fig 8 ) , where a high level of over spraying ( 47% , n = 41 ) was evident .
At present , the operational response to poor quality spraying is severely limited by the cost and lengthy turnaround times for laboratory analysis of field samples . Although a range of new field tests for insecticide quantification are under development , including biosensors for DDT and pyrethroid detection [15 , 18] , and colorimetric tests for cyano-pyrethroids [13 , 14] , none are in routine use . By using ‘off-the-shelf’ reagents we have produced a simple low cost insecticide quantification kit for the detection of DDT that performs favourably against HPLC . As well as being expensive at ~£3 . 0 per assay excluding equipment cost , performing HPLC assays cannot be done on-site in the field , limiting programmatic ability to respond to QA failures . By comparison , material and reagent cost per test of our IQK assay is ~ £1 . 0 , with quantitative results that are easily interpretable by untrained users within ~1 . 5 hours . In this study ~2000 samples were analysed at a rate of ~150–170 samples/ person/ day . The lower capacity and sample run times ( 20 minutes ) of HPLC , in contrast , limited the number of samples analysable per day to ~100 . The development of a simple field assay for measuring the active ingredient of DDT ( p , p’-DDT ) thus presents new capabilities for users to perform routine on-the-spot tests to ensure correct DDT dosing . Quantification is achieved by measuring the Cl- released following alkaline hydrolysis of DDT . The Cl- released can be simply measured using Quantab strips and correlates with the initial DDT concentration in the range 0 . 04–3 mg/ml , equivalent to DDT formulation spray rates of 0 . 08–6 g/m2 , covering the normal target range of 1–2 g/ m2 used for VL and malaria control operations respectively . The assay is highly specific for the biologically active p , p’-DDT molecule against the DDT breakdown products , while simple phased separation of DDT into heptane followed by separation of the Cl- released into aqueous acetic acid alleviates interference by dirt or chemicals . Overall , the assay is far more practical compared with standard HPLC methodology , while the high specificity for p . p’-DDT provides an advantage over alternative immunoassay-based systems[18] . The assay is also adaptable to use with other chloride detection systems . For example , once Cl- is released into the aqueous phase ( Fig 1 ) AQUANAL -professional Chloride reagent test ( FLUKA-UK ) can be used alone or following Quantab strip measurements to produce a colorimetric reading for the Cl- detection step . The main obstacle in adapting the IQK for monitoring IRS is to develop a method to accurately sample insecticide from the walls of sprayed houses . Attaching large filter papers to walls prior to spraying [19] , as recommended by WHO , creates behavioural bias due to the visibility of the papers , although this may be overcome using more discrete attachments such as small filter papers or felt pads [13] . Whilst the IQK can easily be used with pre-spray attachments , anonymous direct sampling of residues off wall surfaces is far more desirable . Since DDT is applied to walls as a wettable powder , Bostik adhesive discs were found to be a reliable method for extracting DDT for post spray measurement . The efficiency of DDT recovery ranged from 25 to 70% depending on surface type , which can be factored into the estimations of DDT application . Water dispensable powders such as DDT have the advantage of low cost and long residuality on porous surfaces [20] . However , this is offset by rapid sedimentation of active ingredient and spray nozzle clogging that can produce non-uniform spraying , potentially impacting spray quality , and reinforcing the need for close monitoring of DDT spray quality . In order to assess the performance of the DDT IQK in measuring DDT , paired samples were taken from Indian households and either pooled ( 160 samples ) for HPLC and DDT IQK analysis or analysed separately ( 1964 samples ) . The results for DDT IQK closely matched those of HPLC and were unaffected by surface contaminants collected from diverse surfaces ( brick walls , wood , thatch and lime painted surfaces ) . Since DDT is applied in swathes using stirrup-pumps in India , the patchiness of the powder residues deposited on the walls was reflected by the different correlation coefficients of pooled vs non pooled paired samples for the HPLC vs IQK analysis ( R2 = 0 . 96 vs . 0 . 64 , respectively ) . However , the correlation of non pooled samples was strengthened ( R2 = 0 . 8 ) when averaged household concentrations were compared . Overall , these data demonstrate the dipstick assay to be a robust method for quantifying DDT spray rates . In this study , the IQK data confirmed that the rate of DDT spraying in Bihar State was consistently below the 1 . 0 g ai/m2 target concentration , with little variation in DDT concentrations within households or across districts [9] . Reasons for the reduced quality of the IRS have yet to be established , but include sub-standard spraying or sub-standard WHO specification formulation , both of which are measurable by IQK . The ability of the IQK to quantify the DDT active ingredient demonstrates that a simple dipstick assay can be used to verify that sprayers have actually sprayed a house , to evaluate the spray coverage in each house and to verify the effective insecticide coverage rates per area . Since DDT is applied at high concentrations ( 1–2 g/ m2 ) as a wettable powder the sticky adhesives work well in recovering the formulation after spraying . However , post spray recovery with adhesives may not be as efficient with microencapsulated formulations or low concentrations of active ingredient ( mg range ) . As a rough estimate we have found the Bostik Discs pulls 40–50% of DDT off surfaces , but for accurate estimations , one must consider differences in surface recovery depending on the types of walls being sprayed . Here , DDT recoveries from common wall surfaces in Bihar have been factored into DDT estimations , but these might vary in different countries or states . The availability of the simple dipstick assay for DDT quantification is an important step forward in enabling programme managers to report on the quality of their operations . From an operational perspective , the simplicity of the IQK means that large numbers of samples can be processed on a daily basis on-site . The next important step will be to develop standard operating procedures to information on spray quality that is relevant to spray operatives that will allow them to take appropriate steps in the case of spray failures | Visceral leishmaniasis ( VL ) is a major parasitic disease on the Indian Subcontinent , with 85% of the disease incidence in India . DDT is being used to control the sandfly vector of VL , but the quality of indoor residual spraying ( IRS ) is not routinely assessed due to a lack of practical assays . Here we have developed a simple DDT insecticide quantification kit ( IQK ) for measuring the amounts of DDT being applied to surfaces during large scale vector control operations . We demonstrate that the IQK provides equivalent data to the gold standard high performance liquid chromatography ( HPLC ) assay . Being truly point-of-care , this test is able to immediately inform the management of spray quality . Integration of the IQK will allow non-laboratory staff such as spray team supervisors to establish robust quality assurance procedures for DDT dosing that are vital to support the South Asian VL elimination efforts . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [] | 2016 | Development of a Simple Dipstick Assay for Operational Monitoring of DDT |
Systemic lupus erythematosus is a complex and potentially fatal autoimmune disease , characterized by autoantibody production and multi-organ damage . By a genome-wide association study ( 320 patients and 1 , 500 controls ) and subsequent replication altogether involving a total of 3 , 300 Asian SLE patients from Hong Kong , Mainland China , and Thailand , as well as 4 , 200 ethnically and geographically matched controls , genetic variants in ETS1 and WDFY4 were found to be associated with SLE ( ETS1: rs1128334 , P = 2 . 33×10−11 , OR = 1 . 29; WDFY4: rs7097397 , P = 8 . 15×10−12 , OR = 1 . 30 ) . ETS1 encodes for a transcription factor known to be involved in a wide range of immune functions , including Th17 cell development and terminal differentiation of B lymphocytes . SNP rs1128334 is located in the 3′-UTR of ETS1 , and allelic expression analysis from peripheral blood mononuclear cells showed significantly lower expression level from the risk allele . WDFY4 is a conserved protein with unknown function , but is predominantly expressed in primary and secondary immune tissues , and rs7097397 in WDFY4 changes an arginine residue to glutamine ( R1816Q ) in this protein . Our study also confirmed association of the HLA locus , STAT4 , TNFSF4 , BLK , BANK1 , IRF5 , and TNFAIP3 with SLE in Asians . These new genetic findings may help us to gain a better understanding of the disease and the functions of the genes involved .
Systemic lupus erythematosus ( SLE ) is a prototype autoimmune disease characterized by auto-antibody production and multi-organ damage . Genetic factors are known to play an important role in the disease , with the monozygotic twin concordance rate between 20–59 , and the risk for siblings of affected individuals 30 times higher than that for the general population [1]–[3] . There are also population differences for the disease both in terms of genetic susceptibility and disease manifestations . African Americans , Hispanics and Asians all have higher disease prevalence than Caucasians; with Asians known to have more lupus nephritis than patients of European ancestry [4] , [5] . Genome-Wide Association studies ( GWAS ) have dramatically changed the landscape of SLE genetics , with a pace of discovery the field has never seen before . In less than two years time , STAT4 [6] , ITGAM [7]–[9] , BLK [8] , PXK and KIAA1542 [7] , BANK1 [10] and TNFAIP3 [11] , [12] and several other genes have been identified as associated with SLE [13]–[16] . More susceptibility loci were reported recently in two other GWAS on this disease [17] , [18] . Despite varied disease prevalence and severity across different populations , it is noteworthy that most previous studies were conducted on patients of European ancestry with under-representation of other ethnicities . We previously examined some of the GWAS findings from populations of European origin [19] . In addition to confirming the association of STAT4 and BLK with SLE in our population , our data indicated differences between the Asian and Caucasian populations . For example , our study did not detect any significant disease association for PXK , a result that was later confirmed by an independent study on a Korean population [20] . Data from our study indicated that , although the risk alleles in ITGAM are rare in Asians ( <2% ) , they are risk factors in our populations and are closely related to lupus nephritis in particular [21] . In this study , we first genotyped 320 SLE patients collected in Hong Kong by the Illumina 610-Quad Beadchip and analyzed the data against 1500 control individuals genotyped on the same platform . Selected SNPs were then replicated in four independent sample collections from Hong Kong , Shanghai and Anhui , ( China ) , as well as Bangkok ( Thailand ) . Genetic variants in and around two genes , ETS1 and WDFY4 , were identified as associated with SLE with genome-wide significance . Functional characterization of the risk alleles also supported potential roles of these genetic variants in disease pathogenesis .
The whole-genome genotyping data was thoroughly examined by quality control measures and by population substructure analysis . Analysis of principal component using Eigenstrat [22] did reveal that the samples collected in Hong Kong clustered together , suggesting that confounding population substructure or admixture is not a major concern if Hong Kong controls were used in association analyses . It did indicate , however , that samples collected in Taiwan ( obtained from deCODE Genetics ) and Han Chinese in Beijing ( HCB , available from HapMap ) cluster very differently from Hong Kong samples , suggesting population substructure among Chinese living in different geographical regions and potential pitfalls in association studies when cases and controls are not well-matched ( Figure 1 ) . Genome-wide association analysis confirmed significant association of some established susceptibility genes in our population , including SNPs in the HLA locus , STAT4 , TNFSF4 , BANK1 , TNFAIP3 , IRF5 and BLK ( Table 1 and Figure 2 ) . Similar to the Caucasian findings [12] , [23] , the risk allele for rs2230926 in TNFAIP3 is low in frequency but with a relatively large effect size in disease association . SNP rs9271366 located between HLA-DRB1 and HLA-DQA1 is the most significantly associated SNP in the whole genome in our study . Although the HLA locus has been consistently shown to be the locus conferring the largest effect size with SLE association , there is little overlap between previous GWAS findings from populations of European ancestry and our results . The most significant SNPs in the Caucasian data [7] , [8] are either monomorphic in our population or are not associated with disease susceptibility based on our GWAS result , something worth further pursuit in future studies . Association of STAT4 , BLK , IRF5 , BANK1 and TNFSF4 with the disease has been reported in our population previously [19] , [21] , [24] , [25] . To answer the question on whether there are still other genetic variants contributing to disease susceptibility , we reexamined the Q-Q plot comparing expected and observed P values by removing all the SNPs in the known susceptibility loci mentioned above . After removal of these SNPs , we still observed an excess of association signal ( Figure 3 ) , suggesting involvement of additional susceptibility loci for this disease . Since our GWAS involved a limited number of patients , and is therefore prone to false positive and false negative findings , we selected SNPs for replication based on both their significance in GWAS results as well as the function and expression pattern of the nearby genes . Selected SNPs were replicated first using Sequenom genotyping on limited number of additional samples ( 360 cases and 360 controls ) , and variants with significant association in the Sequenom data were then examined by TaqMan genotyping on a much expanded sample collection from four independent cohorts . SNPs in and around two genes , v-ets erythroblastosis virus E26 oncogene homolog 1 ( avian ) ( ETS1 ) and the WDFY family member 4 ( WDFY4 ) regions were chosen based both on initial GWAS data as well as their known function ( in the case of ETS1 ) and expression pattern ( in the case of WDFY4 ) . Table 2 displays the SNPs in these two loci that showed disease association with a P value<0 . 01 from our GWAS data . Initial replication by Sequenom showed consistent results with the GWAS trend for these two genes and they were further tested in the remaining samples . Making use of the remaining samples from Hong Kong not included in GWAS , and sample collections from Shanghai and Anhui , China , and Bangkok , Thailand , we went on to replicate the whole-genome findings on these two loci . SNPs in ETS1 listed in Table 2 , rs12223943 , rs7932088 and rs10893872 , were examined in the expanded samples . SNPs rs10893872 and rs4937333 have absolute LD with each other , so only rs10893872 was chosen for replication . SNP rs1128334 was chosen in the place of rs6590330 for replication due to its high LD with rs6590330 ( r2 = 0 . 97 , Figure 4 ) and its relative position to the gene ( 3′-UTR ) and predicted effect on microRNA binding . For SNPs rs12223943 and rs7932088 , the association seen from whole-genome data was inconclusive ( data not shown ) in the replication stage and were not further pursued . We genotyped rs10893872 and rs1128334 in all the samples from the four cohorts and both were found to be highly associated with SLE ( Table 3 ) . Independence test by logistic regression by Plink pointed to a major contribution from rs1128334 . And indeed , the sequence around rs1128334 has high sequence conservation among different species ( Figure 4 ) . Haplotype analysis indicated that the TA haplotype formed by the two SNPs ( rs10893872 and rs1128334 ) is the major risk haplotype , whilst the CG haplotype is the major protective haplotype ( Table 4 ) with other haplotypes having low allele frequencies . Subphenotype analysis was also performed for these SNPs , and both SNPs were found to have larger effect sizes for patients with lupus nephritis in all four cohorts , although no statistical significance was reached in any cases . Analysis of other subphenotypes showed insignificant , or inconsistent results among different cohorts . Since rs1128334 is located at the 3′-UTR region of the gene , it is predicted that it may have an effect on the expression level of ETS1 . Therefore , we examined allelic expression of ETS1 gene for the two alleles of rs1128334 , “A” and “G” , from healthy individuals heterozygous for this SNP ( N = 33 ) . This assay assesses directly whether the two alleles of the SNP correlate with different steady state mRNA levels . Pyrosequencing results from PBMC of healthy individuals heterozygous on the SNP showed a significantly higher expression from the “G” allele than from the risk “A” allele , with a P value<0 . 0001 ( Figure 5 ) . Two of the SNPs in WDFY4 that showed the most significant association with the disease in our GWAS , SNPs rs10857650 and rs877819 , were selected for further replication . In addition , three nonsynonymous SNPs in this gene not genotype by the Illumina 610-Quad Beadchip , rs2170132 ( Ser1528Pro ) , rs7097397 ( Arg1816Gln ) and rs2292584 ( Pro3118Leu ) , were also selected to test for disease association , aiming at identifying functional variants in this gene . SNPs rs2170132 and rs2292584 did not show significant difference between the cases and the controls in the Hong Kong cohort ( Table 5 ) and Thai samples ( data not shown ) and were not further tested in other cohorts . Genotyping results on rs10857650 using TaqMan showed significant discordance with results from the Illumina Beadchip , and was thus removed from further analysis . SNP rs7097397 and rs877819 were confirmed to have significant association with the disease ( Table 3 ) . Conditional logistic regression test indicate that the nonsynonymous SNP coded by rs7097397 is probably the functional variant , with P = 1 . 01×10−5 when controlling the effect of rs877819 . Independent contribution from rs877819 is questionable , with a P value of 0 . 088 considering the effect of rs7097397 in the same test . The two SNPs have intermediate LD ( r2 = 0 . 44 , Figure 6A ) . The genetic result is consistent with the fact that the arginine residue at 1816 in WDFY4 protein is well conserved among orthologs in different mammals ( Figure 6B ) . Preliminary analysis on subphenotype stratification suggests that this nonsynonymous change may be more closely involved in male and early onset patients in a case only analysis ( onset age 12 and below vs . above , OR = 1 . 96 , P = 0 . 0021; male vs . female: OR = 1 . 52 , P = 0 . 023 ) . Association of the SNPs in the two genes with disease risk was also corrected by logistic regression using age and sex as covariates , and the associations found in this study all remain highly significant after all the corrections .
Several GWAS on SLE have been conducted on populations of European ancestry [7] , [8] , but populations of Asian or African ancestry were seriously underrepresented . Only during the process of submission of our current work , a GWAS study on Chinese populations was reported [18] . Considering the population differences in both disease prevalence and clinical manifestations , GWAS on non-Caucasian populations may have novel findings and help to elucidate the differences between populations . An interesting analysis result from our GWAS data is the difference between Hong Kong samples and samples collected in Taiwan and Beijing , shown by principal component analysis ( Figure 1 ) . It suggests population substructure for Chinese living in different regions , which may cause spurious findings in association studies when cases and controls are not well matched . With most of the genetic variants of relatively larger effect sizes already being identified , GWAS becomes more susceptible to effects from mismatches between cases and controls in dealing with SNPs of smaller effect sizes . Our analysis echoed two very recent reports delineating population substructures in Chinese populations living in different geographical regions [26] , [27] . Ets-1 is a member of the ETS family of transcription factors that share a unique Ets DNA binding domain . They control a wide variety of cellular processes including cell proliferation and differentiation [28] . Ets-1-deficient mice develop lupus-like disease characterized by high titers of IgM and IgG autoantibodies , immune complex-mediated glomerulonephritis , and local activation of complement [29] . Ets-1 is also involved in many cellular abnormalities that are known to participate in SLE pathogenesis as illustrated in Figure 7 . Ets-1 is a negative regulator of terminal differentiation of B cells and plays critical roles in maintaining B cell identity [30]–[32] . Ets-1-deficient B cells were present in normal numbers but have a large proportion of IgM plasma cells [33] . Ets-1 blocks the function of B-lymphocyte-induced maturation protein 1 ( Blimp-1 ) , an essential transcription factor for plasma cells [34] . The number and frequency of plasma cells were known to correlate with disease activity and the titer of anti-dsDNA antibodies in SLE [35] , [36] . Ets-1 is also a negative regulator of Th17 cell differentiation , and naïve CD4+ T cells deficient in Ets-1 undergo greatly enhanced differentiation into Th17 cells when cultured in vitro under Th17-skewing conditions [37] . Th17 cells with specificity for self-antigens are known to be highly pathogenic and lead to the development of inflammation and severe autoimmunity [38] . Higher plasma IL-17 , IL-23 and higher number of Th17 cells in SLE patients were reported and correlated positively with SLE disease index ( SLEDAI ) [39]–[41] . IL-17 and IL-21 produced by Th17 cells may also induce B cell terminal differentiation [42]–[44] . In this study , we found that in the PBMC of healthy individuals , expression of ETS1 from the risk “A” allele is reduced comparing to that from the “G” allele . The expression level of ETS1 may be tightly regulated . It was shown that resting T cells express high levels of ETS1 mRNA and protein , which decreased to very low levels upon T cell activation [45] . Lower expression of ETS1 for the risk allele carriers may play a role in disease pathogenesis through increased differentiation and activity of both plasma cells and Th17 cells . It is likely that the association SNPs identified in this study may affect the response of ETS1 gene to other upstream signals . SNP rs1128334 locates in the 3′-UTR of ETS1 , and rs10893872 is in absolute LD with rs4937333 , another SNP that is also located in the 3′-UTR of the gene and both SNPs are on putative microRNA ( miRNA ) binding sites . In a recent study by Du et al , the expression level of a microRNA , miR-326 , was found to be related to disease severity in patients with multiple sclerosis and mice with experimental autoimmune encephalomyelitis . ETS1 was shown to be the major target of miR-326 , through downregulation of which miR-326 promoted the generation of Th17 cells both in vitro and in vivo [46] . Another microRNA , miRNA-146a , was also found to be involved in SLE pathogenesis [47] . WDFY family member 4 ( WDFY4 , NCBI GeneID: 57705 ) codes for a huge protein ( 3184 amino acid residues ) with unknown function . Its closest paralog is WD repeat and FYVE domain containing 3 ( WDFY3 , NCBI GeneID: 23001 ) . Similar to WDFY3 , WDFY4 does contain WD40 domains and a BEACH ( Beige and chediak-kigashi ) domain , although FYVE zinc finger domain is truncated . Its sequence is well conserved among various species . For example , there is an 84% sequence similarity between the human protein and its orthologs from Bos Taurus and Canis lupus familiaris . And there is a 42% identity between the human protein and protein XP_701288 . 3 in Danio Rerio ( zebra fish ) . Although protein XP_701288 . 3 is annotated as WDFY3 in NCBI , it has a higher sequence similarity with human WDFY4 than with WDFY3 . WD40 domain is found in a number of eukaryotic proteins that cover a wide variety of functions including adaptor/regulatory modules in signal transduction , pre-mRNA processing and cytoskeleton assembly , while the BEACH domains are implicated in membrane trafficking . While very little is known about the potential function of this well conserved protein , an interesting phenomenon is that the gene is predominantly expressed in immune tissues such as lymph node , spleen , thymus and tonsil ( UniGene Hs . 287379 , http://www . ncbi . nlm . nih . gov/UniGene/ESTProfileViewer . cgi ? uglist=Hs . 287379 ) , unlike WDFY3 , which is expressed in a wide variety of tissues and organs . After submission of our work , Han et al reported their GWAS work on SLE on Chinese populations [18] and reported genome-wide significant association signals on ETS1 ( rs6590330 , downstream of the gene ) and WDFY4 ( rs1913517 , in the intron of both WDFY4 and LRRC18 , leucine rich repeat containing 18 ) . Independent identification of these two genes in SLE susceptibility underlined the validity of the findings . Identifying susceptibility genes provided a good start in our effort to elucidate the disease mechanisms and the functions of the genes involved , although many questions remain to be answered . Characterization of the functions of the genes in autoimmunity may eventually help us to translate genetic findings into clinical and therapeutic applications .
This study was conducted according to the principles expressed in the Declaration of Helsinki . The Hong Kong study was approved by the Institutional Review Board of the University of Hong Kong and Hospital Authority , Hong Kong West Cluster , New Territory West Cluster , and Hong Kong East Cluster . The study on Shanghai , Anhui and Thai samples was approved by the Institutional Review Board of Renji Hospital , Research Ethics Committee of Anhui Medical University and the Ethics Committee of the Faculty of Medicine , Chulalongkorn University , respectively . All patients provided written informed consent for the collection of samples and subsequent analysis . 1073 SLE samples collected in Hong Kong were from four hospitals in Hong Kong Island and the New Territories: Queen Mary Hospital , Tuen Mun Hospital , Queen Elizabeth Hospital and Pamela Youde Nethersole Eastern Hospital . The patients were all of self-reported Chinese ethnicity living in Hong Kong . The average onset age was 28 years old and the ratio of female to male patients was 9∶1 . About half of the patients had renal involvement , and about 70% tested positive for anti-dsDNA antibodies . The SLE samples collected in Shanghai were patients attending Renji Hospital of Jiaotong University Medical School , a tertiary referral hospital covering Shanghai and the surrounding areas . There is an 8∶1 female to male patient ratio , and about 52% of patients have lupus nephritis . 951 SLE patients collected in Anhui were all self-reported Chinese ethnicity living in Anhui province , central China . They were recruited from the Departments of Rheumatology at Anhui Provincial Hospital and the First Affiliated Hospital of Anhui Medical University , both located in Hefei , Anhui province , about 450 km from Shanghai . The average onset age was 31 years old and the ratio of female to male patients was 17∶1 . 314 Thai patients with SLE ( female∶male ratio = 14∶1 ) attending King Chulalongkorn Memorial Hospital , a tertiary referral center in Bangkok were also recruited in this study . Medical records were reviewed to confirm that all subjects met the revised criteria of the American College of Rheumatology for SLE diagnoses [48] . Controls used in the GWAS stage were from both healthy individuals and from other studies conducted in the same institution genotyped with the same platform . For the replication stage , Hong Kong controls were healthy blood donors kindly contributed by the Hong Kong Red Cross and were all of self-reported Chinese ethnicity living in Hong Kong . Controls from Shanghai and Anhui were selected from a pool of healthy blood donors recruited from Renji Hospital ( Shanghai ) and Hefei City ( Anhui ) , respectively , with an effort to match for the age and sex of corresponding SLE patients . Thai controls were recruited from unrelated voluntary healthy donors from the same ethnic background and geographic area as the Thai SLE patients . 320 ( 27 males , 293 females ) SLE patients were genotyped by Illumina 610-Quad Human Beadchip with a total number of SNPs reaching 620 , 901 . 24 individuals were removed due to low call rate or hidden first degree relationship . A total number of 104 , 395 SNPs were also removed from the initial analysis on the ground of low genotyping call rate ( <90% , 30 , 133 SNPs ) and/or low minor allele frequency ( MAF ) ( <0 . 005 , 102 , 970 SNPs ) . 2 , 285 SNPs were also removed due to violation of Hardy-Weinberg equilibrium in controls ( P values<0 . 0001 ) . After quality control measures , 314 case ( 27 males , 287 females ) and 1484 controls ( 840 males and 644 females ) were analyzed on a total of 514 , 221 SNPs . The call rate for the remaining SNPs reached 0 . 999 with a genome-wide inflation factor of 1 . 03 , which is an indication of good match between the cases and controls . To overcome any potential effect from the heterogeneity in the controls , three independent comparisons were initially conducted by separating the controls into three different subsets and only SNPs reaching significance in all subsets in association analysis were selected for further replication . The SNPs were analyzed for association with the disease by means of comparison of the minor allele frequency in patients and controls ( basic allelic test ) as well as other tests using Plink [49] . Linkage disequilibrium ( LD ) patterns were analyzed and displayed by HaploView [50] . Association of the SNPs with disease risk was also corrected by logistic regression using age and sex as covariates . Average odds ratios ( OR ) and P values jointly analyzed from four sample collections were obtained by Cochran-Mantel-Haenszel ( CMH ) test of disease association conditional on SNP frequency differences among different populations . Test of independent contributions of a SNP controlling for the effect of other SNPs in the same locus was done by conditional logistic regression as well as haplotype analyses . Subphenotype stratification was performed by comparing cases with and without a given subphenotype . SNPs rs1128334 , rs10893872 , rs7097397 , rs10857650 , and rs877819 were genotyped by TaqMan SNP genotyping method using assay-on-demand probes and primers ( Applied Biosystems , Foster City , CA94404 , USA ) . Some of the initial screening was also done using Sequenom MassARRAY iPLEX Gold system . Genotyping accuracy was confirmed by direct sequencing of PCR products for some randomly chosen samples . Genotyping concordance between Illumina Human 610-Quad Beadchip and TaqMan SNP genotyping method was also examined on selected samples and probes: rs877819 has a concordance rate of 99 . 64% ( 1 out of 277 differed by the two platforms ) ; rs10893872 has a concordance rate of 99 . 16% ( 1 out of 119 samples differed ) . SNP rs1128334 and rs7097397 were examined by direct sequencing of selected samples and showed complete consistence between TaqMan and sequencing ( 50 samples each ) . The results of rs10857650 were discarded due to low concordance between the Illumina Beadchip data and the TaqMan results . Thirty-three healthy individuals heterozygous for rs1128334 were chosen to assess the relative ETS1 mRNA levels from the two alleles , “A” and “G” , by pyrosequencing [51] , [52] . In the meantime , DNA detection ratio was used as a control for amplification efficiency . Briefly , total RNA was extracted from peripheral blood mononuclear cell ( PBMC ) from each individual . RNA samples were then treated with DNAase to eliminate genomic DNA contamination before being reverse-transcribed into cDNA using oligo-dT primer . cDNA was then amplified by PCR using transcript-specific primers , together with DNA from the same individuals . The cDNA and DNA PCR products were purified using the Qiaquick PCR purification kit , and then subjected to allele quantitative pyrosequencing . The sequencing primer was designed using Pyrosequencing Assay Design Software v . 1 . 0 . Reactions were performed on a Biotage PSQ96MA machine , and allele quantification was analyzed using PSQMA 2 . 1 software . The average G/A cDNA expression ratio of each individual was normalized by the G/A DNA ratio from the same sample . Paired student' t test was used to compare the normalized expression level from the “A” and “G” alleles . | In this study , we first conducted a genome-wide association study in a Hong Kong Chinese population , followed by replication in three other cohorts from Mainland China and a cohort from Thailand , which totaled 3 , 300 Asian patients and 4 , 200 ethnically and geographically matched controls . We identified novel variants in ETS1 and WDFY4 associated with SLE with genome-wide significance and confirmed the association of HLA locus , STAT4 , BLK , IRF5 , BANK1 , TNFSF , and IRF5 with the disease . ETS1 encodes a critical transcription factor involved in Th17 and B cell development . Allelic expression study showed a significantly lower expression of ETS1 from the risk allele , which provided functional support to the genetic findings . WDFY4 is a huge protein with unknown function but is predominantly expressed in primary and secondary immune tissues , and a nonsynonymous SNP in this gene was found to be highly associated with SLE susceptibility . Our findings shed new light on the function of these genes as well as the mechanism of this devastating disease . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genetics",
"and",
"genomics/genomics",
"genetics",
"and",
"genomics/gene",
"discovery",
"genetics",
"and",
"genomics/gene",
"expression",
"genetics",
"and",
"genomics/genetics",
"of",
"the",
"immune",
"system",
"genetics",
"and",
"genomics/complex",
"traits",
"rheumatology/systemic",
"lupus",
"erythematosos",
"genetics",
"and",
"genomics/genetics",
"of",
"disease",
"immunology/genetics",
"of",
"the",
"immune",
"system",
"immunology/autoimmunity",
"genetics",
"and",
"genomics/bioinformatics",
"genetics",
"and",
"genomics/medical",
"genetics",
"genetics",
"and",
"genomics/population",
"genetics"
] | 2010 | Genome-Wide Association Study in Asian Populations Identifies Variants in ETS1 and WDFY4 Associated with Systemic Lupus Erythematosus |
High resolution structures of antibody-antigen complexes are useful for analyzing the binding interface and to make rational choices for antibody engineering . When a crystallographic structure of a complex is unavailable , the structure must be predicted using computational tools . In this work , we illustrate a novel approach , named SnugDock , to predict high-resolution antibody-antigen complex structures by simultaneously structurally optimizing the antibody-antigen rigid-body positions , the relative orientation of the antibody light and heavy chains , and the conformations of the six complementarity determining region loops . This approach is especially useful when the crystal structure of the antibody is not available , requiring allowances for inaccuracies in an antibody homology model which would otherwise frustrate rigid-backbone docking predictions . Local docking using SnugDock with the lowest-energy RosettaAntibody homology model produced more accurate predictions than standard rigid-body docking . SnugDock can be combined with ensemble docking to mimic conformer selection and induced fit resulting in increased sampling of diverse antibody conformations . The combined algorithm produced four medium ( Critical Assessment of PRediction of Interactions-CAPRI rating ) and seven acceptable lowest-interface-energy predictions in a test set of fifteen complexes . Structural analysis shows that diverse paratope conformations are sampled , but docked paratope backbones are not necessarily closer to the crystal structure conformations than the starting homology models . The accuracy of SnugDock predictions suggests a new genre of general docking algorithms with flexible binding interfaces targeted towards making homology models useful for further high-resolution predictions .
High resolution structures of protein-protein complexes are necessary for understanding mechanisms of protein-protein interactions , analyzing mutations , and manipulating binding affinity [1] . The large gap between the number of experimentally determined complex structures and the available sequences of pairs of protein complexes underscores the challenges , time required and cost of x-ray crystallography or nuclear magnetic resonance approaches . The paucity in complex structures can be alleviated by computational docking , i . e . , the prediction of protein-protein complexes , which potentially provides a fast and efficient alternative route . To predict the structure of a protein-protein complex , computational docking requires the structures of the interacting partners . However , sometimes even the structures of the monomeric units are unavailable , forcing the use of a homology modeled structure for one or both partners [2] , [3] . Given the inaccuracies in a homology model , current computational docking strategies can determine the gross structural features of a complex , but find it exceedingly challenging to successfully predict high resolution structures of such protein-protein complexes , pointing to the need to develop new docking algorithms which incorporate the necessary degrees of freedom to compensate for the inaccuracies . Antibody-antigen ( Ab-Ag ) complexes provide a model system where much needed high-resolution computational docking predictions are challenged by inaccuracies in antibody homology models . The selection of antibodies for exploring homology model docking simplifies the problem by isolating the various degrees of freedom according to prior knowledge of the uncertainties in an antibody homology model , namely the conformation of the complementarity determining region ( CDR ) loops ( L1 , L2 , L3 in the light chain , and H1 , H2 , H3 in the heavy ) [4] , [5] , the hyper-variability of the CDR-H3 loop [6]–[9] , and the relative orientation of the antibody light ( VL ) and heavy ( VH ) chains [6] , [10]–[12] . A recent study by Narayanan et al found that the VL-VH relative orientation has a significant impact on the antigen binding properties of an antibody [10] , suggesting that simultaneous optimization of the VL-VH relative orientation and antibody-antigen relative orientation might capture some of the intramolecular changes undergone by an antibody upon antigen binding . An additional motivation for studying antibody-antigen complexes is that therapeutic antibodies are revolutionizing healthcare [13] . Oncology , arthritis , immune and inflammatory disorder treatments have benefitted from newly developed therapeutic antibodies [14] . Success of several therapeutic antibody drugs has relied on homology modeling . According to Schwede et al . , homology modeling played an important role in the development of 11 of the first 21 marketed antibodies including Zenapax ( humanized anti-Tac or daclizumab ) , Herceptin ( humanized anti-HER2 or trastuzumab ) , and Avastin ( humanized anti-VEGF or bevacizumab ) [15] . High-resolution computational docking can aid in the design of antibody biologics by providing insights into the complex interactions between an antibody and an antigen [16] . The importance of antibodies combined with the detailed knowledge of the flexibility in the various segments of an antibody fragment variable ( FV ) region make them ideal candidates for the development of novel flexible docking algorithms . Although there are currently no flexible docking algorithms tailored for antibodies , there have been several efforts to incorporate some of the relevant modes of internal flexibility during docking . Early [17] , [18] and more recent approaches [19] use hinges to account for internal flexibility . Multi-body docking , which might be useful for optimizing assembly of VL and VH and antigen chains , has been explored in a few case studies [20] , [21] including some targets in the blind prediction challenge known as the Critical Assessment of PRediction of Interactions ( CAPRI ) [22] . Another genre of docking algorithms like HADDOCK [23] allows flexibility in the side chains and backbones to allow for conformational rearrangements in the interaction surface . Wang et al . 's modifications to RosettaDock allow simultaneous gradient-based minimization of the backbone torsional angles and the protein-protein rigid-body orientation [24] . We recently developed a RosettaDock generalization called EnsembleDock [25] that follows the conformer-selection plus induced-fit model [26] to additionally enable docking of a pre-generated ensemble of structures . We also recently developed RosettaAntibody [27] , an antibody FV region homology modeling protocol which incorporates refinement of VL-VH relative orientation , CDR H3 loop modeling and minimization of all the CDR loops . RosettaAntibody generates ten antibody homology models for each input sequence , and this set of models can be used simultaneously with EnsembleDock . However , errors in the CDRs of RosettaAntibody homology models ( particularly H2 and H3 ) can still frustrate docking , and only the ten pre-generated backbone conformations are sampled during ensemble docking [27] . In this paper , we discuss the development and implementation of SnugDock , a new antibody docking algorithm built upon RosettaDock using the sampling components of RosettaAntibody . The new protocol performs multi-body docking by allowing simultaneous structural optimization of the relative orientations of antibody-antigen and VL-VH . SnugDock simulates induced fit by additional simultaneous optimization of the binding interface by allowing flexibility of CDR loops and interfacial side chains . Moreover , we combine SnugDock with EnsembleDock to synthesize a new docking algorithm that encompasses conformer selection , multi-body docking , and a flexible paratope . We test SnugDock using antibody homology models obtained from two accessible public servers , namely RosettaAntibody [28] and Web Antibody Modeling ( WAM ) [29] . Our goal is to achieve reliable high-resolution antibody docking starting from only the antigen unbound structure and the antibody sequence .
A longstanding source of homology models is the Web Antibody Modeling server ( WAM ) created by Whitelegg and Rees [29] . The WAM server grafts antibody components together and models the H3 loop de novo . SnugDock may be able to compensate for the model errors during docking . Ensemble docking is not possible since WAM returns only one model for a given sequence . Table 2 presents the accuracy of docking predictions obtained by using WAM antibody homology models using both standard RosettaDock and using SnugDock . The lowest-interface-energy docking decoy generated by standard rigid-body docking simulations using WAM homology models are almost all incorrect . Subjecting the WAM models to SnugDock resulted in six medium quality lowest-interface-energy docking decoys . Thus SnugDock recovers more accurate docking predictions . The original WAM homology models showed strain in the molecule reflected in high Rosetta scores . By subjecting the homology model to minimization on the paratope degrees of freedom , SnugDock was able to relieve intramolecular and inter-chain steric clashes . Interestingly , the SnugDock results with WAM models are comparable to those with RosettaAntibody models , while the use of EnsembleDock-plus-SnugDock with the RosettaAntibody models achieves higher accuracy . Local docking is often of interest in antibody applications since epitope information can be obtained by a variety of other methods . However , global docking is a computational alternative for producing epitope information when it is unknown . Global docking can be significantly more challenging because of the larger conformational space to search . Further , flexible docking creates additional danger of creating an unrealistic induced fit at a non-native docking location , resulting in a false positive prediction [41] , [42] . Global docking is considerably more computationally demanding , and thus we restricted our tests of global docking to five targets , and to simulate a practical docking application , chose only those targets for which unbound crystal structures were available for both the antibody and the antigen: 1MLC , 1AHW , 1JPS , 1WEJ and 1VFB . The starting structures consisted of the unbound crystal structure of the antigen and the lowest-interface-energy RosettaAntibody homology model . The EnsembleDock and the EnsembleDock-plus-SnugDock protocols used the ten lowest-interface-energy RosettaAntibody homology models . For each target , we generated 5000 candidate structures , with each prediction run beginning from a random global rotation of the antigen and a small perturbation of the antibody ( to keep the paratope generally directed toward the antigen ) . Global docking using standard rigid-body RosettaDock generated no acceptable quality lowest-interface-energy decoys for any targets , and a top-ten ranked acceptable decoy for only one target ( Table 3 ) . Using EnsembleDock or SnugDock independently produced marginal improvement with a few acceptable quality predictions . The combination algorithm of EnsembleDock-plus-SnugDock generated two medium quality lowest-interface-energy predictions exhibiting the synergy already established in the local docking simulations . Additionally , the most native-like decoy in the ten lowest-interface-energy decoys was of medium quality for three of the five targets and acceptable for one target . The results are comparable to global docking using standard rigid-body RosettaDock with unbound crystal starting structures of the antibodies , where one structure had a high quality ( 1JPS ) , two had medium quality , and the others had acceptable quality prediction for the most native-like decoy in the ten lowest-interface-energy decoys . The results can also be compared with local docking using EnsembleDock-plus-SnugDock around the known epitope ( Table 1 ) , where four of the five targets had at least a medium quality prediction for the most native like decoy in the ten lowest-interface-energy decoys , and the fifth target ( 1VFB ) had an acceptable prediction . Thus one target ( 1VFB ) which had succeeded in local docking failed in global docking due to low-scoring non-native decoys ( see docking energy landscapes , Supporting Information Figure S2 ) . In general , addition of SnugDock increases sampling of more native-like decoys , enabling near-natives to be energetically more favorable . SnugDock is the first docking algorithm with targeted antibody flexibility . SnugDock models flexible loop conformations , backbone motions , and inter-chain ( VH-VL ) adjustments . The introduction of flexibility during docking is critical to overcome the inaccuracies inherent in a homology modeled antibody structure . Comparison of algorithms shows that increasing the degrees of freedom in local docking gradually increases the quality of predictions . Ultimately , EnsembleDock-plus-SnugDock with homology models achieves accuracy comparable to docking crystal structures with standard RosettaDock . While the algorithm is limited to antibody-antigen interactions , the results suggest that it is possible for computational predictions to use homology models to bridge the gap between the number of experimentally determined complex structures and the available sequences of pairs of interacting proteins . In CAPRI rounds 1–18 , eleven of the forty targets involved docking of at least one homology modeled partner ( both partners were homology models for Target 35 ) [43]–[45] . For six of the eleven targets , none of the participating groups could predict any medium or higher accuracy structures . When the sequence identity was under 40% , the solutions were of acceptable quality at best ( Targets 20 , 24 ) . High quality predictions were obtained only for two cases ( Targets 14 , 19 ) and in both cases the binding region of the template structure was structurally similar to the co-crystal structure [44] , and the other docking partner was in the bound conformation . The poor performance of homology modeled docking partners in CAPRI highlights the need of docking algorithms like SnugDock which are robust to inaccuracies in a homology model . Targets 20 and 24 with only acceptable predictions had poorly modeled loop and C-terminal regions which were responsible for key contacts in the native binding interface [45] , showing that using homology models with loops at the binding region makes docking with homology models even harder . SnugDock with its loop relaxations at the binding interface addresses the challenge toward accurate high-resolution predictions . The CDR H3 loop of antibodies provides the most diversity and is thus a focus of the conformational sampling in the SnugDock algorithm . In our experience with the RosettaAntibody Server [28] , there are antibodies with non-H3 loops which elude the traditional Chothia classification system [4] and may not fit into canonical CDR templates . The SnugDock algorithm is easily generalizable to include perturbations of loop conformations for any of the six CDR loops . Extra sampling however should be restricted to special cases for efficiency and to avoid issues with over-optimized , non-native induced-fit structures . For approaching non-antibody targets , the SnugDock methods would need to be adapted requiring knowledge of a binding site and appropriate choices of loops to target flexibility . The multi-chain docking methods can be applied to any multi-chain docking partner . The flexible docking methods help in identifying the correct docked complex structure , but unfortunately they do not yet help in refining the monomer homology structures themselves closer to the crystal backbone conformations . This limitation likely arises from the vast conformational space of the backbone and the difficulties with high-resolution refinement of protein structures [46] , [47] . In docking , some of the energetic issues are avoided through the use of the interface energy instead of the total energy . Further advancements in refinement techniques will be needed to address this shortcoming . SnugDock's advancements in the sampling problem also reveal continuing issues with the knowledge of nature's energy function . Missing water molecules affected the prediction of targets 1ZTX and 1VFB . Antibody interfaces in general are polar [48] , and several targets with the most polar interfacial CDRs ( 1VFB , 1JHL , 1NCA ) failed perhaps due to the challenges in modeling electrostatics [49] . Experimental techniques for epitope mapping like hydrogen-deuterium mass spectroscopy [50] can help to pre-orient an antigen for local docking . However , when such data are not available , one must resort to global docking where the docking simulations are started with random orientations of the docking partners . Our limited testing of global docking encouragingly suggests that the EnsembleDock-plus-SnugDock approach can successfully find epitopes . Global searches should still be performed with care as the large conformation space can frustrate the ability to find the native binding interface or obscure it through false positive non-native interaction . One could envision a complete computational antibody engineering pipeline starting from the antibody sequence and ending with accurate predictions for optimized antibody-antigen interactions . In this paper we have been successful in reaching the second step by computational docking using computational models of the monomer antibody . The next stages may be additionally challenging . High-resolution complex structures might next be used for computational alanine scanning [51] , computational affinity maturation [52] or alteration of binding specificity [53] . For antibody therapeutics , structures will help define drug mechanisms for regulatory approval [15] , enable epitope mapping [54] and humanize constructs [55] . These applications require varying amounts of resolution and further testing will reveal the full utility of the SnugDock predictions .
To compare with prior work , we use the set of fifteen antibody-antigen complexes as tested in the original RosettaAntibody publication [27] . The antibody-antigen complex dataset was compiled to ensure: 1 ) a fair representation of unbound-unbound , unbound-bound and bound-bound antibody-antigen docking targets , 2 ) a spectrum of CDR H3 loop lengths ( 7 to 11 residues ) and 3 ) both old and newly released crystal complexes ( PDB release dates 1992–2006 ) . The RosettaAntibody and the WAM structures are as reported previously [27] . SnugDock is implemented in the Rosetta biomolecular modeling suite . Fold trees objects [24] are used to guide the propagation of structural changes during docking with backbone flexibility . One fold tree uses flexible jumps for moving the VL-VH and antibody-antigen pairs relative to each other . A second fold tree for CDR loop relaxation had fixed jumps joining the loop stems , and cuts at the middle of the loops . Move map objects are used to select particular sets of residues for backbone and/or side-chain torsion angle flexibility . Figure 1 depicts the flowchart for the steps in the SnugDock protocol , implemented as follows . Steps 1 and 2 describe the initial setup , Steps 3–6 describe the low resolution stage and steps 7–13 describe the high resolution stage . Each decoy of an independent docking simulation begins from a different random starting position . In local docking , the set of starting positions comprises a diffuse cloud that covers a reasonable area ( ∼20 Å rmsd ) with moderate density around the native ligand position . In local and global docking , 1 , 000 and 5 , 000 candidate docking structures are generated for each target respectively . In our empirical tests , 5 , 000 decoys were sufficient and results were similar for test runs of 10 , 000 decoys in a subset of targets . The energy function used during the course of the simulation is as described previously [35] with ( i ) the interfacial terms of the scoring function including both the antibody-antigen interface and the VL-VH interface , and ( ii ) chain break penalties for six CDR loops . Interface energy [24] is used to rank and discriminate the structures produced by the docking simulations . Ligand rmsd is the deviation of the N , Cα , C and O backbone atoms of the antigen after superposition of the antibody backbone atoms . Interface rmsd is the deviation of the backbone atoms at the interface after optimal superposition of those same atoms , where the interface is defined as all residues within 10 Å of a non-hydrogen atom of the other docking partner . Interface energy is the component of the total docking score that arises from inter-molecular residue-residue interactions at the antibody-antigen interface . For fnat calculations , residue-residue contacts are defined when a residue is within 5 Å of a non-hydrogen atom from the other docking partner . The docking models are assigned CAPRI [22]-style “high” , “medium” , “acceptable” or “incorrect” rankings that depend on the rmsd-to-native of the ligand position , the interface rmsd to native and the fraction of native residue-residue contacts ( fnat ) that are recovered in the docked model [59] . Convergence of a docking simulation is indicated by the presence of a docking funnel , which is defined to exist if there are at least five medium quality predictions in the ten lowest-energy docking decoys . The SnugDock method presented here is freely available for academic and non-profit use as part of the Rosetta structure prediction suite at www . rosettacommons . org . The distribution includes documentation and full source code . The Rosetta version numbers and command lines used to generate the data are provided in Supporting Information Text S1 . | Antibodies are proteins that are key elements of the immune system and increasingly used as drugs . Antibodies bind tightly and specifically to antigens to block their activity or to mark them for destruction . Three-dimensional structures of the antibody-antigen complexes are useful for understanding their mechanism and for designing improved antibody drugs . Experimental determination of structures is laborious and not always possible , so we have developed tools to predict structures of antibody-antigen complexes computationally . Computer-predicted models of antibodies , or homology models , typically have errors which can frustrate algorithms for prediction of protein-protein interfaces ( docking ) , and result in incorrect predictions . Here , we have created and tested a new docking algorithm which incorporates flexibility to overcome structural errors in the antibody structural model . The algorithm allows both intramolecular and interfacial flexibility in the antibody during docking , resulting in improved accuracy approaching that when using experimentally determined antibody structures . Structural analysis of the predicted binding region of the complex will enable the protein engineer to make rational choices for better antibody drug designs . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Methods"
] | [
"immunology/antigen",
"processing",
"and",
"recognition",
"biophysics/protein",
"folding",
"biochemistry/protein",
"folding",
"biophysics/theory",
"and",
"simulation",
"pharmacology/drug",
"development",
"molecular",
"biology/bioinformatics",
"biophysics/biomacromolecule-ligand",
"interactions",
"biochemistry/bioinformatics",
"computational",
"biology/protein",
"structure",
"prediction",
"biochemistry/theory",
"and",
"simulation",
"biochemistry/drug",
"discovery"
] | 2010 | SnugDock: Paratope Structural Optimization during Antibody-Antigen Docking Compensates for Errors in Antibody Homology Models |
Efficient acquisition of extracellular nutrients is essential for bacterial pathogenesis , however the identities and mechanisms for transport of many of these substrates remain unclear . Here , we investigate the predicted iron-binding transporter AfuABC and its role in bacterial pathogenesis in vivo . By crystallographic , biophysical and in vivo approaches , we show that AfuABC is in fact a cyclic hexose/heptose-phosphate transporter with high selectivity and specificity for a set of ubiquitous metabolites ( glucose-6-phosphate , fructose-6-phosphate and sedoheptulose-7-phosphate ) . AfuABC is conserved across a wide range of bacterial genera , including the enteric pathogens EHEC O157:H7 and its murine-specific relative Citrobacter rodentium , where it lies adjacent to genes implicated in sugar sensing and acquisition . C . rodentium ΔafuA was significantly impaired in an in vivo murine competitive assay as well as its ability to transmit infection from an afflicted to a naïve murine host . Sugar-phosphates were present in normal and infected intestinal mucus and stool samples , indicating that these metabolites are available within the intestinal lumen for enteric bacteria to import during infection . Our study shows that AfuABC-dependent uptake of sugar-phosphates plays a critical role during enteric bacterial infection and uncovers previously unrecognized roles for these metabolites as important contributors to successful pathogenesis .
Competition between host cells and pathogenic bacteria for diverse substrates ranging from metals to vitamins can often influence the outcome of infection . As a result , pathogens have developed a variety of nutrient uptake mechanisms to increase their competitiveness and hence their ability to colonize and successfully infect their hosts [1] . One such example of these nutrient uptake mechanisms are the binding-protein dependent transporters ( BPDTs ) . BPDTs are ubiquitous in bacteria and rely on the presence of a soluble ligand binding protein that directs a specific substrate to an integral membrane permease/ATPase complex . The ensuing tripartite complex then couples the hydrolysis of cytosolic ATP to the influx of the ligand into the cytosol for downstream uses [2] . BPDTs target diverse ligands such as metal ions , amino acids and sugars and are recognized as both virulence determinants and therapeutic targets in many bacterial infections [2] . In this study , we investigate AfuABC ( Actinobacillus ferric uptake ABC ) , a BPDT originally identified as an iron-specific transporter in Actinobacillus pleuropneumoniae , a Gram-negative upper respiratory tract pathogen [3] . The locus afuABC encodes 3 polypeptides corresponding to a conventional BPDT–AfuA ( a periplasmic binding protein or PBP ) , AfuB ( an inner membrane permease ) , and AfuC ( a cytosolic ATPase ) . Although AfuABC is annotated as an Fe3+-binding transporter , its ligand has been the subject of controversy [4–6] . Initially of interest given the central role of iron piracy to the success of bacterial pathogens , AfuABC has not been investigated on a functional level and the role of currently annotated AfuABC homologues in virulence of other bacterial species has not been characterized . Since several studies have identified afuABC as transcriptionally upregulated during bacterial infection [7–9] , we investigated AfuABC on a structural and functional basis to definitively identify its ligand and further illuminate its role in pathogenesis . We found that AfuABC is not an iron transporter as previously described and instead that this BPDT is specific for phosphorylated sugars . AfuABC contributed to virulence and was key to transmission success in a mouse model of C . rodentium infection , highlighting its role in nutrient uptake during bacterial pathogenesis .
To structurally characterize the ligand of AfuA , a 1 . 6Å crystal structure of ligand-bound A . pleuropneumoniae AfuA purified from E . coli was solved by use of sulfur single wavelength anomalous diffraction . AfuA is a class II/cluster D periplasmic binding protein with two globular α/β domains linked by a dual β-stranded hinge [10] ( S1 Table and Fig 1A ) . In the binding cleft , we unexpectedly observed positive electron density and identified glucose-6-phosphate ( β-G6P ) as the bound ligand with a small ( ~5% ) portion of density at carbon 2 corresponding to mannose-6-phosphate ( β-M6P ) ( Fig 1B ) . In the binding cleft , His205 , Asp206 and Glu229 interact multivalently with the sugar ring while Ser37 and Thr150 form hydrogen bonds with the phosphate moiety of G6P ( Fig 1B ) . Isothermal titration calorimetry ( ITC ) validated the binding affinity of A . pleuropneumoniae AfuA for β-G6P at 24 nM with favourable values of enthalpy and entropy change ( S2 Table and Fig 1C and 1D ) . In order to probe the ligand-binding site of AfuA , site-directed mutants were generated for putative binding residues . The mutations T150A and S37A lowered but did not eliminate binding to G6P , whereas mutations to the sugar-binding residues ( H205A , D206A and E229A ) completely abrogated the interaction ( Table 1 ) . This indicates that the hydrogen-bonding capacity of the sugar-coordinating residues is the critical determinant in mediating binding of potential sugar phosphates . To gain insight into the mechanism of ligand capture by this PBP , a structure of apo ( open ) -AfuA was solved to 1 . 6Å . The globular domains of apo-AfuA are held at a 29 . 7° angle relative to the closed form ( Fig 1A ) . The binding pocket of AfuA contains both electropositive and electronegative regions , corresponding to the regions of the pocket that bound the phosphate or sugar moiety of G6P respectively ( Fig 2A , 2B and 2C ) . A subsequent ligand screen by ITC with structurally similar molecules revealed that three other sugar-phosphates including fructose-6-phosphate ( F6P ) , sedoheptulose-7-phosphate ( S7P ) and M6P had affinities of 8 , 57 and 960 nM respectively for AfuA ( Table 2 and S1 Fig ) . F6P binding affinities of the single site AfuA mutants displayed the same trend as for G6P , suggesting a similar binding mechanism between the two ligands ( S3 Table ) . Structures of AfuA bound to F6P ( 2 . 0Å ) and S7P ( 1 . 3Å ) were also solved to confirm ligand coordination . Similar to the G6P structure , both F6P and S7P were found in the β-anomer conformation ( Fig 2D and 2E ) . Between the three holo-AfuA structures , strand and amino acid placement and conformation were identical ( RMSD < 0 . 4Å between Cα chains ) . Interestingly , S7P was captured by AfuA in its furan form and not pyran as typically depicted . S7P and its derivatives can exist as furan forms in vivo [11] , indicating that this ligand structure is biologically relevant . AfuA did not bind other sugar-phosphates such as glucose-1-phosphate and ribose-5-phosphate , demonstrating a high level of selectivity for ligands ( Table 2 and S1 Fig ) . No interaction between AfuA and glucose , fructose or inorganic phosphate could be detected ( Table 2 and S1 Fig ) . These data collectively show that AfuA is a PBP that is highly specific for four sugar-phosphates involved in central metabolism . Since AfuA binds sugar-phosphates , we queried whether AfuABC could recapitulate sugar-phosphate transport in a bacterium that could not transport these substrates . The only other sugar-phosphate uptake system identified thus far in Gram-negative bacteria is the two component system UhpABCT . UhpT is an inner membrane antiporter that couples the influx of hexose phosphates such as G6P and F6P to the efflux of inorganic phosphate [12] . The expression of UhpT is regulated by the transcription factor UhpA , which is activated in response to periplasmic sugar-phosphates through the inner membrane sensor kinase complex UhpB/UhpC [13] . E . coli ΔuhpT were unable to replicate in medium where G6P or F6P were the only carbon source , confirming UhpT is the only hexose-phosphate specific transporter in BW25113 ( Figs 3A , 3B and S2 ) [14] . When ΔuhpT cells were transformed with a vector carrying the A . pleuropneumoniae afuABC operon , a significant increase in growth rate on M9 + G6P or F6P agar media compared to empty vector-transformed cells was observed ( S4 Table and Figs 3A and S2A ) . This growth complementation was reproduced with liquid media of the same supplementation ( Fig 3B ) . These results demonstrate that AfuABC is sufficient to transport G6P and F6P as substrates for bacterial metabolism and growth . We did not observe complementation with afuA or afuBC alone , consistent with a model where AfuABC acts as a BPDT to transport substrates from the periplasm into the cytosol ( Fig 3A ) . Given the novel function of AfuABC as a sugar-phosphate transporter , we next sought to identify whether this machinery was conserved in other bacteria . The three crystal structures provided a means to identify the sugar-coordinating residues ( H205 , D206 and E229 ) as the critical determinants in mediating binding of sugar-phosphates . These residues were subsequently used to define a potential sugar-phosphate binding motif , which in turn was used to search for and identify putative homologues of AfuA and hence the AfuABC operon ( S5 Table and Fig 4 ) . The majority of hits were found in Gram-negative bacteria . In particular , AfuABC was identified in many human pathogens from the Pasteurellaceae ( e . g . Haemophilus influenzae ) , the Vibrionaceae ( e . g . Vibrio cholerae ) and the Enterobacteriaceae ( e . g . E . coli O157:H7 ) . Surprisingly , a minority of putative afuABC loci were also found in Gram-positive bacteria such as Clostridium tetani and other members of the Firmicutes . The conservation of AfuABC over a broad spectrum of bacterial genera suggests sugar-phosphate transport is required across a wide range of colonization niches and conditions . To examine whether AfuABC plays a role in bacterial pathogenesis , we utilized the intestinal murine pathogen C . rodentium , a commonly used model for EHEC O157:H7 as well as other attaching and effacing pathogens [15] . In EHEC , afuABC is located on the putative pathogenicity island OI-20 . OI-20 also contains a recently identified fucose-sensing two-component regulatory system , suggesting this genomic island may play a role in carbohydrate sensing and/or acquisition [16] . OI-20 is conserved to >83% amino acid identity in C . rodentium ICC168 ( Fig 5A ) , and ITC and growth assays confirmed that C . rodentium AfuABC functioned as a sugar-phosphate transporter [17 , 18] ( S6 Table and S3 Fig ) . We generated an AfuA knockout strain of C . rodentium ( ΔafuA ) which did not display significant changes in planktonic growth , intestinal epithelial adherence/pedestal formation , type III effector secretion and bacterial tissue localization ( Figs 5B–5E and S4 ) . Oral infection of C57BL/6 mice by C . rodentium WT or ΔafuA yielded no significant difference in bacterial colonic tissue burdens between the two groups at 10 days post infection ( dpi ) ( Fig 5F ) . However , C . rodentium is a highly adapted mouse pathogen and strains lacking important virulence factors can still reach high colonic burdens in individual infections [19 , 20] . This prompted us to perform a competitive index ( CI ) assay to measure the comparative fitness between WT and ΔafuA in a simultaneous infection . ΔafuA was significantly impaired compared to WT in stool pellets at 6 dpi ( median CI = 0 . 42 , p = 0 . 031 ) as well as in colonic tissues at 10 dpi ( CI = 0 . 29 , p = 0 . 015 ) ( Fig 5G ) indicating that AfuABC is required for C . rodentium optimal colonization of the intestine . A competitive defect of this magnitude is comparable to those previously reported for C . rodentium strains lacking known secreted effectors [19–21] . The primary role of sugar-phosphates as cytosolic metabolites suggests that they would not be present in the extracellular environment and hence available for an extracellular pathogen such as C . rodentium . However , recent evidence has emerged indicating that they are present in the intestine and their availability is dependent on the presence of the host microbiota [22 , 23] . Since previous studies have not defined the concentration of these molecules , we used targeted liquid chromatography/tandem mass spectrometry to quantify the level of AfuABC substrates accessible to C . rodentium in control and C . rodentium-infected mice colons . Both validated substrates of AfuABC ( G6P and F6P ) as well as the other AfuA ligands ( M6P and S7P ) were present in all locations sampled ( luminal contents , stool pellets and mucus scrapings ) at micromolar quantities ( Table 3 and Fig 6 ) . These amounts exceed the Kd of C . rodentium AfuA for G6P or F6P , indicating that these sugar-phosphates are present in the mouse intestine at levels sufficient to be utilized by the bacteria . Typical assays with C . rodentium involve the oral delivery of large doses of bacteria to ensure reproducible infections . As such , they fail to mimic how C . rodentium and other enteric bacterial pathogens typically spread via the oral-fecal route in wild host populations . The high levels of sugar-phosphates present in shed stool pellets led us to examine a potential role for AfuA in the transmission of C . rodentium in a murine population . Index mice were orally infected with either WT or ΔafuA and after six days , co-housed with two naïve mice for 48 hours to allow transmission via coprophagy ( Fig 7A ) . Seven out of eight co-housed mice exposed to WT-infected index mice were successfully infected , with heavy pathogen burdens recovered from both their colonic tissues and lumen ( Fig 7B and 7C ) . In contrast , only half of the ΔafuA exposed mice showed any detectable C . rodentium within their luminal contents , and only two out of eight mice showed any C . rodentium adherent to their colonic tissues ( Fig 7B and 7C ) . These results indicate that AfuABC does not directly affect the growth or infective capacity of C . rodentium but rather impacts its ability to compete and survive within the host intestinal environment and ultimately transmit to new hosts .
We conclude that AfuABC has been misannotated and is a sugar-phosphate transporter that contributes to the virulence of an enteric pathogen . We have defined a potential sugar-phosphate binding motif in AfuA and used it to identify a broad conservation of AfuABC across bacteria , including both Gram-negative and-positive species . Although AfuA was not required for in vivo colonization by C . rodentium or functional type III secretion , loss of AfuA significantly impacted its ability to be transmitted between hosts , indicating that sugar-phosphate uptake is required for a functional C . rodentium infective life cycle . Competition between the host and bacteria for iron is an established concept in bacterial pathogenesis [24] . Therefore , a locus such as afuABC that had been misannotated as a Fe3+ transporter might have been assumed to contribute to virulence without direct experimental evidence indicating such a function . Previous data implicating AfuABC in iron uptake was based on the afuABC-mediated rescue of ΔaroB ( siderophore-deficient ) E . coli on iron-limited media [3] . However , since aroB encodes a metabolic enzyme that has dual roles in both carbon metabolism ( the shikimate pathway ) and siderophore synthesis , it is not clear whether the afuABC complementation of ΔaroB observed was only attributable to a complementation of iron uptake in this strain . Since the sole function of UhpT is in hexose-phosphate uptake , the assay we used in this study is more specific and directly implicates AfuABC in transport of these substrates instead of iron . Based on the results presented here , AfuABC does contribute to bacterial pathogenesis but this is achieved through sugar-phosphate and not iron uptake . It remains an open question whether other predicted iron transporters may , akin to AfuABC , be misannotated . The structures shown here are the first structural data for a non-enzymatic sugar-phosphate ( i . e . G6P ) binding protein . In AfuA , the essential residues His205 and Asp206 form multivalent hydrogen bonds to the sugar moiety of ligands and function as the primary determinants of ligand selectivity . On the opposing end of the pocket , redundant phosphate-coordinating residues such as Ser37 cooperate to hold the phosphate moiety of ligands . This bipartite binding cleft is able to discriminate between isomers of the same sugar-phosphate ( G6P vs . G1P ) and between differing orientations of side chains around the carbohydrate ring ( F6P vs . R5P ) . Our observation that both furan ( F6P and S7P ) and pyran ( G6P ) ligands could bind AfuA suggests the position of hydroxyl determinants is more important than steric fitting of the overall ring structure into the binding pocket . We propose a two-step mechanism for ligand capture by AfuA . Substrates such as G6P or F6P first dock to the protein via hydrogen bonding to residues in the electronegative region of the binding cleft . Next , the negatively charged phosphate group of the ligand serves as the target of a simultaneous rotation and closure of the two globular domains of AfuA , generating a stable AfuA-ligand complex . This cinching mechanism is driven by the electrostatic attraction of the Arg/Lys-rich regions of the binding pocket ( which also includes residues such as Ser37 and Thr150 ) . This mechanism is compatible with our mutagenesis data , as all mutations to sugar-coordinating residues completely eliminated binding , suggesting initial recognition of the sugar ring is the determining step in ligand capture . Other homologues of AfuA may display altered substrate specificity and thus might provide more insight into the mechanism of sugar recognition and coordination by PBPs . We have confirmed AfuABC functions as a novel sugar-phosphate uptake mechanism . The only other known transporter in Gram-negative bacteria for sugar-phosphates is UhpABCT [4] . The primary distinction between these two systems is that AfuABC is an active transport system that relies on cytosolic hydrolysis of ATP , whereas UhpT-mediated transport is dependent on the maintenance of a phosphate/sugar-phosphate concentration gradient . Since AfuABC does not rely on a concentration gradient nor the concomitant phosphate efflux which is required by UhpT , uptake of sugar-phosphates through AfuABC may be preferable in a location where nutrients are constantly in flux ( such as the intestinal lumen and mucosa ) . AfuABC-mediated transport could potentially allow a pathogen to simultaneously conserve phosphate stores and take advantage of nutrients for which other bacteria may lack an appropriate transport mechanism [4] . It is established that sugar-phosphate uptake is important for intracellular pathogen replication [25–27] . Our study provides evidence that the same is true for extracellular pathogens and provides reference values for these ligands in the mouse intestine . While the most likely downstream fate of transported sugar-phosphates through AfuABC is to support bacterial metabolism , recent studies have identified both phosphorylated and cognate sugars as molecules that can act as signaling intermediaries in the host-bacterial interaction [16 , 28] . The close genetic proximity of AfuABC to one of these systems in both C . rodentium and EHEC O157:H7 , the FusKR system , implies that the substrates targeted by AfuABC could also be involved in other roles besides pure carbon and phosphate metabolism . The identification of AfuABC homologues in bacteria that colonize a variety of niches , such as the gastrointestinal and respiratory tracts , suggests uptake of sugar-phosphates plays a role in their competitive establishment on these surfaces . Since the microbiota is known to modulate the nutritional environment of these areas , we propose that AfuABC could aid in pathogen competition with host-associated commensal microbes during colonization . Notably , the clear dependence of successful pathogen transmission on the presence of AfuABC supports this concept since our assays were performed at early time points , when C . rodentium would need to overcome colonization resistance by the resident commensal microbes in order to successfully infect the host [29] . The major contributing source of sugar-phosphates to the gut metabolome remains unclear . It is likely that a variety of host and microbial factors contribute to the pool of these substrates ( e . g . from epithelial cell turnover and bacterial lysis ) . One potential source of sugar phosphates is the mucin family of glycosylated cell surface proteins . The contributions of mucin-derived substrates , in particular fucose and sialic acid , to bacterial infections have recently been recognized in several studies [16 , 30–32] . Intriguingly , when EHEC is grown in the presence of mucus , the OI-20 pathogenicity island is upregulated , indicating a potential connection between AfuABC and the presence of mucin [33] . Mice lacking the predominant intestinal mucin Muc2 mice are hypersusceptible to C . rodentium infection due to decreased mucin-dependent bacterial clearance and it will be interesting to determine if this and other mucins are involved in generation of free sugar-phosphates in the gut [34] . Further characterization of the source and flux of concentration of sugar-phosphates during dysbiosis will help to illuminate the potential role of the microbiota in generating substrates for AfuABC and the contribution of sugar-phosphate transport within mammalian hosts .
All bacteria were grown at 37°C on Luria-Bertani ( LB ) agar plates or in LB broth with shaking . A list of strains and plasmids used in this study is included in S7 Table . Where appropriate , antibiotics were supplemented at the following concentrations: kanamycin – 50μg/mL , ampicillin – 50μg/mL , streptomycin—100μg/mL . AfuA was initially cloned into a pET26b expression vector from A . pleuropneumoniae genomic DNA . For the growth assays , complementation vectors with low ( PSC101 ) copy numbers were used . afuA or afuABC were amplified from A . pleuropneumoniae genomic DNA and cloned into each backbone by restriction free or exponential megaprimer cloning , respectively [35 , 36] . Plasmid insert sequences were confirmed by forward and reverse capillary sequencing ( The Centre for Applied Genomics , Toronto , Canada ) . E . coli transformed with an AfuA expression plasmid were grown overnight in LB-kanamycin and used to initiate liter-scale volumes of the same media . Population density was estimated by OD600 and IPTG was added to a concentration of 1mM when the OD600 reached 0 . 6 . Cells were induced at 20°C for 16 hours , after which they were spun down at 4000rpm at 4°C . Pellets were resuspended in a 20% sucrose solution to initiate osmotic shock and spun down again at 4000rpm . Sucrose pellets were resuspended in a 5mM MgCl2 lysis solution with added protease inhibitors ( Roche Applied Sciences , Canada ) and allowed to incubate on ice for at least 20 minutes . The fluid was then centrifuged at 30 , 000rpm for 40 minutes at 4°C in an ultracentrifuge ( Beckman Coulter , USA ) . The supernatant was passed through a 0 . 45μM syringe filter and batch bound for >1 hour to Ni-NTA resin ( Pierce , USA ) at 4°C with shaking . The solution was then passed through a gravity column and washed with wash buffer ( 20mM imidazole , 50mM Tris pH 8 and 300mM NaCl ) . A 6M solution of guanidine hydrochloride was used to denature bound protein after which a 3M Gd-HCl solution and finally pure wash buffer was applied to allow bound protein to refold . Protein was eluted with elution buffer ( wash buffer with 400mM imidazole ) and purity assessed with denaturing SDS-PAGE followed by Coomassie blue staining . Fractions with detectable AfuA were pooled and dialyzed with added thrombin ( Sigma-Aldrich , USA ) in wash buffer to simultaneously remove excess imidazole and cleave the N-terminal 6xHis tag from AfuA . The solution was spun down and concentrated in a 10kDa cutoff centrifugal filter ( Millipore , USA ) . Samples were further purified on a Superdex S75 10/300 gel filtration column ( GE Healthcare , USA ) pre-equilibrated with a buffer consisting of 20mM Tris ( pH 8 ) and 100mM NaCl . Fraction purity was assessed by SDS-PAGE with Coomassie Blue staining . Since we initially investigated AfuA in the context of iron capture , initial protein purifications in this study leading to the structure of G6P-AfuA were performed in the absence of guanidine . All other purifications used for ITC and crystallography were denaturing preparations . To generate single-residue mutants of AfuA , site-directed mutagenesis was performed on the original pET26b-APAfuA vector . A short ( 17 cycle ) PCR reaction was used to generate the desired mutation in the template plasmid . PCR products were digested with DpnI at 37°C for 4–6 hours to cleave undesired template plasmid and were then used to transform competent E . coli ( NEB Turbo ) for selection . Mutations were confirmed by forward and reverse capillary sequencing ( The Centre for Applied Genomics , Toronto , Canada ) . ITC was performed using a VP-ITC instrument ( Microcal , GE Healthcare ) . Runs consisted of ligand loaded into the injection syringe titrated against purified protein loaded into the sample cell . All solutions were pre-dialyzed in buffer for at least 12 hours to mitigate effects of dilution . The buffer used was the same as the gel filtration buffer described above and power changes were measured relative to a reference cell filled with distilled water . Runs consisted of a 10:1 ligand:protein ratio , with efforts made to adhere to a 200μM:20μM concentration standard . Data used to calculate binding constants were referenced against runs performed with ligand alone to control for the heat of ligand solvation . For crystallization trials , AfuA was purified as described above and concentrated to at least 10mg/mL . A common crystallization condition consisting of 0 . 2M MgCl2 , 0 . 1M Tris pH 6 . 5 and 25% PEG3350 was found for AfuA bound to either G6P , F6P or S7P . Apo-AfuA crystallized best in 0 . 2M MgCl2 , 0 . 1M MES:NaOH pH 6 . 0 and 27 . 5% PEG 3350 . Candidate crystals were looped in a cryoprotectant solution consisting of the mother liquor with 20% glycerol and flash frozen in liquid N2 . The sulfur anomalous diffraction data for AfuA bound to G6P was collected at the Structural Genomics Consortium ( SGC ) ( Toronto , Canada ) using a chromium rotating anode X-ray source ( λ= 2 . 29 Å ) and R-AXIS IV++ detector ( Rigaku ) . To improve the signal-to-noise ratio of the anomalous signal of sulfur , the crystal was scanned over 326° and a dataset was collected with an average I/σI of 59 with 11 fold redundancy to 2 . 3 Å resolution . Data was collected for AfuA-G6P , AfuA-F6P and apo-AfuA at beamline 08ID-1 ( CMCF-ID ) at the Canadian Light Source ( CLS ) ( Saskatchewan , Canada ) . For AfuA-S7P , data was collected at NECAT beamline 24-ID-E at the Advanced Photon Source ( APS ) ( Chicago , USA ) . Sulfur single wavelength anomalous dispersion ( SAD ) phasing was used to phase the initial G6P-bound structure of AfuA . This structure was then used as a molecular replacement ( MR ) search model for the solutions of AfuA-F6P and AfuA-S7P . For apo-AfuA , the AfuA-G6P structure was split into two domains ( correlating to the two lobes of the protein ) , each of which was used as an independent search model for MR . All MR and automated refinement was performed with the PHENIX suite [37] . Manual refinement was undertaken with Coot [38] . Chimera and PyMol were used for molecular modeling , graphics ( PyMol ) and electrostatic surface mapping ( Chimera ) [39] . Atomic coordinate files were deposited in the Protein Data Bank under the accession numbers 4R72 ( Apo ) , 4R73 ( G6P ) , 4R74 ( F6P ) and 4R75 ( S7P ) . For growth assays , both the mutant ( ΔuhpT ) and the WT strain ( BW25113 ) were obtained from the Keio collection of single gene knockouts [40] . Overnight cultures were grown in 5mL of M9 minimal media + 10mM glucose + antibiotics . Cells were washed twice in 1mL M9 salts ( no carbon ) and brought to a density of 109 cells/mL . For solid agar growth assays , a dilution series of 5μL spots from 107 to 103 cells/mL was used . Plates were incubated at 37°C for 30 hours and were imaged with a GelDoc XR system ( BioRad ) . For liquid growth assays , cells were plated at dilutions ranging from 108 to 104 cells/mL in wells containing 100μL of the appropriate minimal media . Each well was duplicated to account for pipetting errors . Plates were incubated at 37°C with shaking in a Bioscreen C microplate reader ( Growth Curves USA ) . Growth was measured by OD600 readings every 15 minutes for at least 24 hours . For data analysis , cell density of duplicate wells was averaged . All assays were performed in at least triplicate . A . pleuropneumoniae AfuA was used as a template for a BlastP search . The results were filtered by removing homologs that did not possess the substrate specificity-determining residues . To reduce the tree size further , AfuA hits from multiple strains of the same organism were removed , keeping the top hit with highest sequence similarity . The selected hits used for further analysis came from 268 different organisms . The multiple sequence alignments and phylogenetic tree ( neighbor-joining ) construction was performed with Geneious R7 ( Biomatters ) . MUSCLE [41] was used to generate the multiple sequence alignment . The tree was re-sampled 100 times using the bootstrap module and braches with less than 60% confidence were trimmed . The tree was confirmed by using the MrBayes [42] module . Strains from an overnight culture were diluted 1/100 into fresh LB , and 200 μl transferred into a sterile 96-well plate in triplicate ( Costar ) . The optical density at 600 nm was measured every 15 min after 5s of orbital shaking using a Trilux Scintillation Counter ( Wallec ) . DBS100 in-frame deletion mutants were generated using the suicide vector pRE118 via sacB-based allelic exchange [43] . A 1 . 5kb fragment upstream of afuA was amplified by primers ( GTGGTACCTGCGCGAGCGCGTCAGCGCG and CCGCTAGCCGCCGCCAGCGCTACGGCAGAG ) with flanking KpnI and NheI sites , while a 1 . 5kb fragment downstream of afuA was amplified by primers ( CCGAGCTCTGCTGCCGTAGAGCAGGGCG and CCGCTAGCTACGGCTCAACGGAGGTGC ) with flanking SacI and NheI sites ( italicised text indicates restriction sites ) . Upstream and downstream fragments were digested alongside pRE118 with appropriate restriction enzyme and cloned into E . coli SY327 λpir to generate deletion vector pΔafuA . DBS100 was electroporated with pΔafuA and grown on kanamycin . Resulting colonies were inoculated on sucrose plates and a third set of primers ( GTCGCTGGTTATTGAACGC and GGATGCAGAGCGAGTGTCTG ) was used to confirm deletion of the gene for sucrose resistant , kanamycin sensitive colonies . C57BL/6 mice ( 8–12 weeks old ) were bred under specific pathogen-free conditions at the Child and Family Research Institute . C . rodentium WT ( DBS100 ) and ΔafuA strains were grown shaking overnight at 37°C in 3 ml of LB . Cell density was measured by OD600 readings . Mice were orally gavaged with 100 μl of either WT , ΔafuA or a 1:1 mix of WT to ΔafuA ( total 2 . 5 × 108 cfu ) . Stool was collected at day 6 pi , homogenized , serial diluted and plated on LB agar supplemented with 100 μg/ml of streptomycin . Animals were euthanized at 10 dpi and colonic tissues were removed , homogenized and plated . For CI , single colonies were picked and used as templates for colony PCR with deletion screening primers from ΔafuA generation . CI was calculated as the ratio of ΔafuA to WT colonies divided by the ratio of ΔafuA to WT from the input . All mouse experiments were performed in accordance with protocols approved by the University of British Columbia’s Animal Care Committee and in direct accordance with the Canadian Council on Animal Care’s guidelines . Assessment of the ability of C . rodentium to transmit between hosts was adapted from the protocol used by Wickham et al [44] . Index C57BL/6 mice were infected with WT or ΔafuA C . rodentium as outlined above . At 6 dpi , the index mouse plus two naïve mice ( referred to as “co-housed” ) were added to a new cage . Naïve mice remained co-housed with the index mouse for 48 h , at which time they were euthanized . Their colon tissues and luminal contents were then homogenized , serial diluted and plated on LB agar supplemented with 100 μg/ml of streptomycin for enumeration of C . rodentium burdens . As previously described [45 , 46] , C . rodentium WT , ΔafuA and ΔescN were streaked onto LB-streptomycin plates for single colony isolation . 5 ml LB-streptomycin was inoculated with a single colony of above mentioned strains and grown shaking O/N at 37°C . The strains were then subcultured at a 1:50 dilution into Dulbecco's modified Eagle's medium ( DMEM ) . The bacteria were incubated statically at 37°C , 5% CO2 until the optical density of the cultures reached 0 . 7 ( OD600 ~0 . 7 ) . Bacteria were pelleted ( 13200 rpm , 4°C , 10 minutes ) and proteins in the supernatant were precipitated using 10% trichloroacetic acid ( TCA , Sigma ) O/N at 4°C . Precipitated proteins were pelleted by centrifugation ( 13200 rpm , 4°C , 10 minutes ) and resuspended in Laemmli buffer . Samples were resolved on 12% SDS-polyacrylamide gels and visualized by Coomassie R-250 Blue staining . HeLa cells were seeded at a concentration of 5x104 cells per well and incubated for 24 h prior to infection . To initialize bacterial infection , tissue cultures were pre-incubated in 1 ml of DMEM ( Life Technologies ) supplemented with 2% FBS for 30 min . Cells were then infected with an overnight culture of C . rodentium DBS100 at a MOI of 1:100 for 8 hours . Monolayers were washed three times with Dulbecco's PBS ( Life Technologies ) to remove nonadherent bacteria , and treated with 200μl of 0 . 1% Triton X-100 PBS for 5 min at room temperature . Samples were serially diluted and incubated on LB-streptomycin plates overnight at 37°C . Values are the mean CFU of three independent experiments repeated in triplicate , and statistical comparisons between groups were performed with two-tailed Student’s t tests . Sterile round coverslips were also seeded and infected as above . After infection , coverslips were washed three times with Dulbecco's PBS and fixed in 4% paraformaldehyde ( Fisher Scientific ) overnight . The fixed cells were washed three times in PBS and permeabilized by incubation in 0 . 1% Triton X-100 in PBS for 10 min . Cover slips were incubated in anti-Tir rat polyclonal IgG antibody diluted 1:2000 in 1% BSA in TBST for 1 hour . The coverslips were then washed three times in TBST and incubated in Alexa Fluor 568 goat anti-rat IgG ( diluted 1:1000 ) and phalloidin-Alexa Fluor 488 conjugate ( Pierce ) in 1% BSA in TBST for 1 hour . After three washes in PBS , cells were mounted with ProLong Gold Mountant with DAPI ( Life Technologies ) . Images were acquired on a Zeiss AxioImager Z1 with an AxioCam HRm camera operating through AxioVision software . Immunofluorescence staining of infected tissues was performed using previously described procedures [47] . In brief , colon tissues were fixed in 10% neutral buffered formalin ( Fisher ) for 16 hrs , rinsed in 70% ethanol and paraffin embedded . Sectioning was completed by the histology laboratory at the Child and Family Research Institute . Serial 5 μm sections were cut and deparaffinized by heating at 60°C , xylene treatment and rehydration through an ethanol gradient to water . Sections were treated with PBS containing 0 . 1% Triton X-100 , followed by blocking buffer ( PBS containing 0 . 05% Tween 20 and 1% normal donkey serum ) . Tissues were incubated with goat anti-cytokeratin 19 ( 1:300 , Santa Cruz Biotechnology ) and rabbit anti-C . rodentium Tir ( 1:5 , 000; gift from W . Deng ) and subsequently probed with Alexa Fluor 568-conjugated donkey anti-goat IgG ( 1:1000; Life Technologies ) and Alexa Fluor 488-conjugate donkey anti-rabbit IgG ( 1:1000; Life Technologies ) . Tissues were mounted and imaged in the same manner as the in vitro pedestal coverslips . C57BL/6 mice were infected with C . rodentium WT or ΔafuA as described above for single-strain infections . At 6 dpi , shed stool pellets were collected and mice sacrificed . From each individual , colons were extracted , cut longitudinally , and luminal contents removed . Mucus was collected via gentle scraping along the apical colonic surface with a glass slide . Samples were immediately resuspended in 25mM HEPES pH 7 . 0 at a ratio of 1μL:1mg tissue , gently shaken to dissolve soluble extracellular species and stored on ice . Suspensions were centrifuged and equal volumes of supernatant from each sample were subjected to a conventional MeOH/CHCl3 extraction . Aqueous layers were kept and dried in a vacuum centrifuge and stored at -80°C for further analysis . Samples were analyzed with targeted tandem liquid chromatography/mass spectrometry ( LC-MS/MS ) at the Analytical Facility for Bioactive Molecules ( Hospital for Sick Children , Canada ) using an optimized protocol for LC separation of sugar-phosphates . Samples were separated on a Synergi Hydro-RP column ( Phenomenex ) and quantified with an API 4000 triple quadrupole mass spectrometer ( Sciex ) . Quantification was performed against a standard curve subjected to the extraction process . The mean values ± standard errors of the means ( SEM ) for at least two independent experiments is shown in all figures unless stated otherwise . p values were calculated using a Wilcoxon Rank Sum Test ( CI ) or a nonparametric Mann-Whitney t-test using GraphPad Prism software ( v6 . 05 ) . A p value less than 0 . 05 was considered statistically significant . | Essentially all Gram-negative pathogens are reliant on specific transport machineries termed binding protein-dependent transporters ( BPDTs ) to transport solutes such as amino acids , sugars and metal ions across their membranes . In this study we investigated AfuABC , a predicted iron-transporting BPDT found in many bacterial pathogens . We show by structural and functional approaches that AfuABC is not an iron transporter . Instead , AfuABC is a trio of proteins that bind and transport sugar-phosphates such as glucose-6-phosphate ( G6P ) . In doing so , we present the first structural solution of a G6P-specific transport protein and add to the few known unique machineries for sugar-phosphate uptake by bacteria . Furthermore , we show that AfuABC is required by the intestinal pathogen C . rodentium to effectively transmit between mice and re-establish infection , leading us to propose that the transport of sugar-phosphates is an important part of general bacterial pathogenesis . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | Active Transport of Phosphorylated Carbohydrates Promotes Intestinal Colonization and Transmission of a Bacterial Pathogen |
Transforming growth factor ( TGF ) -β inhibits hepatitis B virus ( HBV ) replication although the intracellular effectors involved are not determined . Here , we report that reduction of HBV transcripts by TGF-β is dependent on AID expression , which significantly decreases both HBV transcripts and viral DNA , resulting in inhibition of viral replication . Immunoprecipitation reveals that AID physically associates with viral P protein that binds to specific virus RNA sequence called epsilon . AID also binds to an RNA degradation complex ( RNA exosome proteins ) , indicating that AID , RNA exosome , and P protein form an RNP complex . Suppression of HBV transcripts by TGF-β was abrogated by depletion of either AID or RNA exosome components , suggesting that AID and the RNA exosome involve in TGF-β mediated suppression of HBV RNA . Moreover , AID-mediated HBV reduction does not occur when P protein is disrupted or when viral transcription is inhibited . These results suggest that induced expression of AID by TGF-β causes recruitment of the RNA exosome to viral RNP complex and the RNA exosome degrades HBV RNA in a transcription-coupled manner .
Hepatitis B virus ( HBV ) is recognized as the major causative factor of severe liver diseases such as cirrhosis and hepatocellular carcinoma . The clinical outcomes and development of hepatocellular carcinoma and cirrhosis are modulated by viral replication and antiviral immunity against HBV [1] . After entry into the host hepatocyte , HBV forms covalently closed circular DNA ( cccDNA ) in the nucleus and it initiates the transcription of viral RNAs , including a replicative intermediate known as pregenomic ( pg ) RNA . Two viral proteins ( core and P protein ) encapsidate pgRNA to form nucleocapsids , where P protein reverse-transcribes pgRNA to produce relaxed circular ( RC ) -DNA . These nucleocapsids associate with three types of viral surface proteins for secretion as infectious virions [1 , 2] . Although the mechanism of HBV replication has been well studied , the mechanisms of antiviral immunity against HBV remain unclear . Several members of the apolipoprotein B mRNA editing enzyme catalytic polypeptide ( APOBEC ) family were recently identified as new types of antiviral factors [3–5] . In humans , the APOBEC family comprises at least 11 members , including activation-induced cytidine deaminase ( AID ) , APOBEC 1 , 2 , 3A , 3B , 3C , 3D , 3F , 3G , 3H , and 4 . Most family members deaminate cytidine bases on DNA and/or RNA to generate uridine [3–5] . Accumulating evidence from in vitro experiments has further revealed that A3 proteins can inhibit the replication of various types of viruses , including human immunodeficiency virus type 1 ( HIV-1 ) and HBV [4 , 5] . Among APOBEC deaminases , the molecular mechanism of A3G antiviral activity has been well characterized . In cases of HBV , A3G restricts viral replication through hypermutation and inhibition of reverse-transcription [4 , 5] . AID is another member of the APOBEC family [4 , 5] and was originally isolated as a cytidine deaminase that triggered class switch recombination ( CSR ) and somatic hypermutation ( SHM ) of transcribed immunoglobulin genes in B cells [6–9] . AID expression was recently shown to be upregulated in human hepatocytes in vitro after stimulation with cytokines , including TGF-β1 , TNFα , and IL-1β and in the liver in chronic hepatitis patients , and AID involvement in viral infection was suggested [10–17] . Higher serum TGF-β1 levels were reported in some HBV infections in vivo [18 , 19] , and TGF-β1 reduces HBV replication in vitro [18 , 20] . However , the precise mechanisms remain elusive . In the present study , we examined the involvement of AID in TGF-β1-mediated restriction of HBV replication . We have demonstrated that TGF-β1 induces AID expression in hepatocytes , which leads to the downregulation of HBV transcripts and inhibition of nucleocapsid formation . AID-dependent downregulation of HBV transcripts requires a viral RNA binding protein ( P protein ) and RNA exosome components . These data suggest a novel antiviral pathway in which AID recruits the RNA exosome to downregulate viral RNA in HBV infected hepatocytes .
To investigate the involvement of APOBEC deaminases in TGF-β1-mediated antiviral activity against HBV , human hepatocytes ( Huh7 ) were transfected with a HBV replicon plasmid ( pPB ) [21] and the cells were then treated with TGF-β1 . Concentrations of 5–20 ng/mL TGF-β1 were used to match the range reported in chronic HBV and hepatocellular carcinoma patients [19] . HBV replication was evaluated by measuring HBV transcript levels using quantitative reverse transcription-polymerase chain reaction ( qRT-PCR ) ( Fig 1A ) and Northern blotting ( Fig 1D ) . Viral DNA in secreted virions was determined using qPCR ( Fig 1B ) , and nucleocapsid formation was estimated using native agarose gel electrophoresis ( NAGE ) . Subsequently , cytoplasmic nucleocapsid core protein and nucleocapsid associated DNA ( NC-DNA ) levels were determined using western blotting and Southern blotting , respectively ( Fig 1C ) . Collectively , TGF-β1 dose-dependently inhibited the production of HBV transcripts , nucleocapsid core protein , and nucleocapsid NC-DNA in both cytoplasmic and secreted samples . In further experiments , qRT-PCR was used to determine the expression of APOBEC deaminases in the presence and absence of TGF-β1 . Initially , relative expression levels of APOBEC deaminases in non-stimulated Huh7 cells were determined . Huh7 cells expressed all APOBEC3 deaminases . A3G and A3C were highly expressed among A3 deaminases ( Fig 1E ) , whereas APOBEC1 expression was not detected in Huh7 cells . In TGF-β1-treated Huh7 cells , expression of most APOBEC deaminases , including A3A , A3B , A3C , A3F , and AID ( Fig 1F , upper and lower ) was upregulated . Western blotting also detected AID protein in TGF-β1-stimulated Huh7 cells ( Fig 1G ) . It has been demonstrated that APOBEC3 proteins suppress HBV replication in vitro [1 , 4 , 5] . HBV plasmids and APOBEC deaminase expression vectors were transfected into Huh7 cells , and nucleocaspid formation was estimated using NAGE followed by Southern and western blotting ( NAGE assay ) . The expression of A3G and A3F , but not A3A , reduced NC-DNA levels in cytoplasmic nucleocapsids but did not reduce nucleocapsid core protein levels ( Fig 2A ) . HBV virion DNA was also reduced by A3C , A3G and A3F expression , whereas total HBV transcript levels were not affected by A3C , A3G or A3F ( Fig 2B and 2C ) . It was proposed that minus-strand DNA synthesis was the primary target of A3G-mediated anti-HBV activity in hepatocytes that were transiently transfected with HBV plasmids [1 , 4 , 5] . Our results support this proposed mechanism of A3G antiviral activity . In contrast with A3 deaminases , the overexpression of AID reduced HBV transcript levels , nucleocapsid formation , and virion secretion ( Fig 2A–2C and S1A and S1B Figs ) . Nucleocapsid NC-DNA levels were also reduced in AID-expressing cells , as indicated by Southern blotting using purified nucleocapsid NC-DNA ( Fig 2D ) . Importantly , AID expression did not suppress host cell gene transcripts ( S2 Fig ) , suggesting that AID expression may specifically suppress viral RNA . In accordance with the HBV life cycle , these data suggest that AID-mediated reduction of HBV transcripts leads to the downregulation of nucleocapsid core protein and NC-DNA . To investigate the contributions of APOBEC deaminases to TGF-β1-mediated anti-HBV activity , small interfering ( si ) RNAs targeting specific deaminases were transfected with the HBV plasmid into Huh7 cells . Cells were further treated with TGF-β1 to assess the effects on TGF-β1-mediated reduction of HBV transcripts . TGF-β1 stimulation in siGFP-transfected control cells reduced HBV transcript levels by 76% compared with non-stimulated cells ( Fig 2E , top , lane 4 vs . 8 ) . Transfection of siAID , siA3A , or siA3G suppressed the corresponding endogenous genes by up to 51% , 40% , and 56% , respectively . However , the knockdown of A3A and A3G did not affect TGF-β1-mediated reduction of HBV RNA in comparison with the siGFP control . In contrast , TGF-β1-mediated downregulation of HBV RNA was significantly attenuated by the knockdown of AID ( Fig 2E , top , lane 1 vs . 4 ) . These data suggest that TGF-β1-mediated downregulation of HBV transcripts is dependent on endogenous AID expression . Partial rescue of HBV transcript levels in siAID-transfected cells also suggests the involvement of either residual AID or other unidentified effectors in TGF-β1-mediated reduction of HBV transcripts . We previously demonstrated that the induction of AID in B cells triggers class switch recombination ( CSR ) in immunoglobulin genes [7–9] , which validates B cells as a model to study AID functions . In addition , it is anticipated that peripheral blood mononuclear cells and B cells can be extrahepatic reservoirs for HBV infection [22 , 23] . Thus , we investigated whether endogenous AID expression that could trigger CSR is also sufficient to trigger a reduction in HBV transcripts . AID expression and IgA class switching can be induced in CH12F3-2 mouse B cells following co-stimulation with CD40 ligand , IL-4 , and TGF-β1 ( designated CIT ) [6 , 24] . CH12F3-2 cells transiently transfected with the HBV plasmid were divided into two groups , and were treated with ( or without ) CIT to induce IgA switching , a GFP expression vector was co-transfected to verify transfection efficiency . At three days post-transfection , HBV replication and CSR were determined ( Fig 3A–3D ) , and showed that CIT induced AID protein expression and initiated IgA class switching , as previously reported [6 , 24] . Moreover , NAGE assays and qRT-PCR revealed that HBV transcripts , nucleocapsid NC-DNA , and core protein were downregulated in CIT-stimulated cells , whereas the expression of GFP remained intact after CIT stimulation ( Fig 3B and 3C ) . These data indicate that CIT stimulation specifically inhibits HBV replication in mouse B cells . We further used siRNAs against mouse AID ( simAID-1 and -2 ) to assess the contribution of AID to the suppression of HBV products in CIT-stimulated cells . Although simAIDs knocked down endogenous AID transcripts to only 39% determined by qRT-PCR ( Fig 3E ) , western blotting revealed clear suppression of endogenous AID protein levels ( Fig 3F ) . Furthermore , flow cytometric analyses revealed that IgA class switching is attenuated by the knockdown of AID ( Fig 3G ) , and qRT-PCR revealed that HBV transcript levels are inversely correlated with AID expression and IgA switching efficiency ( Fig 3G and 3H ) . To avoid artifacts due to the transfection process , a tetracycline-dependent stable line of the HBV replicon plasmid was established in CH12F3-2 cells ( CH12-HBV; Fig 3I ) . CH12-HBV cells were treated with CIT to induce IgA switching , and HBV transcript levels were determined . Subsequent qRT-PCR analyses demonstrated significant reductions of HBV transcript levels upon IgA switching ( Fig 3J and 3K ) . These data clearly demonstrate that endogenous AID expression sufficient to trigger CSR is also sufficient to downregulate HBV transcripts . Another putative activity of AID involves the initiation of somatic hypermutation ( SHM ) in immunoglobulin variable genes [8 , 9 , 25] previously demonstrated that human BL2 B cells autonomously induce SHM , which is absent following AID gene disruption by gene targeting . Thus , we transiently transfected the HBV replicon plasmid into BL2 cells and compared HBV replication in Aicda+/+ and Aicda−/− BL2 cells . We previously demonstrated that nucleocapsid NC-DNA and core protein are suppressed in Aicda+/+ in comparison with Aicda−/− BL2 cells , although co-transfected GFP expression levels were similar in both cell types [26] . Using identical samples , we here showed that HBV transcript levels in Aicda+/+ BL2 cells were almost 50% of those in Aicda−/− BL2 cells ( Fig 3L ) . Both mouse and human B cell lines collectively demonstrated that endogenous AID activity that can initiate either CSR or SHM of immunoglobulin genes is sufficient to trigger downregulation of HBV transcripts . To investigate the mechanism of AID-mediated downregulation of HBV transcripts , we initially focus on the viral P protein , because AID , P protein and HBV transcripts form RNP complex [26] . In these experiments , we applied a mutant HBV replicon plasmid ( pPB-ΔP , Fig 4A ) that expresses a mutant P protein lacking the C-terminal half including catalytic DNA polymerase and RNase H domains [26] . Transfection with pPB-ΔP did not support nucleocapsid DNA synthesis due to inhibition of reverse-transcription , although HBV transcription and core protein synthesis remained intact in Huh7 cells ( Fig 4C , lanes 1 and 4 ) . AID-mediated downregulation of HBV transcripts was compared between pPB- and pPB-ΔP-transfected Huh7 cells . As shown in Fig 4C , AID-mediated downregulation of HBV transcripts was not observed in pPB-ΔP-transfected Huh7 cells , indicating that AID-mediated downregulation of HBV transcripts requires intact viral P protein . The requirement of cytidine deaminase activity for AID was also investigated . AID mutant P19 was isolated from a class switch deficient patient and the deaminase activity was negligible owing to a missense mutation in catalytic cytidine deaminase domain [27] . P19 was then co-transfected with the wild-type HBV plasmid , and HBV transcript levels were compared with that in wild-type AID controls . These experiments showed that the P19 mutant significantly reduced HBV transcript level , although less effectively than wild-type AID ( Fig 4C ) . Therefore , under experimental conditions of AID over-expression , cytidine deaminase activity is not exclusively required for AID-mediated downregulation of HBV transcripts . In subsequent experiments , we generated an expression vector ( pFLAG-PΔC ) for the mutant P protein which was a corresponding mutant P protein produced from pPB-ΔP-transfected cells ( Fig 4B ) . Then the physical association between AID and the mutant P protein was examined . Immunoprecipitation analyses showed that wild type P protein co-precipitated AID in an RNase A-sensitive manner ( Fig 4D , lane 5 , 8 , 9 ) , whereas the mutant P protein ( FLAG-PΔC ) precipitated only trace levels of AID protein , suggesting that AID may not efficiently form RNP complex with the mutant P protein in pPB-ΔP-transfected cells . To explore which subcellular sites are responsible for AID and P protein interaction , cells were biochemically fractionated into three fractions ( cytoplasmic , soluble nuclear , and insoluble nuclear ) ( S3 Fig ) . Immunoprecipitation analyses using cytoplasmic and soluble nuclear proteins revealed that AID can associate with P protein in both nucleus and cytoplasm . It is of note that robust signals of AID and P proteins were found in the insoluble fraction that contains chromatin and other nuclear proteins . AID was recently shown to physically interact with RNA exosome proteins and promote CSR in transcribed immunoglobulin genes [28 , 29] . The RNA exosome comprises a ring-like structure and two catalytic components , and plays a major role in various RNA processing and degradation pathways [30 , 31] . Exosome component 3 ( Exosc3 , also known as Rrp40 ) is non-catalytic but is essential for the degradation and processing of target RNA , and the knockdown of Exosc3 severely diminished the RNA exosome function [32] . Thus , we investigated whether Exosc3 is involved in TGF-β1-mediated downregulation of HBV transcripts in Huh7 cells . As shown in Fig 5A , immunoprecipitation of AID co-purified Exosc3 , but did not precipitate GAPDH or GFP . Exosc3 immunoprecipitation also co-purified AID but not GAPDH or GFP ( Fig 5B ) , indicating a physical association between AID and Exosc3 proteins . This study found a physical association between AID and the RNA exosome proteins ( Exosc 2 , 3 , 7 ) in Huh7 cells in the absence of HBV replication ( Fig 5D ) . As expected , Exosc3 immunoprecipition also copurified with other RNA exosome proteins ( Exosc2 and 7 ) in Huh7 cells ( Fig 5D ) . Furthermore , we found that AID can also associate with RNA exosome in both nucleus and cytoplasm ( S4A Fig ) . Consistent with AID-RNA exosome interaction , RNA exosome proteins localized to both cytoplasm and nucleus ( S5A and S5B Fig ) . We previously demonstrated a physical association between HBV transcripts and AID in HBV-replicating Huh7 cells [26] . In current study , we examined whether Exosc3 associates with HBV transcripts . As shown in Fig 5C , qRT-PCR analysis demonstrated enrichment of HBV but not HPRT transcripts in Exosc3 immunoprecipitates , which was observed only when AID was present ( Fig 5C , lane 1 ) . This is also true when nuclear or cytoplasmic Exosc3 was separately precipitated ( S4B Fig ) . AID-mediated downregulation of HBV transcripts was observed in both nucleus and cytoplasm , and efficiency of downregulation was comparable between nucleus , cytoplasm , and whole cell samples ( S6A and S6B Fig ) . These results suggest that AID recruits the RNA exosome proteins to HBV transcripts and AID downregulates HBV RNA in nucleus . To further confirm that the RNA exosome is involved in AID-mediated downregulation of HBV transcripts , we used the siRNA knockdown of Exosc3 , which is essential for the RNA exosome function [32] . In these experiments , siRNAs against Exosc3 were co-transfected with the HBV plasmid and AID ( or GFP ) expression vectors , and HBV replication was determined . Northern blotting , NAGE assays , and qRT-PCR analyses showed the attenuation of AID-mediated downregulation of HBV transcripts and nucleocapsid formation in siExosc3 transfectants ( Fig 5E–5G ) . In contrast , AID , GFP , and GAPDH expression were not affected by Exosc3 depletion ( Fig 5E , bottom ) . Importantly , knock down of Exosc3 did not increase HBV RNA levels in GFP transfected samples . Moreover , siExosc3 transfection attenuated TGF-β1-mediated downregulation of HBV transcripts and nucleocapsid formation in a similar manner to that observed after transfection with siAID ( Fig 6A–6F ) . In further experiments , knockdown of another RNA exosome component Exosc6 also attenuated TGF-β1-mediated downregulation of HBV transcripts and nucleocapsid formation , albeit less effectively than the knockdown of siExosc3 and AID ( Fig 6A–6F ) . Similarly , the contributions of AID and Exosc3 to TGF-β1-mediated downregulation of HBV transcripts were examined in stably HBV-transfected Huh7 cells ( 7T7-8 ) [26] . The short hairpin ( sh ) RNA expressing lentivirus was transduced into 7T7-8 cells , and two stable transfectants ( shAID and shExosc3 ) and a control transfectant ( shLuc ) were established after puromycin selection . These cells were then cultured in the presence or absence of TGF-β1 ( Fig 7A ) . Subsequent qRT-PCR and western blotting showed reduced endogenous AID and Exosc3 expression ( Fig 7B–7E ) . Comparison of HBV transcript levels between TGF-β1-treated and non-treated 7T7-8 cells revealed that TGF-β1-mediated reduction of HBV transcripts is restored by the knockdown of AID and Exosc3 ( Fig 7F ) . Taken together , these data indicate that RNA exosome proteins ( Exosc3 and Exosc6 ) and AID are required for TGF-β1-mediated downregulutation of HBV transcripts . Immunoglobulin gene diversification triggered by AID is coupled with the transcription of immunoglobulin locus [8 , 9] . Here we examined whether AID-mediated HBV RNA downregulation is also coupled with transcription using a transcription inhibitor actinomycin D ( ActD ) . Using a stable HBV transfectant ( 7T7-8 ) , we generated experimental conditions in which endogenous or ectopic AID is expressed in HBV-replicating cells . ActD was then added to evaluate whether it could downregulate HBV RNA even in ActD-treated cells . As shown in Fig 8A and 8B , no significant synergistic reduction in HBV RNA levels by ActD and AID was observed in TGF-β1-treated and AID-overexpressing cells , indicating that AID was unable to reduce HBV RNA levels in ActD-treated cells . These results suggest that AID-mediated HBV RNA downregulation depends on transcription , similar to the immunoglobulin gene diversification triggered by AID .
AID is a key molecule involved in the diversification of immunoglobulin genes [8 , 9] , and thus its role in B cells is well understood . AID expression has been also found in non-B cells [11–13] , however , its role in non-B cells remains elusive . In the present study , we assessed AID involvement in TGF-β1-dependent anti-HBV activity and demonstrated the following: ( 1 ) AID expression is upregulated in TGF-β1-stimulated hepatocytes and reduces HBV RNA levels ( Figs 1 and 2 ) ; ( 2 ) TGF-β1-mediated downregulation of HBV transcripts is inhibited by AID knockdown ( Fig 2 ) ; and ( 3 ) endogenous AID protein levels in B cells capable of inducing immunoglobulin diversification also downregulate HBV transcript levels in a transcription-coupled manner ( Figs 3 and 8 ) . These data indicate that AID is involved in a TGF-β1-mediated anti-HBV pathway . Which part of the virus life cycle that is targeted by AID-mediated downregulation of HBV transcripts ? Another APOBEC protein , A3A , which was previously proposed to hypermutate transfected plasmids in human peripheral monocytes [33] . However , AID did not change HBV transcript levels in hepatocytes transfected with the mutant HBV replicon ( pPB-ΔP ) ( Fig 4C ) . In contrast , HBV transcripts in hepatocytes transfected with the wild-type replicon ( pPB ) were specifically downregulated by following the expression of AID expression ( Figs 2 and 4 ) . Intact HBV transcript levels in AID-expressing pPB-ΔP transfectants suggest that AID-mediated reduction of HBV transcripts is not due to plasmid targeting or promoter interference by AID activity . Otherwise , targeting of HBV plasmid or promoter activity would result in reduction of HBV transcripts in both pPB- and pPB-ΔP-transfectants because those HBV plasmids share the exactly same DNA sequences except 4 base insertion within P gene in pPB-ΔP . Previous our study demonstrated that chicken AID can downregulate cccDNA of duck hepatitis virus in a uracil-DNA glycosylase ( UNG ) -dependent manner [34] , therefore , the next obvious candidate for AID target is cccDNA of HBV . We determined cccDNA levels of transfectants using the rolling circle amplification ( RCA ) assay , which specifically amplifies circular DNA , including cccDNA . As per our results , cccDNA was clearly detected in a cccDNA-producing control cell line ( HepG2 . 2 . 5 ) [10–15 , 35]; however , the HBV-replicating transfectants used in this study rarely produced cccDNA ( S7A and S7B Fig ) . Therefore , the majority of the HBV transcripts produced from HBV transfectants in the present experimental systems are derived from HBV replicon plasmids and not from cccDNA . That means that targeting of cccDNA does not explain the observed downregulation of HBV transcripts in the present experimental systems . AID over-expression was previously shown to deaminate nucleocapsid NC-DNA and encapsidated pgRNA [10 , 13 , 26] . However , because NC-DNA is reverse transcribed from HBV pgRNA , AID activity against NC-DNA fails to explain the downregulation of HBV transcripts . Reduction of HBV RNA by the catalytically dead mutant AID ( p19 ) indicates that encapsidated pgRNA editing is distinct from AID-mediated reduction of HBV RNA . Thus , we concluded that AID directly targets HBV transcripts . The viral P protein is a reverse transcriptase that binds 5′-ε RNA structure in pgRNA and encapsidates pgRNA to the nucleocapsid [1 , 2] ( see also S1 Fig ) . It is demonstrated that P protein can also bind 3′-ε RNA structure present in 2 . 4- , 2 . 1- , and 0 . 7-kb viral mRNAs [36] , indicating that P protein binds all types of HBV transcripts . AID and TGF-β downregulate HBV transcripts containing 3′-ε but not cellular transcripts ( Figs 1D , 4C , 5E and 6A , and S1 and S2 Figs ) . AID-mediated HBV RNA reduction did not occur in hepatocytes expressing a mutant P protein ( Fig 4C ) . We demonstrated that AID can physically associate with viral RNP complexes comprising P protein [26] ( Fig 4 ) . Therefore , AID-mediated HBV RNA reduction is dependent on the presence of intact P protein and P protein may determine the target specificity for AID-mediated HBV RNA reduction . Mao et al . recently reported that zinc finger antiviral protein ( ZAP ) inhibits the replication of HBV by binding the 5′-ε RNA structure of HBV and degrading viral RNA [37] . To explore the molecular mechanism of AID-mediated downregulation of HBV transcripts , we first investigated the possible involvement of ZAP . RT-qPCR revealed that AID expression did not affect the level of ZAP mRNA ( S2 Fig ) . Knocking down of ZAP by transfection of siRNAs against ZAP increased HBV RNA levels , which indicates that ZAP reduces the basal level of HBV RNA; however , AID-mediated downregulation of HBV transcripts was not affected by knocking down of ZAP expression ( S8 Fig ) . These results imply that the ZAP antiviral pathway is dispensable for AID-mediated downregulation of HBV transcripts . Next , we explored the possible involvement of the RNA exosome . Basu et al . [29] demonstrated that AID binds and recruits the RNA exosome complex to R-loop structures in immunoglobulin genes . Here , we investigated whether AID forms a complex with RNA exosome proteins in hepatocytes . The immunoprecipitation of AID and Exosc3 revealed the formation of a RNP complex comprising AID and RNA exosome proteins in both nucleus and cytoplasm of hepatocytes , and that HBV transcripts formed a specific complex with the RNA exosome in an AID-dependent manner ( Fig 5 and S4 Fig ) . Furthermore , AID-dependent downregulation of HBV transcripts was inhibited in the absence of the essential RNA exosome component Exosc3 ( Fig 5 ) . We also demonstrated that AID-mediated downregulation of HBV transcripts does not occur when P protein loses the C-terminus domain , which is essential for AID binding ( Fig 4C ) . Inhibition of transcription resulted in blocking of AID-mediated downregulation of HBV transcripts ( Fig 8 ) . Taken together , we suggest that AID recruits the RNA exosome to transcribing HBV RNA through an association with the P protein , and thereby downregulates HBV transcripts ( Fig 8C ) . AID has been shown to reduce the transpositioning of the reverse transcriptase-dependent retroelement L1 [14 , 15] . Moreover , MacDuff et al . demonstrated that a catalytically dead mutant and wild-type AID suppress L1 transpositioning . Here , we showed that the AID-mediated HBV RNA reduction depends on HBV reverse transcriptase ( P protein ) , and catalytically dead mutant AID ( p19 ) reduces HBV transcript levels ( Fig 4 ) . It would be interesting to examine whether suppression of transpositioning by AID is also dependent on the RNA exosome . To our knowledge , this is the first study to show that AID mediates the downregulation of viral RNA through the RNA exosome complex . However , further studies are required to elucidate the mechanisms of AID-mediated HBV RNA downregulation , and to investigate the involvement of AID in anti-HBV activity in vivo .
NAGE assays were performed as previously described [20 , 26 , 38 , 39] . In brief , intact nucleocapsid particles were separated from crude extracts of HBV-replicating cells using agarose gel electrophoresis . Nucleocapsid particles within the gel were then denatured under alkaline conditions , and were transferred onto nitrocellulose membranes ( Roche ) . Nucleocapsid DNA and core proteins were detected using Southern and western blotting with a double-stranded HBV DNA probe spanning the whole viral genome and an anti-core antibody , respectively . Plasmids were transfected into Huh7 cells using CalPhos ( Clontech ) or Fugene 6 ( Roche ) . The total transfected plasmid per sample was normalized by supplementation with empty or GFP expression plasmids . Co-transfection of plasmid and siRNA was performed using lipofectamine 2000 according to the manufacturer’s instructions . Stealth-grade siRNA for mouse and human AID , A3A , A3G , Exosc3 , Exosc6 , and control were purchased from Invitrogen . In all transfection experiments , control siRNA was designed to differ from all mammal transcripts . BL2 [25] and CH12F3-2 cell culture , CIT stimulation , and transfection by electroporation were performed as previously described [24–26 , 40] . The HBV-replicating Huh7 cell line ( 7T7-8 ) was established and described previously [26] . The pTre-HBV [41] vector was transfected into tetracycline activator expressing CH12F3-2 cells ( FTZ14 [42] ) to establish the CH12-HBV cell line . Subsequently , shLuc , shAID , and shExosc3 expressing 7T7-8 cells were established by infection with recombinant lentivirus followed by puromycin selection . Recombinant lentiviruses were generated by transient transfection of shLuc- , shAID- , and shExosc3-pLKO1-puro and packaging plasmids ( pMD2 . D and psPAX2 , Addgene plasmid 12259 and 12260 , respectively , kind gifts of Dr . Trono ) in 293T cells according to the manufacturer’s instructions . Human TGF-β1 and IL-4 were purchased from R&D systems . Actinomycin D was purchased from Sigma-Aldrich . The HBV replicon plasmid ( pPB ) contains 1 . 04 copies of HBV genomic DNA and expresses pgRNA under the control of the CMV promoter [21] . The pPB-ΔP plasmid contains a frame-shift mutation in codon 306 of the P gene , leading to loss of the C-terminal 539 amino acids , which comprise catalytic and RNase H domains [26] . Probe labeling and northern and Southern blots were developed using the AlkPhos direct labeling system ( Amersham ) . Signals for northern , Southern , and western blots were analyzed using a LAS1000 Imager System ( FujiFilm ) . Other expression vectors are listed in S1 Table and the gene accession numbers were listed in S3 Table . Cells were lysed in buffer containing 50-mM Tris-HCl ( pH 7 . 1 ) , 20-mM NaCl , 1% NP-40 , 1-mM EDTA , 2% glycerol , and protease inhibitor cocktail ( Roche ) . After centrifugation , supernatants were incubated with the indicated antibodies and protein G sepharose ( GE Healthcare ) or anti-FLAG M2 agarose beads ( Sigma , A2220 ) . For IP-qRT-PCR experiments , cells were lysed with PBS containing 0 . 1% Tween 20 , 1% triton-X , 1-mM EDTA , protease inhibitor cocktail ( Roche ) , and 2% glycerol . After centrifugation , crude lysates were subjected to anti-FLAG M2 beads for 4 h . Immune complexes were washed in lysis buffer 10 times and were then washed in lysis buffer containing an additional 100-mM NaCl . FLAG-Exosc3 and RNA complexes were eluted using free 3 × FLAG peptides ( Sigma , F4799 ) . Western blotting was performed using standard methods with rabbit anti-GAPDH ( Sigma , G9545 ) , mouse anti-FLAG ( Sigma , F3165 ) , rabbit anti-GFP ( Clontech , 632376 ) , anti-rabbit Igs HRP ( Biosource , ALI3404 ) , anti-rat Igs HRP ( Jackson ImmunoResearch , 712-035-153 ) , rabbit and mouse IgG TrueBlot ( eBioscience , 18–8816 , 18–8877 ) , rat monoclonal anti-AID ( MAID2 , eBioscience , 14–5959 ) , rabbit anti-A3G[38] , anti-core ( Dako , B0586 ) , anti-human Exosc3 ( GenWay Biotech , GNB-FF795C , F8130F ) , and isotype control ( eBioscience 14–4321 ) antibodies . To generate a polyclonal antibody against AID , the C-terminal AID peptide ( EVDDLRDAFRMLGF ) was conjugated with cysteine and rabbits were immunized using keyhole limpet hemocyanin ( KLH ) . Subsequently , anti-AID rabbit serum and rat monoclonal anti-AID were isolated and used in IP experiments . IgA class switching was determined by detecting surface IgA expression using flow cytometry as previously described[7 , 24 , 40] . Total RNA was extracted using TRIsure ( Bioline ) , was treated with DNase I ( Takara ) to eliminate transfected plasmids , and was then re-purified using TRIsure . For qRT-PCR analyses , 1 μg of total RNA was treated with amplification grade DNase I ( Invitrogen ) and was then reverse-transcribed using oligo-dT or random primers with SuperScript III ( Invitrogen ) . Subsequently , cDNA was amplified using SYBR green ROX ( Toyobo ) with MX3000 ( Stratagene ) according to the PCR protocol . A1 , AID , A3A , A3B , A3C , A3D , A3F , A3G , A3H , Exosc3 , Exosc6 , 18S ribosomal RNA , HPRT , and β-actin expression and HBV transcription were determined using PCR conditions of 95°C for 1 min followed by 40 cycles of 95°C for 15 s , 55°C for 30 s , and 70°C for 30 s , and one cycle of 95°C for 1 min , 55°C for 30 s , and 95°C for 30 s . For A3A amplification , an annealing temperature of 60°C was used . Copy numbers of APOBECs were determined using plasmid standard curves for each APOBEC ( Fig 2A ) . Fold induction of APOBEC expression following treatment of cells with TGF-β1 was determined using the ΔΔCT method [43] . To eliminate transfected plasmids , purified NC-DNA from secreted virions and cytoplasmic lysates was obtained after serial DNase I digestion , proteinase K and SDS digestion , phenol–chloroform extraction , and isopropanol precipitation . NC-DNA copy numbers were determined using a HBV plasmid standard curve . Transcript expression levels in this study ( except Fig 2A ) are presented as fold induction relative to unstimulated cells . In transfection experiments , expression levels of mock- , GFP- , siGFP- , and siLuc-transfected cells were defined as one . Expression levels in qRT-PCR analyses were normalized to the amplification of internal controls ( HPRT , β-actin , or 18S ribosomal RNA ) . Primers are listed in S2 Table . Differences were identified using the two-tailed unpaired Student’s t-tests and were considered significant when P < 0 . 05 . | HBV is one of the causative factors of hepatocellular carcinoma . Recent studies have shown that the members of the APOBEC deaminase family are antiviral factors that suppress the replication of viruses , such as HIV-1 and HBV . APOBEC3G suppresses viral replication by either hypermutation of nascent DNA or inhibition of reverse transcription . Recent studies have been suggested that AID , another APOBEC family member , restricts viruses and retrotransposons that use reverse transcription for their replication . However , little is known about the antiviral mechanisms of AID . TGF-β is a pleiotropic cytokine involved in the suppression of HBV replication , but the mechanism underlying its anti-HBV activity is unclear . In this study , we found that AID plays a role in the anti-HBV activity of TGF-β . Further study revealed that AID physically associates with a viral RNP complex containing reverse transcriptase and recruits the RNA degradosome ( RNA exosome ) to the RNP complex to degrade the viral RNA . To the best of our knowledge , this study is the first to reveal a novel antiviral pathway in which AID triggers viral RNA degradation by tethering the RNA exosome to the viral reverse transcriptase/RNA complex . Viral RNA may be another target for APOBEC antiviral activity . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | TGF-β Suppression of HBV RNA through AID-Dependent Recruitment of an RNA Exosome Complex |
Many Gram-negative bacterial pathogens express contact-dependent growth inhibition ( CDI ) systems that promote cell-cell interaction . CDI+ bacteria express surface CdiA effector proteins , which transfer their C-terminal toxin domains into susceptible target cells upon binding to specific receptors . CDI+ cells also produce immunity proteins that neutralize the toxin domains delivered from neighboring siblings . Here , we show that CdiAEC536 from uropathogenic Escherichia coli 536 ( EC536 ) uses OmpC and OmpF as receptors to recognize target bacteria . E . coli mutants lacking either ompF or ompC are resistant to CDIEC536-mediated growth inhibition , and both porins are required for target-cell adhesion to inhibitors that express CdiAEC536 . Experiments with single-chain OmpF fusions indicate that the CdiAEC536 receptor is heterotrimeric OmpC-OmpF . Because the OmpC and OmpF porins are under selective pressure from bacteriophages and host immune systems , their surface-exposed loops vary between E . coli isolates . OmpC polymorphism has a significant impact on CDIEC536 mediated competition , with many E . coli isolates expressing alleles that are not recognized by CdiAEC536 . Analyses of recombinant OmpC chimeras suggest that extracellular loops L4 and L5 are important recognition epitopes for CdiAEC536 . Loops L4 and L5 also account for much of the sequence variability between E . coli OmpC proteins , raising the possibility that CDI contributes to the selective pressure driving OmpC diversification . We find that the most efficient CdiAEC536 receptors are encoded by isolates that carry the same cdi gene cluster as E . coli 536 . Thus , it appears that CdiA effectors often bind preferentially to "self" receptors , thereby promoting interactions between sibling cells . As a consequence , these effector proteins cannot recognize nor suppress the growth of many potential competitors . These findings suggest that self-recognition and kin selection are important functions of CDI .
Contact-dependent growth inhibition ( CDI ) systems mediate the transfer of protein toxins between Gram-negative bacteria . CDI was first discovered and characterized in Escherichia coli EC93 , which uses CdiBEC93/CdiAEC93 two-partner secretion proteins to inhibit the growth of other E . coli isolates [1] . CdiBEC93 is an Omp85 family β-barrel protein that exports and presents CdiAEC93 on the cell surface . CdiAEC93 carries toxic effector activity and is predicted to form a β-helical filament that projects several hundred angstroms from the inhibitor cell [2 , 3] . CdiAEC93 binds to BamA on the surface of neighboring E . coli cells , then delivers its C-terminal toxin domain ( CdiA-CTEC93 ) into the target cell to inhibit growth [4] . The proton gradient is rapidly dissipated in intoxicated cells [5 , 6] , suggesting that the CdiA-CTEC93 toxin forms pores in the cytoplasmic membranes of target bacteria . Because E . coli EC93 cells also express BamA receptors , they receive toxin domains from neighboring siblings . However , inter-sibling toxin exchange does not induce growth arrest because the cdi locus encodes the CdiIEC93 immunity protein , which neutralizes CdiA-CTEC93 toxicity [1 , 7] . Thus , E . coli EC93 deploys CDI to inhibit competing bacteria , but sibling cells are immune to growth inhibition . Since their discovery in E . coli EC93 , cdi genes have been identified and characterized in several other proteobacteria [8–13] . CDI systems are typically encoded on plasmids and genomic islands and therefore are not necessarily found in all isolates of given species . This is exemplified by E . coli , for which ~25% of sequenced isolates carry cdi gene clusters . By contrast , every sequenced strain of Neisseria meningitidis and Burkholderia pseudomallei encodes at least one CDI system . CDI is also characterized by toxin diversity [3 , 14] . CdiA-CT sequences vary widely between bacteria , with toxins exhibiting several distinct activities [3 , 15] . CdiI immunity proteins are also diverse and specifically protect against cognate CdiA-CT toxins . Thus , cdi genes collectively encode a network of toxin/immunity protein pairs that appear to be rapidly diversifying . Toxin diversity and the specificity of immunity protection suggest that CDI systems are deployed to compete for environmental resources and growth niches . The molecular mechanisms of CDI toxin delivery have been explored largely using the E . coli EC93 system as a model . Early work showed that CDIEC93 expression in E . coli K-12 strains promotes auto-adhesion [1] and that CdiA-CTEC93 delivery is receptor-dependent [4] . The CdiAEC93 receptor , BamA , was identified from selections for CDI-resistant ( CDIR ) target cells . E . coli bamA101 mutants carry a transposon insertion that reduces BamA expression five-fold and confers partial resistance to CDIEC93-mediated growth inhibition [4] . BamA is an essential outer-membrane protein that forms the core of the β-barrel assembly machine ( BAM ) complex in Gram-negative bacteria [16–18] . The BAM complex inserts β-barrel proteins into the outer membrane , and BamA orthologues are found in all Gram-negative bacteria and eukaryotic plastids . Though BamA is conserved , its extracellular loops vary considerably between closely related enterobacterial species [19] . For example , E . coli BamAEco shares ~94% overall sequence identity with BamAECL from Enterobacter cloacae ATCC 13047 , but the surface-exposed portions of extracellular loops L4 , L6 and L7 are unrelated in sequence [20] . Consequently , E . coli EC93 does not recognize E . cloacae cells as targets and cannot inhibit the growth of this species . Similarly , replacement of bamAEco with bamA genes from other enterobacterial species confers CDI-resistance to E . coli cells [20] . Thus , receptor polymorphism restricts the CDIEC93 target-cell range to E . coli and Shigella isolates , which share identical BamA proteins . Recognition of "self" receptors appears to be a general feature of CDI , because CdiA effectors from Dickeya dadantii 3937 , Burkholderia thailandensis E264 , Neisseria meningitidis B16B6 and Pseudomonas aeruginosa PAO1 all deliver toxins to sibling cells [8–13] . This exchange of toxins provides no apparent competitive advantage because siblings are immune . However , the accompanying cell-cell adhesion likely imparts significant benefits by promoting auto-aggregation and biofilm formation [7 , 11 , 13 , 21–25] . Thus , in addition to suppressing the growth of competing bacteria , CDI also contributes to bacterial fitness by mediating cooperative behaviors between sibling cells . Bacterial surfaces interact directly with the environment and are therefore under strong selective pressure . Bacteriophages and adaptive immune systems exploit surface epitopes to identify bacteria for infection and elimination , respectively . In response , bacteria have evolved mechanisms to rapidly alter their repertoire of surface molecules such that different isolates of the same species commonly express unique extracellular antigens [26–29] . Therefore , CdiA effectors from different bacterial species must presumably recognize distinct receptors . Here , we begin to explore CDI receptor diversity using the CdiAEC536 effector from uropathogenic E . coli 536 ( EC536 ) . We find that E . coli ompC and ompF mutants are resistant to CDIEC536-mediated growth inhibition , suggesting that the OmpC and OmpF osmoporins function as receptors for CdiAEC536 . CDIEC536 inhibitor cells only bind to target bacteria that express OmpC and OmpF , indicating that both porins are required for receptor function . Using single-chain OmpF dimers , we provide evidence that heterotrimeric OmpC-OmpF constitutes the receptor for CdiAEC536 . Because OmpF and OmpC exhibit considerable sequence variability in their extracellular loops [30 , 31] , we examined the effects of these polymorphisms on receptor function . We find that several OmpC proteins from uropathogenic E . coli isolates are not recognized by CdiAEC536 , suggesting that many potential competitors are resistant to CDIEC536 . By contrast , OmpC proteins from E . coli UTI89 and F11 are effective receptors for CdiAEC536 . Notably , E . coli UTI89 and F11 also encode CdiA proteins that are >98% identical to CdiAEC536 , indicating that their effectors also recognize OmpC-OmpF receptors and mediate self-recognition . In these instances , CdiA effector proteins preferentially recognize "self" receptors , suggesting that auto-adhesion and kin selection are important functions of CDI .
We previously reported that E . coli bamA101 mutants are resistant to CDIEC536 [32] , though the level of protection is less than that observed against CDIEC93 inhibitors [4] . Because the bamA101 mutation confers incomplete resistance , we revisited these experiments using E . coli target strains that express E . cloacae bamA ( bamAECL ) , which provides full resistance to CDIEC93 [20] . We first confirmed that E . coli bamAECL target cells are resistant to CDIEC93 using E . coli EPI100 inhibitor cells that express the cdiBAIEC93 gene cluster from a cosmid vector . As expected , E . coli bamAECL targets grew to the same level in co-culture with either CDIEC93 inhibitors or CDI−cells that carry an empty cosmid vector ( Fig 1 ) . By contrast , bamAEco target-cell viability decreased ~600-fold during co-culture with CDIEC93 inhibitors ( Fig 1 ) . We then tested bamAEco and bamAECL cells in competitions against E . coli EPI100 that express the E . coli 536 cdiBAIEC536 genes from a cosmid and found that both target strains were inhibited to the same extent ( Fig 1 ) . These results indicate that BamAECL does not protect against CDIEC536 , suggesting that the originally observed bamA101 resistance may be due to decreased expression of an unidentified outer-membrane receptor that requires BamA for assembly . We used a genetic approach to identify the CdiAEC536 receptor , reasoning that its disruption should confer CDI resistance ( CDIR ) . We subjected CDI-sensitive E . coli target cells to transposon mutagenesis , then selected CDIR mutants from co-cultures with E . coli 536 cells that express the cdi locus under control of an arabinose-inducible promoter . Additionally , because CysK ( O-acetylserine sulfhydrylase A ) is required to activate the CdiA-CTEC536 toxin domain [33 , 34] , we provided target cells with plasmid-borne cysK to avoid isolating CDIR cysK null mutants . CDIR populations were enriched after three rounds of selection , and randomly isolated clones exhibited complete resistance to growth inhibition . Transposon insertions from CDIR clones were transduced into CDI-sensitive cells and genetic linkage between insertions and CDIR phenotypes was established for each transductant . The disrupted genes were then identified from ten randomly selected CDIR mutants , revealing five independent insertions each in ompC and ompF ( Fig 2A ) . The ompC and ompF genes encode closely related osmoporins , which are trimeric β-barrel proteins that facilitate the diffusion of small hydrophilic molecules across the outer membranes of Gram-negative bacteria [35] . To confirm the roles of OmpC and OmpF in CDIEC536 , we generated an E . coli ΔompC ΔompF double mutant and performed complementation analysis . As expected , the growth of ΔompC ΔompF cells was not inhibited in co-culture with E . coli 536 ( Fig 2B ) . By contrast , wild-type ompC+ ompF+ target cells showed a 50-fold loss in viability under the same co-culture conditions ( Fig 2B ) . This latter growth inhibition is due to CDIEC536 , because the growth of wild-type E . coli was not affected by E . coli 536 mock inhibitors that carry a deletion of the cdiA-CT coding region ( Fig 2C ) . Complementation of the ΔompC ΔompF mutant with a plasmid expressing either ompC or ompF had no effect on the resistance phenotype , but the strain became CDI-sensitive when it harbored plasmids expressing both osmoporins ( Fig 2B ) . Together , these findings demonstrate that OmpC and OmpF expression is required to render bacteria sensitive to the CDIEC536 system . CdiA-mediated adhesion between inhibitor and target cells is a hallmark of CDI . To determine whether OmpC and OmpF are required for CDI-dependent adhesion , we monitored cell-cell binding using a flow-cytometry approach . GFP-labeled inhibitor cells were mixed with DsRed-labeled target bacteria at a 5:1 ratio , and the suspensions analyzed by flow cytometry . Flow events registering both green and red fluorescence correspond to aggregates containing inhibitor and target bacteria . The flow cytometric assay for CDI-mediated adhesion generally requires CdiA over-expression [4 , 36] , therefore we used inhibitor cells that express cdi genes from cosmids for these experiments . Under these conditions , CDIEC93 inhibitor cells bind to 60–80% of target bacteria ( Fig 3A ) . This cell-cell adhesion is dependent on CDI , with only about 5% of target cells associated with CDI−mock inhibitors ( Fig 3A ) . We then tested whether CDIEC536 expression also promotes stable adhesion to target cells . As expected , we did not detect CDIEC536-dependent adhesion to ΔompC target bacteria; but surprisingly there was little increase in adhesion upon complementation with ompCK-12 from E . coli K-12 ( Fig 3A ) . Despite the lack of detectable cell-cell adhesion , ompCK-12-complemented cells were inhibited during co-culture with E . coli 536 cells ( Fig 3B ) and immunoblotting confirmed that they produced OmpCK-12 ( Fig 3C ) . OmpC porins vary significantly between E . coli isolates [30] , and the E . coli K-12 and 536 proteins carry different surface-exposed loops ( Fig 3D ) . Because extracellular loop polymorphism could affect receptor function , we tested whether ompCEC536 better supports CDIEC536-dependent cell adhesion . Strikingly , over 40% of ompCEC536 complemented bacteria adhered to CDIEC536 inhibitors , and this adhesion was specific because these target cells did not associate with CDI−mock inhibitors ( Fig 3A ) . These results suggest that CdiAEC536 binds to OmpCEC536 with higher avidity than OmpCK-12 . We also found that ompCEC536 target cells were inhibited to a greater extent than ompCK-12 targets in co-cultures with E . coli 536 ( Fig 3B ) . We next tested OmpC proteins from Salmonella Typhimurium LT2 and E . cloacae ATCC 13047 ( Fig 3D ) to determine whether other species are potential targets for the CDIEC536 system . E . coli ΔompC cells complemented with S . Typhimurium ompCLT2 did not adhere to CDIEC536 inhibitors ( Fig 3A ) and were not inhibited in competition co-culture ( Fig 3B ) , though they produced substantial OmpCLT2 protein ( Fig 3C ) . By contrast , CDIEC536 inhibitors bound to ompCECL complemented cells and inhibited their growth ( Fig 3A and 3B ) . Together , these results support a role for OmpC as the CdiAEC536 receptor and indicate that OmpC polymorphism determines whether bacteria are susceptible to CDIEC536-mediated growth inhibition . We next examined the contribution of OmpF to CDIEC536-dependent adhesion by testing whether porins from different bacteria influence receptor activity ( Fig 4A ) . Because OmpCK-12 does not support detectable adhesion to CDIEC536 inhibitor cells , we replaced the native ompCK-12 locus with the ompCEC536 allele to generate a target-cell strain for cell-cell binding studies . The resulting ΔompF ompCEC536 cells did not form aggregates with CDIEC536 inhibitors , but adhered normally to CDIEC93 inhibitor cells ( Fig 4B ) . We then expressed plasmid-borne ompF from E . coli K-12 , E . coli 536 , S . Typhimurium and E . cloacae and found that each allele supported adhesion to CDIEC536 inhibitors ( Fig 4B and 4C ) . These results demonstrate that both OmpC and OmpF are required for CDIEC536-dependent cell adhesion , indicating that each porin contributes to receptor function . At least two models can explain the roles of OmpC and OmpF in CDI . CdiAEC536 could interact simultaneously with adjacent OmpC and OmpF homotrimers on the target-cell surface . Alternatively , the effector could recognize OmpC and OmpF displayed as heterotrimers , which have been reported previously [37] . To differentiate between these models , we generated and tested the function of covalently constrained OmpC-OmpF heterotrimers . Because osmoporins are obligate trimers , we reasoned that porin dimers produced from a gene fusion should require an additional monomer to assemble properly in the outer membrane . We tested this prediction using dimeric OmpF proteins tethered by 8- and 16-residue linker peptides . OmpF assembly and function was assessed by monitoring sensitivity to colicin E5 , which requires OmpF to translocate its C-terminal toxin domain into E . coli cells [38–41] . E . coli ΔompF ΔompC mutants are completely resistant to colicin E5 , but become susceptible when complemented with plasmid-borne ompFEC536 ( Fig 5A ) . By contrast , neither of the dimeric ompF constructs restored colicin sensitivity ( Fig 5A ) , even though immunoblot analysis confirmed the presence of OmpF dimers in these cells ( Fig 5B ) . We then expressed ompF-ompF fusions in the ΔompF ompCEC536 background to test whether monomeric OmpC supports heterotrimer formation . E . coli ΔompF ompCEC536 cells are somewhat sensitive to colicin E5 ( Fig 5C ) , consistent with a prior report showing that OmpC can support colicin entry in the absence of OmpF [42] . However , colicin E5 had a greater inhibitory effect on ΔompF ompCEC536 cells that were complemented with plasmids that produce either monomeric or dimeric OmpF ( Fig 5C ) . Together , these results indicate that OmpF2-OmpC heterotrimers assemble in the outer membrane and are active in colicin translocation . Having established a strategy to generate heterotrimeric porin complexes , we tested whether OmpF2-OmpC is a receptor for CdiAEC536 . Flow cytometry showed that each OmpF dimer supported CDIEC536-dependent cell-cell adhesion to the same extent as monomeric OmpF ( Fig 6A ) . Moreover , we also found that ΔompF ompCEC536 cells expressing dimeric OmpF were inhibited 10- to 100-fold by E . coli 536 cells in competition co-cultures ( Fig 6B ) . Taken together , these results indicate that heterotrimeric OmpF-OmpC is the receptor for CdiAEC536 . We next used natural variations between E . coli OmpC porins to identify potential CdiAEC536 recognition epitopes . We surveyed predicted OmpC proteins from E . coli isolates and identified over 220 distinct sequences ( S1 Table ) . Though some E . coli OmpC variants differ by single amino-acid residue substitutions , much of the sequence variation is localized to extracellular loops L4 ( 33 sequences ) , L5 ( 30 sequences ) and L7 ( 13 sequences ) ( S2 Table ) . Moreover , these loops are found in many different combinations ( S1 Table ) . To examine the influence of OmpC polymorphism on CDIEC536 , we tested a subset of E . coli OmpC proteins with unique extracellular loop combinations ( Figs 7A and S1 ) . E . coli ΔompC target cells were provided with plasmid-borne ompC alleles and co-cultured with E . coli 536 inhibitor cells . The ompC alleles from E . coli UTI89 , F11 and A33H rendered ΔompC cells sensitive to growth inhibition to roughly the same extent as ompCEC536 ( Fig 7B ) . Targets that express ompCEC93 from E . coli EC93 were also inhibited in co-culture with E . coli 536 , but ompC alleles from uropathogenic isolates CFT073 , 6104H , A54H , A42H and A35H were all associated with resistance ( Fig 7B and 7C ) . Immunoblot analysis confirmed that OmpC was produced in all of the complemented strains ( Fig 7D ) , indicating that resistance is not due to a lack of ompC expression . Thus , E . coli ompC allele polymorphism appears to dictate cell susceptibility to CDIEC536-mediated intoxication . To determine which OmpC sequences are critical for recognition by CdiAEC536 , we generated chimeric porins using OmpCCFT073 as a framework upon which to graft heterologous OmpC loops . OmpCCFT073 and OmpCEC536 share loop L4 , but differ in their L5 , L7 and L8 sequences ( Figs 7A and S1 ) . OmpCCFT073 chimeras carrying L7 or L8 from OmpCEC536 still provided resistance to CDIEC536 ( Fig 8A ) , indicating that these loops probably have no role in binding CdiAEC536 . However , replacement of loop L5 converted OmpCCFT073 into a functional receptor that supports the same level of inhibition as OmpCEC536 ( Fig 8A ) . Similarly , OmpCCFT073 chimeras carrying OmpCEC536 L7 or L8 were functional receptors , provided that they also carried loop L5 from OmpCEC536 ( Fig 8A ) . We tested the role of loop L4 by grafting the sequence from OmpCF11 ( which is a receptor for CdiAEC536 ) onto OmpCCFT073 . Targets that express the resulting chimera were inhibited by E . coli 536 to about the same extent as targets that express OmpCF11 ( Fig 8A ) . This latter observation shows that OmpCCFT073 loop L5 is not an anti-determinant , because it supports CdiAEC536 recognition in other contexts . Collectively , these results indicate that OmpC loops L4 and L5 contribute to target-cell selection , suggesting that the loops may be recognized directly by CdiAEC536 .
The results presented here establish a critical role for OmpC and OmpF in the CDIEC536 pathway . CDIEC536 inhibitor cells only adhere to target bacteria that express OmpC and OmpF , indicating that both porins contribute to receptor function . Though osmoporins are generally assumed to form homotrimers , we show that heterotrimeric OmpC-OmpF constitutes the CdiAEC536 receptor . Heterotrimeric porins were first reported by Gehring and Nikaido , who found that OmpC , OmpF and PhoE can form mixed trimers in E . coli K-12 [37] . OmpCK-12 and OmpFK-12 share 58 . 5% sequence identity ( S2 Fig ) , and their β-barrels superimpose with root-square mean deviation of 0 . 78 Å [43] . The trimer interaction surface is dominated by hydrophobic contacts between β-barrel strands β1 , β2 , β3 , β4 , β5 and β1´ , and these regions are highly conserved between OmpCK-12 and OmpFK-12 ( S2 and S3 Figs , S3 Table ) . The porins also share an inter-protomer ion-pair between a conserved Glu residue ( Glu66 in OmpCK-12 and Glu71 in OmpFK-12 ) at the tip of loop L2 and an Arg residue within L3 of the adjacent β-barrel ( S3 Fig and S3 Table ) . Thus , the structures of OmpF and OmpC are compatible with heterotrimer formation . The targeting of heterotrimeric OmpF-OmpC receptors has a number of implications . Firstly , residues from both porins should be directly recognized by CdiAEC536 . The results presented here suggest that OmpC loops L4 and L5 are important CdiAEC536-binding determinants , but the contribution of OmpF is less clear . OmpF porins from E . cloacae and S . Typhimurium are effective co-receptors in conjunction with OmpCEC536 , so the CdiA-binding epitope is presumably shared by all of the tested OmpF variants . Additionally , because OmpC and OmpF are regulated reciprocally , it is possible that receptor availability fluctuates in response to environmental conditions . OmpF is expressed preferentially in response to low osmolarity , whereas OmpC dominates under hyperosmotic conditions [35] . We find that target cells are effectively inhibited when they over-express ompC alleles , suggesting that OmpF-OmpC2 heterotrimers are recognized by CdiAEC536 . We have also shown that covalently constrained OmpF2-OmpC heterotrimers are functional receptors . Because both heterotrimeric forms support CDIEC536 , we conclude that CdiAEC536 effectors probably make contact with one OmpC and one OmpF protomer in the receptor complex . The mechanism ( s ) by which CDI toxins are translocated across the target-cell outer membrane is unknown , but in principle , receptors could mediate this transport . Group A colicins appear to exploit the central pore of OmpF to cross the outer membrane , and this translocation is powered by the proton-motive force via the Tol protein complex [40 , 41 , 44 , 45] . However , CDI toxin translocation is not Tol-dependent , nor does it require the Ton-Exb system , which energizes the import of group B colicins [6 , 44] . Further , CDI toxin domains are modular and can be exchanged between CdiA proteins for delivery through different receptor pathways [8 , 10 , 46] . Therefore , if CdiA receptors mediate outer-membrane transport , then BamA must also be capable of CDI toxin translocation . Like OmpF and OmpC , BamA is a 16-strand β-barrel protein , but its lumen is not exposed to the extracellular milieu . Instead , extracellular loops L3 , L4 , L6 , L7 and L8 form a dome that shields the BamA lumen from the environment [47 , 48] . It is possible that the BamA barrel opens upon binding to CdiAEC93 , consistent with work showing that extracellular loop L6 is dynamic during the OMP assembly duty cycle [49 , 50] . Buchanan and colleagues have also suggested that the BamA β-barrel opens a lateral gate during OMP biogenesis [51] . These conformational changes could be exploited to translocate toxin across the outer membrane , but we note that inactive BamA lacking the POTRA-3 ( polypeptide-transport associated ) domain is able to complement the bamA101 mutation and restore sensitivity to CDIEC93 [4] . Therefore , OMP biogenesis activity is not required for CDI , and thus it remains unclear whether BamA dynamics contribute to toxin translocation . Putative CdiA-binding epitopes have been localized on BamA and OmpC , but the corresponding receptor-binding regions within CdiAEC93 and CdiAEC536 have not yet been identified . Given that BamA and OmpF/OmpC have distinct structures , the two CdiA effectors should contain unique receptor-binding domains . Alignment of CdiAEC93 and CdiAEC536 shows several regions of sequence divergence . Prominent among these are the C-terminal toxin region and the sequence corresponding to residues Ser1379 –Tyr1636 of CdiAEC93 ( S4 Fig ) . Additionally , the CdiAEC536 effector contains a 69-residue insertion ( Ser1988 –Ser2056 ) relative to CdiAEC93 ( S4 Fig ) . We recently reported that truncated CdiAEC93 proteins lacking up to ~1 , 000 C-terminal residues retain BamA-binding function [7] . Furthermore , CdiAEC93 lacking internal residues Ala722 –Gly1624 is still secreted and assembled onto the cell surface , but does not bind to BamA [7] . Together , these observations suggest that the receptor-binding domain lies between residues Ser1300 –Asn1600 of CdiAEC93 . There are four distinct classes of E . coli CdiA proteins based on sequence comparisons of the 1300–1600 region , suggesting that there may be additional CDI receptors in E . coli . Alternatively , the other uncharacterized E . coli CdiA effectors could bind to BamA or OmpC/OmpF using epitopes that are distinct from those recognized by CdiAEC93 and CdiAEC536 ( respectively ) . Intriguingly , the putative receptor-binding domain is located in the middle of the CdiA primary sequence , between the FHA-1 and FHA-2 peptide repeat regions ( S4 Fig ) . This position is unexpected based on current models of CdiA structure and cell-surface presentation . According to the β-helix structural model [2] , the receptor-binding region would be positioned near the center of the CdiA filament rather than the distal tip . It is unclear how a central location is compatible with receptor binding , because the distal end of filament would presumably have to perforate the target-cell outer membrane . Penetration of target bacteria could facilitate toxin delivery , but this model raises the question of why specific receptors should be required for CDI . Bacterial surface antigens are under intense pressure to diversify , yet E . coli BamA and OmpC proteins show remarkably different responses to this selective pressure . BamA is invariant in over 2 , 500 E . coli isolates for which sequence information is available . By contrast , E . coli strains encode at least 220 distinct OmpC proteins , many of which are distinguished by unique combinations of extracellular loops L4 , L5 and L7 . Crystal structures are available for the OmpCK-12 , OmpCCFT073 and OmpC6104H porins studied here , and each exhibits a distinct surface topology and electrostatic potential ( S5 Fig ) [43 , 52] . The large number of extracellular loop combinations seen in E . coli OmpC proteins suggests that these elements are modular and can be rearranged independently to generate novel porins . However , OmpC loops L4 , L5 , L6 , L7 and L8 engage in a series of intra-protomer interactions ( S4 Table ) , and therefore some loop combinations may be incompatible with a properly folded porin structure . Most intra-protomer loop interactions ( e . g . L5-L7 and L7-L8 ) are mediated by invariant residues , though the contacts between L4 and L5 in OmpCCFT073 and OmpC6104H involve non-conserved residues ( S4 Table ) . We note that chimeric OmpCCFT073 carrying L4 from OmpCF11 represents a L4-L5 combination that has yet to be found in any E . coli isolate . Because this OmpC chimera is functional as a CdiAEC536 receptor , it must be assembled properly in the outer membrane . Therefore , in this instance , loop L4 is clearly modular . This combinatorial complexity has implications for CDI , and almost certainly governs bacterial susceptibility to bacteriophages and immune recognition [53 , 54] . CDI receptor polymorphism restricts the number of susceptible target bacteria , yet CdiAEC536 recognizes multiple OmpC variants . OmpC proteins from E . coli EC93 , 536 , F11 , UTI89 and A33H are all recognized efficiently by CdiAEC536 though each receptor carries a distinct combination of extracellular loops . E . coli F11 and UTI89 encode CdiA proteins that share >98% identity with CdiAEC536 . Taken together , these observations indicate that the CdiAF11 and CdiAUTI89 effectors also bind to OmpF-OmpC and preferentially recognize self-receptors . These findings are consistent with numerous reports that CDI+ bacteria tend to auto-aggregate [1 , 7 , 11 , 13 , 21–23] . Intriguingly , E . coli CFT073 also encodes a CdiA effector that shares 89 . 7% identity with CdiAEC536 over 3 , 037 N-terminal residues , yet OmpCCFT073 does not function as a CdiAEC536 receptor . Inspection of the cdiB gene from E . coli CFT073 reveals that it contains an in-frame stop codon [55] , and therefore this strain presumably cannot secrete its CdiA effector . It is tempting to speculate that the loss of CdiACFT073 surface expression has allowed ompCCFT073 to diverge in response to other selective pressures . However , there are other E . coli isolates with intact cdi loci that also share ompC and ompF alleles with CFT073 . Based on the data presented here , it appears that the CdiA proteins from these latter isolates would not preferentially recognize self and therefore may function primarily in competition . Though auto-adhesion is beneficial for bacteria , these phenomena can also be viewed from the perspective of cdi genes as selfish elements . CDI systems are encoded on mobile genetic elements and their toxin/immunity activities act as "gene-addiction" modules [56] . Cells that lose CDI-encoding plasmids or genomic islands are suddenly susceptible to inhibition and risk elimination by neighboring CDI+ siblings . Thus , self-recognition would ensure continual selective pressure to retain cdi genes in a population . We speculate that CdiA effectors may have first evolved as adhesins used exclusively for cooperation . Toxin-delivery capabilities could then be grafted onto the ancestral CdiA protein , providing a strong positive selection for cdi retention . Thus , the antagonistic activity of CDI systems may have arisen from evolutionary pressure to retain cdi genes in fluid bacterial genomes .
Bacterial strains are listed in Table 1 . Bacteria were grown in lysogeny broth ( LB ) or on LB agar unless otherwise noted . Where indicated , media were supplemented with antibiotics at the following concentrations: ampicillin , 150 μg mL−1; kanamycin , 50 μg mL−1; chloramphenicol , 66 μg mL−1; spectinomycin , 50 μg mL−1; rifampicin , 200 μg mL−1; streptomycin , 100 μg mL−1 . The bamA::cat allele was introduced into MC4100 pZS21-bamA+ cells by bacteriophage P1-mediated transduction [20] . The ΔompC::kan , ΔompF::kan , Δwzb::kan and ΔrecA::kan disruptions were obtained from the Keio collection [57] and transduced into E . coli strain CH10013 . Kanamycin-resistance cassettes were subsequently removed with FLP recombinase expressed from plasmid pCP20 [58] . The ompCK-12 gene of E . coli CH10013 was replaced with ompCEC536 using Red-mediated recombination . ompCEC536 was amplified by PCR with oligonucleotides 3322 and 3323 , and a fragment upstream of ompCK-12 was amplified using oligonucleotide pair 3320/3321 ( all oligonucleotide primers are listed in S5 Table ) . The two DNA fragments were combined into one fragment by overlap-extension PCR ( OE-PCR ) using primers 3320/3323 and ligated to pKAN [59] using SacI/BamHI sites to generate plasmid pCH11341 . A fragment downstream of ompCK-12 was then amplified with primers 3324/3325 and ligated to pKAN using EcoRI/KpnI sites to generate plasmid pCH11342 . Fragments from pCH11341 and pCH11342 were amplified with primer pairs 3320/3505 and 3504/3325 ( respectively ) and the resulting products combined into one fragment by OE-PCR using primers 3320/3325 . The final product was introduced into E . coli CH10013 that express phage λ Red proteins from plasmid pSIM6 [60] and recombinant bacteria selected on LB agar supplemented with kanamycin . Plasmids constructs used in this study are listed in Table 2 . The E . coli cysK gene was amplified with primer pair 2111/2110 and ligated to pBR322 using EcoRV and BamHI restriction sites to generate plasmid pCH9763 . The ompFK-12 allele was amplified from E . coli JCM158 with primer pair 2385/2386 , and the fragment ligated to pZS21 using BamHI and XbaI restriction sites to generate plasmid pCH10202 . Additionally , ompFK-12 and ompFEC536 were amplified with primers 2729/2734 , and the fragments ligated to pTrcKX using KpnI and XhoI restriction sites to generate plasmids pCH11850 and pCH10780 . The ompF genes from S . Typhimurium LT2 and E . cloacae ATCC 13047 were amplified with primer pairs 3468/3469 and 2729/3470 ( respectively ) and ligated to pTrcKX as described for the E . coli genes . The ompC genes from S . Typhimurium LT2 and E . cloacae were amplified with primers 3326/2388 , digested with NcoI/XhoI and ligated to plasmid pTrc99a to generate pCH11852 and pCH11853 , respectively . All E . coli ompC alleles were amplified with primer pair 2387/2388 , followed by ligation to plasmid pTrcKX using KpnI and XhoI restriction sites . The coding sequences for OmpCCFT073 loops L4 , L5 , L7 and L8 of OmpCCFT073 were replaced using megaprimer PCR [61] . Plasmid pCH10778 template DNA was amplified with reverse primer 2388 in conjunction with primers 2935 ( L5EC536 ) , 2936 ( L7EC536 ) and 2937 ( L8EC536 ) . The resulting products were used as megaprimers to amplify ompCCFT073 in conjunction with forward primer 2387 . The chimeric products were digested with KpnI/XhoI and ligated to pTrcKX to generate plasmids pCH11334 , pCH11765 and pCH11335 . The same procedure and primers were used to generate ompCCFT073 constructs encoding two loops from ompCEC536 . The loop L4 exchange megaprimer was generated with primers 2387/2807 and used with 2388 to generate the final product . All PCR products were digested with KpnI and XhoI and ligated to plasmid pTrcKX . Dimeric OmpF expression plasmids were constructed through sequential ligation of ompF coding sequences . The ompFEC536 allele was amplified with primers 2733/2734 and ligated to pCH9181 using SpeI/XhoI restriction sites to generate intermediate construct pCH10796 . A second ompF fragment was amplified with primers 2732/2729 and ligated to pCH10796 using KpnI/SpeI restriction sites . The resulting pCH10797 ( ompF2 ) construct produces OmpFEC536 dimers linked by a Ser-Gly-Thr-Thr-Ser-Thr-Gly-Gly peptide . To introduce a longer linker sequence , ompF was amplified with primers 2729/2765 and ligated to pCH9181 using KpnI/SpeI sites to generate plasmid pCH11893 . The second ompF module was amplified with primers 3008/2734 and ligated to pCH11893 using BamHI/XhoI restriction sites . The resulting pCH12054 ( ompF2* ) construct produces OmpFEC536 dimers linked by a Ser-Gly-Thr-Gly-Ser-Asp-Thr-Ser-Gly-Gly-Thr-Asp-Gly-Thr-Gly-Gly peptide . Inhibitor and target bacteria were grown to mid-log phase in LB media supplemented with the appropriate antibiotics , then mixed at a 10:1 ratio in fresh LB medium without antibiotics and incubated for 4–5 h at 37°C with vigorous shaking in baffled flasks . Viable target-cell counts were enumerated on selective LB-agar as colony forming units ( cfu ) per mL . E . coli EPI100 carrying pDAL660Δ1–39 [1] or pDAL866 [33] were used as inhibitor cells for the comparison of CDIEC93 and CDIEC536 systems . The mock CDI−inhibitors were E . coli EPI100 carrying the pWEB::TNC vector . All other competitions used the E . coli 536 derivative DL6536 , which expresses cdiBAIEC536 under the control of an arabinose-inducible promoter [8] . The mock inhibitor strain DL6381 is a derivative of DL6536 that carries a deletion of the cdiA-CT coding sequence [8] . DL6536 and DL6381 were cultured in 0 . 2% arabinose to induce cdiBAI expression and competition co-cultures were supplemented with 0 . 2% arabinose . Plasmid pSC189 ( encoding the mariner transposon ) was introduced into E . coli CH10013 cells carrying pCH9763 ( pBR322::cysK ) through conjugation with E . coli MFDpir donors [62 , 63] . Donor and recipient cells were grown to mid-log phase in LB media supplemented with ampicillin and 30 μM diaminopimelic acid ( for donors ) . Donors ( ~6 . 0 ×108 colony forming units , cfu ) and recipients ( ~3 ×108 cfu ) were mixed and collected by centrifugation for 2 min at 9 , 000 rpm in a microcentrifuge . The supernatant was removed by aspiration and the cell pellet resuspended in 100 μL of 1× M9 salts . Cell mixtures were spotted onto 0 . 45 μm nitrocellulose membranes and incubated on LB-agar without inversion for 4 h at 37°C . Cells were then harvested into 2 mL of 1× M9 salts and transposon-insertion mutants selected on LB-agar supplemented with rifampicin and kanamycin . Over 50 , 000 colonies from each mating were harvested into 1× M9 salts , and inoculated into 50 mL of LB medium in a 250 mL baffled flask . Inhibitor cells ( E . coli 536 derivative DL6536 ) were grown in parallel in LB medium supplemented with L-arabinose until mid-log phase . Inhibitors and targets were mixed at a 10:1 ratio in fresh LB medium supplemented with L-arabinose and cultured for 4 h with shaking at 37°C . Viable target cells were enumerated as cfu/mL on LB agar supplemented with rifampicin and kanamycin . Survivors from the first round of CDIEC536 selection were harvested into 1× M9 salts and inoculated into 50 mL of LB medium supplemented with L-arabinose for a second round of selection . After the third round of selection , target-cell populations were completely resistant to the CDI inhibitor cells . Individual colonies were screened for CysK activity in MOPS minimal media containing 3 mM 1 , 2 , 4-triazole , which inhibits the growth of cysK+ bacteria [64] . CDIR phenotypes for all cysK+ clones were confirmed in competition co-cultures with E . coli 536 inhibitors . Transposon mutations were transduced into CDI-sensitive cells , and the resulting transductants tested in competition co-cultures for CDIR phenotypes . Overnight cultures of GFP-labeled E . coli DL4905 cells carrying pDAL660Δ1–39 ( CDIEC93 ) , pDAL866 ( CDIEC536 ) or pWEB::TNC ( CDI− ) were diluted into fresh tryptone broth ( TB ) and grown to mid-log phase at 30°C . Inhibitor cells were then mixed at a 5:1 ratio with DsRed-labeled E . coli CH10099 ( ΔompC ) or CH10100 ( ΔompF ) cells complemented with the indicated ompC and ompF expressing plasmids . Cell suspensions were incubated with aeration for 15 min at 30°C , diluted 1:50 into filtered 1× phosphate-buffered saline ( PBS ) , and then analyzed on an Accuri C6 flow cytometer using FL1 ( 533/30 nm , GFP ) and FL2 ( 585/40 nm , DS-Red ) fluorophore filters ( Becton Dickinson ) as described [36] . Bacteria were grown to mid-log phase in LB media supplemented with ampicillin . Cells were collected by centrifugation at 6 , 000 ×g . Cell were frozen at –80°C , then broken by freeze-thaw cycles in 8 M urea , 150 mM NaCl , 50 mM Tris-HCl ( pH 8 . 0 ) . Extracts were quantified by Bradford assay and equal amounts of protein were resolved on 6 M urea/10% SDS polyacrylamide gels for 4 h at 100 V . Gels were electroblotted onto nitrocellulose membranes and OMPs detected using polyclonal antisera raised against E . coli OmpC/OmpF and BamA . Immunoblots were visualized using IRDye 680 ( LI-COR ) labeled anti-rabbit secondary antibodies and an Odyssey infrared imager . Colicin E5 was purified in complex with its His6-tagged immunity protein as described previously [6 , 65] . E . coli strains CH12180 and CH12345 carrying pCH11850 ( ompF ) , pCH10796 ( ompF-ompF , 8-residue linker ) , or pCH12054 ( ompF-ompF , 16-residue linker ) were grown to mid-log phase in LB media supplemented with ampicillin , then collected by centrifugation and re-suspended in fresh pre-warmed media at OD600 = 0 . 05 . After 30 min of incubation with shaking , purified colicin E5 was added to a final concentration of 1 μM and cell growth monitored by measuring the OD600 every 60 min . | Bacterial pathogens often live in crowded communities where cells reside in close contact with one another . Many of these bacteria possess contact-dependent growth inhibition ( CDI ) systems , which allow cells to touch and inhibit each other using toxic CdiA proteins . CDI+ bacteria also produce immunity proteins that specifically protect the cell from the CdiA toxins of neighboring sibling cells . The CDI system from Escherichia coli EC93 was the first to be characterized and its CdiA toxin recognizes a receptor ( BamA ) that is identical in virtually all E . coli isolates . Here , we describe a different CDI system from uropathogenic E . coli 536 , which causes urinary tract infections . In contrast to E . coli EC93 , CdiA from E . coli 536 binds to receptor proteins ( OmpC/OmpF ) that vary widely between different E . coli isolates . Thus , uropathogenic E . coli preferentially bind and deliver toxins into sibling cells and other closely related E . coli strains . These results suggest that CDI systems distinguish between "self" and "non-self" cells . Moreover , because sibling cells are immune to CdiA-mediated growth inhibition , these findings raise the possibility that toxin exchange may be used for communication and cooperative behavior between genetically identical bacteria . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"sequencing",
"techniques",
"complement",
"system",
"medicine",
"and",
"health",
"sciences",
"toxins",
"immune",
"physiology",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"immunology",
"microbiology",
"toxicology",
"toxic",
"agents",
"plasmid",
"construction",
"membrane",
"proteins",
"bacterial",
"diseases",
"outer",
"membrane",
"proteins",
"enterobacteriaceae",
"dna",
"construction",
"molecular",
"biology",
"techniques",
"sequence",
"motif",
"analysis",
"cellular",
"structures",
"and",
"organelles",
"bacteria",
"bacterial",
"pathogens",
"salmonella",
"typhimurium",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"immune",
"system",
"proteins",
"infectious",
"diseases",
"sequence",
"alignment",
"proteins",
"medical",
"microbiology",
"microbial",
"pathogens",
"molecular",
"biology",
"cell",
"membranes",
"salmonella",
"immune",
"system",
"biochemistry",
"cell",
"biology",
"physiology",
"biology",
"and",
"life",
"sciences",
"organisms"
] | 2016 | CdiA Effectors from Uropathogenic Escherichia coli Use Heterotrimeric Osmoporins as Receptors to Recognize Target Bacteria |
The balance of quiescence and cell division is critical for tissue homeostasis and organismal health . Serum stimulation of fibroblasts is well studied as a classic model of entry into the cell division cycle , but the induction of cellular quiescence , such as by serum deprivation ( SD ) , is much less understood . Here we show that SS and SD activate distinct early transcriptional responses genome-wide that converge on a late symmetric transcriptional program . Several serum deprivation early response genes ( SDERGs ) , including the putative tumor suppressor genes SALL2 and MXI1 , are required for cessation of DNA synthesis in response to SD and induction of additional SD genes . SDERGs are coordinately repressed in many types of human cancers compared to their normal counterparts , and repression of SDERGs predicts increased risk of cancer progression and death in human breast cancers . These results identify a gene expression program uniquely responsive to loss of growth factor signaling; members of SDERGs may constitute novel growth inhibitors that prevent cancer .
Quiescence , also termed G0 , is defined as reversible cell cycle arrest where cells are poised to re-enter the cell cycle . Most eukaryotic cells spend the majority of their lifespan in the state of quiescence . In response to injury or specific extracellular stimuli , many types of somatic cells can quickly leave the quiescent state and enter the cell division cycle . For instance , in the skin , dermal fibroblasts and hair follicle stem cells are for the most part quiescent [1 , 2] . Injury to the skin stimulates fibroblasts and epidermal stem cells to rapidly proliferate; once tissue repair has been accomplished , the cells exit the cell cycle and reenter quiescence . Similarly , memory lymphocytes are quiescent as they circulate and survey the body , dividing only when stimulated by cognate antigenic stimuli to mount an immune response [3] . In addition to the absence of cell division , quiescent cells exhibit systematic differences in their metabolism and propensity for differentiation , which may help to ensure the reversibility of quiescence [4] . The ubiquity of quiescence as a central feature of cell life suggests that its regulation may be critical to normal development , degenerative diseases , and cancer [1 , 3 , 4] . Serum , the soluble fraction of clotted blood , is an important mitogenic signal in wound healing and tissue homeostasis . Many key genes involved in cell cycle entry were initially identified by their unique temporal patterns of expression in response to serum stimulation ( SS ) and are dysregulated in cancer [5] . In addition to cell cycle entry , serum induces a transcriptional program activating many aspects of wound healing [6] . This wound response program is recapitulated in many human cancers and is a strong predictor of tumor progression for these cancers [7 , 8] . While much is known about the signal transduction pathways , transcription factors , and immediate early genes that mediate the exit from quiescence and entry to the cell division cycle [5 , 6 , 9 , 10] , comparatively much less is known about the mechanisms by which cells enter the quiescent state . Growth factor deprivation , contact inhibition , and loss of adhesion can each induce a shared set of quiescence genes [4] , indicating the potential existence of multiple pathways to quiescence . Several tumor suppressor genes , such as Rb and PTEN , are required for quiescence maintenance in low serum conditions [11–13] . Yamamoto and colleagues have identified a set of antiproliferative genes whose repression requires ongoing activity of the mitogen-responsive kinase ERK during cell cycle progression [10] . The induction of these genes during quiescence is therefore simply a consequence of the absence of mitogen-induced signaling . If this mode of balanced regulation were generally applicable , then one might predict a symmetric network of gene regulation during quiescence entry and exit . Alternatively , an inducer of quiescence may engage a unique transcriptional program that is not regulated by cell cycle entry . Such quiescence entry-specific genes may represent novel growth inhibitors that link extracellular stimuli to the physiologic state of quiescence . In this report , we characterize the genomic expression program of serum deprivation ( SD ) in fibroblasts and identify the predominance of asymmetric regulation in quiescence induction . We identify two putative tumor suppressor genes , SALL2 and MXI1 , as key regulators of the serum deprivation early response genes ( SDERGs ) and demonstrate the roles of the SDERGs in cell cycle exit and human cancer progression .
To understand the genomic program of entry into quiescence , we characterized the temporal pattern of the genome-wide transcriptional profiles of fibroblasts in response to SD . We employed the same diploid fibroblast culture and experimental time points that we previously used to delineate a detailed transcriptional response to SS [7] , thereby enabling a systematic comparison of entry and exit of quiescence . Human foreskin fibroblasts were grown in media containing 10% fetal bovine serum ( FBS ) for 48 hours , and following switch to low serum media containing 0 . 1% FBS , harvested at 15 time points ranging from 15 minutes to 48 hours after SD . Total RNA was extracted , amplified , and hybridized along with human universal reference RNA onto human cDNA microarrays containing ~43 , 000 elements , representing ~23 , 000 unique genes . Comparison of the temporal expression profiles of SD and SS revealed two dynamic programs with marked asymmetry in the early responses ( Figure 1A ) . For each gene , we quantified the similarity of its pattern of expression during SD and SS by a Pearson correlation . ( A correlation of −1 indicates exact opposite pattern; a correlation of 0 indicates no relationship; a correlation of +1 indicates identical pattern . ) Strikingly , genes with an induction or repression onset after eight hours of treatment showed symmetric regulation by SS and SD with a Pearson correlation of −0 . 80 ( i . e . , genes induced by SS are repressed by SD and vice versa ) . In contrast , genes regulated within the first three hours of SS or SD were asymmetrically governed by these two stimuli ( Pearson correlation −0 . 2 to +0 . 2 ) . Genes that were regulated within three to eight hours of the treatments had an intermediate level of symmetry . These results thus suggest an asymmetric regulation of quiescence entry and exit with distinct sets of early response genes . To gain a higher order view of the transcriptional programs of SS and SD , we next identified the functional groups and upstream cis-regulatory sequences of genes that are regulated in expression by these two stimuli ( Figure 1B and 1C ) . Using the module map method [14] , we identified for each array the coordinate induction or repression of 1 , 735 gene sets , each defined as a group of genes encoding proteins possessing a shared biological function , biochemical process , or subcellular localization by gene ontology ( p < 0 . 05; false discovery rate [FDR] < 0 . 05 ) ( Figure 1B ) [15] . This higher order view confirmed that many functions known to be regulated by serum [6] , such as cell proliferation , RNA metabolism , and sterol synthesis , are symmetrically regulated by SS and SD in the late phase of each program ( Figure 1B; Figure S1 ) . In contrast , while coordinate induction of genes encoding transcription factors and signaling proteins characterized the early response of SS , the genes that were regulated in the early response of SD were enriched for functions in immune response , redox , and extracellular matrix metabolism ( Figure 1B ) . The paucity of gene ontology terms describing functional groups characterizing early SD may reflect the current scarcity of knowledge about this cellular state . Transcriptional regulation is mediated in large part by binding of trans-acting factors to cis-regulatory elements upstream of genes . To better understand the regulation of SS and SD , we therefore mapped the genome-wide occurrence of 175 phylogenetically conserved cis-regulatory motifs [16] in four kilobases surrounding transcriptional start sites genome-wide . For each array , we identified cis-regulatory motifs that are significantly enriched in the genes that are induced or repressed , yielding a map of the regulatory motifs active in quiescence entry and exit ( Figure 1C ) . This unbiased approach highlighted many known transcription factors that play key roles in this process and again reveals the marked asymmetry in the early response of SS and SD ( Figure 1C ) . For instance , the gene set defined by enrichment of motifs for E2F and DP transcription factors contained many cell cycle genes and was symmetrically regulated late in the response by SS and SD ( Figure 1C; Figure S1 ) . In contrast , gene sets defined by the motif of AP-1 , MEF2 , or NF-κB were induced early in response to SS but were not substantially regulated by SD . Once again , we noted a relative paucity of cis motifs that identify early response genes to SD . This result may reflect incomplete information of the relevant transcription factors and their cognate cis motifs or may reflect additional post-transcriptional mechanisms in regulating mRNA levels in SD . To begin to understand the asymmetric regulation of quiescence entry and exit , we focused on genes that are induced early in response to either SS or SD . Approximately half of the well-known early response genes to SS , such as EGR1 , CYR61 , and GADD45B , were correspondingly repressed by SD , albeit with lower amplitude and delayed kinetics that peak at approximately three hours after SD ( Figure 2A ) . In contrast , other key SS early response genes , such as FOS , JUNB , and MYC , are transiently induced by SS and showed little change in transcript level in SD ( Figure 2A ) . Conversely , a group of 135 genes defined by early induction in response to SD , which we term SDERGs , showed immediate induction within the first three hours after SD , but demonstrated substantially less or delayed regulation by SS ( Figure 2B and 2C; Table S1 ) . Some SDERGs , including SALL2 , MXI1 , and TNKSBP1 , are induced in a sustained manner by SD while other SDERGs , such as SPRY4 and SMAD7 , are induced by SD in a transient manner . Intriguingly , SALL2 and MXI1 are both putative tumor suppressors that can suppress cell growth when overexpressed [17 , 18] . SALL2 encodes a zinc finger transcription factor and is a homolog of Drosophila homeotic gene Spalt . Human SALL2 resides in a chromosomal region frequently deleted in ovarian cancers , and SALL2 protein is a binding target of the oncogenic large T-antigen from polyoma virus [19] . MXI1 is a member of the MAD family of potent antagonists of MYC oncoproteins [20] . MXI1 resides on a locus in human Chromosome 10 that is deleted in several types of human cancers , including prostate cancer , and deletion of MXI1 in mouse leads to a cancer-prone phenotype [18] . We noted that several SDERGs are well-known interferon-induced genes , such as STAT1 , ILIR1 , BDKRB2 , and PLSCR1; several additional SDERGs , such as JUND , IFIT2 , and G1P2 are predicted by our cis-regulatory map to contain a motif for the interferon regulatory factor ( IRF ) family of transcription factors . A likely candidate is IRF1 because IRF1 can induce several of these genes , is itself induced by SD , and possesses antiproliferative properties [21 , 22] . The asymmetric regulation of early response genes suggests that the transition to a quiescent state may be enforced by employing signaling pathways unique to SD . To address the functional role of the SDERGs , we used RNA interference to examine the requirement of specific SDERGs for cell cycle exit . We selected four candidate genes ( SALL2 , MXI1 , IRF1 , and TNKS1BP1 ) that encode transcription factors or signaling proteins that may regulate quiescence induction . TNKS1BP1 , a putative telomere binding protein , was included because of the reported roles of telomerase in enhancing S-phase progression [23 , 24] . We transfected primary human foreskin fibroblasts grown in high serum ( 10% FBS ) media with silent interfering RNA ( siRNA ) pools corresponding to each of four candidate genes or a control siRNA targeting GFP . The cells were switched to 0 . 1% serum media 72 hours after transfection to induce quiescence , and DNA synthesis was measured 16 hours later by 5-bromo-2′-deoxyuridine ( BrdU ) incorporation . Reverse transcription-PCR confirmed decreased expression of the target mRNAs ( Figure S2 ) . In control cells treated with siRNAs targeting GFP , SD lowered the percentage of BrdU+ cells from 65% to approximately 25% , indicating induction of quiescence and efficient cell cycle exit ( Figure 3A and 3B ) . Strikingly , cells treated with siRNAs targeting SALL2 doubled the number of BrdU+ cells after SD ( p < 0 . 0001 ) , while siRNAs targeting MXI1 showed modest but consistent increase in BrdU+ cells ( p < 0 . 02 ) . siRNAs to IRF1 or TNKS1BP1 did not significantly affect quiescence induction under the conditions tested . To test the potential functional relationships between SALL2 , MXI1 , and IRF1 , we treated cells with pair-wise combinations of siRNAs . Silencing of IRF1 strongly cooperated with silencing of MXI1 , but silencing of neither IRF1 or MXI1 cooperated with silencing of SALL2 to prevent cell cycle exit . Fluorescence-activated cell sorting analysis of DNA content confirmed that depletion of SALL2 or MXI1 blocked the ability of cells to arrest in G0–G1 after growth factor deprivation and instead led to inappropriate progression through S and G2/M phases of the cell cycle ( Figure 3C ) . These results suggest that several of the SDERGs identified by our microarray screen are required for entry to cellular quiescence and that SALL2 and MXI1 may trigger different pathways to enforce quiescence . To further delineate the mechanisms of SALL2 and MXI1 action , we identified genes that required SALL2 and MXI1 for proper regulation during quiescence induction . RNA from cells transfected with silent interfering ( si ) SALL2 and siMXI1 was extracted , amplified , and compared with RNA from cells treated with siGFP on cDNA microarrays . Genes whose expression levels were consistently changed by loss of SALL2 or MXI1 were identified , and their temporal regulation by SS and SD were systematically organized by hierarchical clustering . We observed three main patterns of gene regulation . First , both SALL2 and MXI1 are individually required for the induction of a cluster of SD middle response genes ( Figure 4 , cluster 1 ) . After knockdown of either SALL2 or MXI1 in SD , these genes reverted to a pattern of expression more closely resembling their normal behavior in SS than in SD . Second , in contrast to this shared role in SD gene induction , SALL2 and MXI1 acted to repress mutually exclusive sets of SD-repressed genes ( Figure 4 , cluster 3 ) . Third , SALL2 appears to have a unique role in limiting expression of a set of middle response genes to SD , as their expression became super-induced when SALL2 was silenced ( Figure 4 , cluster 2 ) . These results confirm the distinct roles of SALL2 and MXI1 in quiescence induction and suggest multiple roles for SALL2 in gene regulation throughout quiescence entry and maintenance . Having discovered the SDERGs as a set of 135 genes specifically induced when fibroblasts enter quiescence , we next tested whether the SDERGs might have broad roles in growth inhibition . We reasoned that if SDERGs were generally required to induce cell quiescence , then SDERGs might be coordinately repressed in conditions of excessive cell proliferation , such as in cancer . We therefore interrogated a compendium of 1 , 973 microarrays representing 22 human tumor types to search for enriched coregulation of the 135 SDERGs , using the gene module map method [14] . The SDERGs were indeed coordinately repressed in many conditions that represent pathologic proliferation , specifically the subset of fast doubling cell lines in the NCI60 collection of tumor cell lines ( p < 10−10 ) and several human cancers relative to their normal tissue counterpart including cancers of prostate ( p < 10−6 ) , blood ( p < 10−12 ) , and lung ( p < 10−4 ) ( Figure 5A ) . These results suggest that SDERGs likely antagonize cell proliferation in many cell types . Cancer consists of a broad range of clinical behaviors ranging from indolent tumors to aggressive metastatic disease . To further dissect the potential molecular variation underlying this clinical heterogeneity and to extend and validate our results , we tested the prognostic power of the SDERG gene set in independent datasets and different subtypes of human cancer . Analyzing DNA microarray data from a study of 103 prostate tissues and cancer [25] , we found that coordinate repression of SDERGs could identify over 90% of prostate tumors relative to normal prostate , a finding very unlikely due to chance alone ( p < 10−11 ) ( Figure 5B ) . Furthermore , expression of SDERGs in a set of 295 breast cancers [26] naturally divided the breast tumors into two groups ( Figure 5C ) . Patients with breast cancers that diminished expression of SDERGs had significantly worse survival ( p < 10−5 ) and significantly increased probability of metastasis ( p < 10−4 ) . This group of tumors with repression of SDERGs also tended to be of the grade 3 tumors ( p < 10−9 ) , which are defined by higher cell proliferation and less differentiation . Interestingly , the mRNA levels of SALL2 and MXI1 , either alone or in combination , were insufficient to predict overall survival or metastasis-free survival ( p > 0 . 05 , Cox-Mantel test ) ; conversely , removal of SALL2 and MXI1 from SDERGs did not affect the prognostic power of the SDERG gene set ( unpublished data ) . These results further suggest a role for SDERGs to prevent excessive and pathologic proliferation . By reflecting the propensity for quiescence , the expression level of the SDERG gene set as whole may aid in predicting tumor behavior in two of the most common human cancers . To better understand the transcriptional regulation of quiescence initiation versus maintenance , we compared SDERGs ( 135 genes ) with 116 genes previously found to be concordantly induced by prolonged entry into quiescence , four days after growth factor deprivation , contact inhibition , or loss of adhesion [4] . A total of nine genes were in common between these two gene sets , while only one overlap gene is expected by chance alone ( p < 10−7 , hypergeometric distribution ) . The overlap genes include MXI1 and STAT1 , thus indicating an interesting but limited overlap between the SDERGs and genes expressed during quiescence maintenance ( Figure 6A ) . We next examined the coordinate expression of these 116 quiescence maintenance genes in 1 , 973 microarray representing 22 human tumor types ( Figure 6B ) . SDERGs and quiescence maintenance genes showed overlapping but distinct patterns of expression , with some tumors coordinately repressing both gene sets but many that repress only one of the two sets . In human prostate cancer , quiescence maintenance genes are typically repressed but in far more haphazard fashion compared to SDERGs ( p value of the separation is five logs of magnitude worse ) ( Figure 6C ) . Similarly , coordinate repression of quiescence maintenance genes is modestly predictive of primary breast cancer survival but not predictive of metastasis-free survival ( Figure 6D and unpublished data ) . These results suggest that genes mediating entry into quiescence are largely distinct from those associated with quiescence maintenance , and the two programs may be repressed in distinct fashions to facilitate the progression of specific types of human cancers . The general association between SDERGs and cell cycle exit raises the possibility that SDERGs may be induced by additional stimuli . For instance , in response to variety of noxious stress , cells will exit the cell cycle as part of the stress response . To test the possibility that SDERGs may be induced by stress , we queried the expression pattern of SDERGs in the published gene expression data of fibroblasts exposed to multiple types of stress [27] . In contrast to the strong induction of SDERGs by SD , exposure of fibroblasts to the reducing agent dithiothreitol ( causing protein unfolding and endoplasmic reticulum stress ) , heat shock , or menadione ( inducing oxidative stress ) did not induce the SDERGs ( Figure 7A ) . These results reaffirm SDERGs as a program uniquely responsive to loss of growth factor signaling as represented by SD .
By determining the genomic transcriptional program in response to SS and SD , we observed the asymmetric regulation of quiescence entry and exit ( Figure 1 ) . While the late transcriptional responses to these two opposing stimuli are largely symmetric , this symmetric program results from two distinct early transcriptional responses . These data suggest that in addition to previously identified antiproliferative genes that require ongoing growth factor signaling for their suppression [10] , a major mechanism of quiescence entry is the induction of a set of unique quiescence entry genes ( Figure 2 ) . By simple analogy , a speeding car may be slowed by releasing the gas pedal , but the car can also be brought to a screeching halt by releasing the gas pedal and stepping on the brakes . We suggest that the SDERGs may be a set of brakes for cell cycle progression and growth factor-induced gene expression . Just as a gain of growth factor signaling activates the classic immediate early genes to induce cell cycle entry , a loss of growth factor signaling uniquely activates the SDERGs to induce cell quiescence . The decision of cell proliferation or quiescence is thus determined by the balance of growth factor-induced genes and SDERGs ( Figure 7B ) . SDERGs are also actively involved in the repression of growth factor-induced genes . For instance , SALL2 and MXI1 are required to repress distinct sets of serum-inducible genes ( Figure 3C ) ; in the case of MXI1 this may occur by direct competition with MYC for binding to promoters of serum-inducible genes [20] . Thus , there is likely cross regulation of the early transcriptional responses to growth factor stimulation and deprivation . One potential reason for this dual transcriptional response to growth factor gain and loss is to achieve tight regulation . It has been well known that many early response genes to SS are induced as a precise pulse that then rapidly decays despite continued mitogen presence ( Figure 2A ) , thereby providing a check against unlimited proliferation and the risk of cancer . For instance , the classic mitogen-induced proto-oncogene MYC is regulated by transcriptional autorepression [28] , mRNA instability [29] , and rapid protein turnover [30 , 31]; enforced MYC expression is sufficient to induce ectopic DNA replication and DNA damage within just one cell cycle [32] . The low level of MYC and other growth factor early response genes at steady state would render a system based on their further decrement an insensitive strategy to detect growth factor deprivation . Instead , a decrement of growth factor signaling triggers a robust transcriptional response of SDERGs , leading to additional quiescence gene induction and exit of cell cycle . Intriguingly , a large number of SDERGs ( including SALL2 and MXI1 , which are required for entry to quiescence ) are induced and maintained in stable expression in response to SD ( Figure 2B ) . The longevity of SDERG expression in contrast to the transient expression of growth factor early response genes may provide an explanation for cell quiescence as the default state of most eukaryotic cells . Among the several the SDERGs we tested , silencing of SALL2 expression had the most substantial effect on cell cycle exit in response to growth factor deprivation . Previously , Benjamin and colleagues had identified SALL2 as an antiproliferative gene by virtue of a tumor host range selection procedure for the oncogenic polyoma virus [19] . Enforced SALL2 expression in ovarian cancer cells inhibits tumor xenograft growth in vivo and can induce the expression of cyclin-dependent kinase inhibitor p21 , although it is unclear whether p21 is the sole mechanism by which SALL2 elicits cell cycle arrest [17] . The biological context in which SALL2 might exert its antiproliferative effect was also not known . Our results suggest that SALL2 is induced in response to loss of growth factor signaling ( Figure 2B ) . Acute loss of SALL2 during SD blocked the ability to stop DNA synthesis and induce additional quiescence-associated genes , suggesting that SALL2 is required for quiescence induction in response to growth factor deprivation ( Figure 3C ) . Among the mammalian SALL family of zinc finger transcription factors , mutation of SALL1 leads to developmental abnormalities [33] , and SALL4 is required for maintenance of pluripotency in embryonic stem cells [34 , 35] . Surprisingly , SALL2 knockout mice are viable and have no obvious phenotype [36] , raising the possibility that SALL2's function may be redundant or compensated by other SALL family members . Indeed , we found that knockdown of SALL4 also blocks cessation of DNA synthesis in response to SD even though SALL4 mRNA level does not change in response to SD ( H . Liu and H . Y . Chang , unpublished data ) . Thus , it may be the total pool of SALL transcription factors in the cell that determines cell quiescence , and SALL4 may compensate for the chronic loss of SALL2 during development . Coller et al . have shown that quiescence in fibroblasts inhibits their trans-differentiation ( such as into muscle cells in response to enforced MyoD expression ) [4]; the role of SALL transcription factors in cell quiescence may therefore be intimately linked to their roles in stem cell pluripotency [34 , 35] . The mechanisms by which the SALL2 message accumulates during SD and the functional roles of newly discovered SALL-regulated genes in SD should be addressed in future studies . Because cell quiescence has been postulated to be an important safeguard against cancer [1] , we reasoned that a transcriptional program mediating entry into quiescence might be systematically repressed in human cancers . Our survey of nearly 2 , 000 microarrays representing diverse types of human cancers identified multiple tumor types in which SDERGs are coordinately repressed ( Figure 4 ) . In addition , we found that repression of SDERGs unambiguously distinguished prostate cancers from adjacent prostate tissues , and repression of SDERGs in human breast cancers further predicted aggressive clinical course of early stage tumors . These properties are specific to SDERGs and are present to a much lower extent in genes associated with quiescence maintenance ( Figure 6 ) . Combined with the evidence that several SDERGs are required for cell cycle exit , these results highlight a potentially important role for the ability of cells to sense decrements of growth factor signaling and respond by quiescence . SDERGs and other genes induced by stimuli that induce cell quiescence may represent previously unrecognized tumor suppressor pathways; better understanding of these transcriptional programs may lead to new avenues of cancer diagnosis and treatment .
Primary human foreskin fibroblasts ( CRL 2091; American Type Culture Collection , http://www . atcc . org ) were cultured in DMEM plus 10% FBS . Cells were plated at 10% confluence . Cells were switch to DMEM plus 0 . 1% FBS 48 h after the last passage and harvested at the indicated time points . Construction of human cDNA microarrays containing approximately 43 , 000 elements , representing approximately 23 , 000 different genes , and array hybridizations were as previously described [8] . Total RNA was extracted using Trizol according to the manufacturer's instructions ( Invitrogen , http://www . invitrogen . com ) and amplified using the Ambion MessageAmpII aRNA kit ( Ambion , http://www . ambion . com ) . For time course experiments , Human Universal Reference RNA ( Stratagene , http://www . stratagene . com ) was used as reference RNA to compare with RNA from individual time points . We took four independent samples at time zero , which functioned as the baseline for other sample time points . For siRNA samples , siGFP was used as the reference RNA to compare with RNA from cells transfected with siSALL2 or siMXI1; each comparison was performed in duplicate . Full microarray data are available for download at Stanford Microarray Database ( http://genome-www5 . stanford . edu ) or Gene Expression Omnibus ( http://www . ncbi . nlm . nih . gov/geo ) . We selected genes for which the corresponding array elements had fluorescent hybridization signals at least 1 . 5-fold greater than the local background fluorescence in the reference channel and further selected genes for which technically adequate measurements were obtained from at least 80% of the samples in a given dataset . The zero time point for the SD experiment was performed in quadruplicate , and the four gene expression measurements were averaged and subtracted from those of the subsequent time points in order to visualize gene induction or repression over time . The gene expression profiles of the same cells to SS over 15 time points were similarly transformed by subtraction of expression value from each time point by that of the zero time point . The two datasets were then merged by matching Stanford clone IDs of the cDNA probes . We next focused on genes that exhibited substantial variation in expression and selected the subset of genes that were induced or repressed by at least four-fold in at least one array in either time course , yielding 444 cDNA probes ( henceforth genes ) . These genes were organized and grouped based on the similarity in their patterns of expression by average linkage clustering using the Cluster software [37] . Clustering of genes revealed three main transition points of gene expression variation—immediately after SS or SD , three hours after the stimuli , or eight hours after the stimuli . We therefore defined a time for the induction or repression of each gene cluster as the time point at which the gene expression reaches half maximal induction or repression , and classified each cluster as being regulated early ( <3 h ) , middle ( 3–8 h ) , and late ( >8 h ) . For gene clusters that exhibit biphasic or more complex patterns of regulation , we defined the onset of activity based on the first peak of expression variation . To quantify the degree of divergence among early- , mid- , and late-response genes , we calculated the Pearson correlation between the expression pattern of the SS and SD time courses for each gene . After ordering genes by hierarchical clustering , we visualized the Pearson correlation of gene clusters by displaying the moving average of correlation values of the ten nearest genes along the y-axis of the heat map . The gene module map of functional groups of genes was as described using the software package Genomica [14] . Briefly , for each microarray , we identified genes that were induced or repressed by at least 2-fold and tested for their enrichment in each of 1 , 735 gene sets defined by gene ontology terms [15] using the hypergeometric distribution . An FDR calculation was used to account for multiple hypothesis testing . Enrichments that had p < 0 . 05 and FDR < 0 . 05 were considered significant and are shown in Figure 1B , yielding a higher order view of each gene expression profile as sets of activated and repressed functions . For the cis-regulatory motif map , we first defined motif modules from the data of Xie et al . [16] , where each motif module is a group of genes that shared enrichment in an evolutionarily conserved cis-regulatory motif in their upstream regulatory sequences . The upstream regulatory region of each gene is defined as the 4 , 000 base pairs centered at the annotated transcription start site , as was done by Xie et al . ; 175 motif modules were defined . Second , for each array , we identified genes that were induced or repressed by 2-fold and tested for their enrichment in each of the motif modules using the hypergeometric distribution as described above . Modules with significant enrichment ( p ≤ 0 . 05 ) were identified and shown in Figure 1C , yielding a higher level view of each expression profile as a combination of activated and repressed cis-regulatory motifs . To conduct microarray analysis of siRNA treated cells we employed a type I microarray design where mRNA of cells treated with siSALL2 or siMXI1 ( labeled with Cy5 ) was compared to mRNA of cells treated with siGFP ( labeled with Cy3 ) by competitive hybridization to cDNA microarrays . We selected for analysis genes for which the corresponding array elements had fluorescent hybridization signals at least 1 . 5-fold greater than the local background fluorescence in the reference channel , and we further restricted our analyses to genes for which technically adequate measurements were obtained from both duplicate arrays . We further selected genes that were induced or repressed by at least 2-fold in two arrays by siSALL2 or siMXI1 , and the genes were organized by average linkage clustering . Regarding the cancer compendium and clinical outcome data , a cancer compendium of 1 , 973 microarrays was as described [14] . Gene probes from different microarray platforms were mapped by LocusLink identification numbers ( ( http://www . ncbi . nlm . nih . gov/entrez/query . fcgi ? db=gene ) . To test for the coordinate regulation of SDERGs in human cancers , we defined the 135 SDERGs as one gene set and tested each expression profile in the compendium for coordinate induction or repression of the SDERGs using the module map method in Genomica [14] . Specifically , for each microarray , we identified genes that were induced or repressed by at least 2-fold , and tested for their enrichment in SDERGs over that expected by chance alone ( p < 0 . 05 , FDR < 0 . 05 ) . Next , to identify enriched clinical annotations among samples that exhibit coordinate SDERG induction or repression , for each annotation we compared the frequency of SDERG induction or repression among the samples versus that expected by chance alone . We found many significant enriched annotations of SDERG repression but not of SDERG induction in cancer ( p < 0 . 05 , FDR < 0 . 05 , hypergeometric distribution ) . Several of the top ten enriched clinical annotations are reported in Figure 5A . The exact same procedure was repeated for 116 quiescence maintenance genes defined by Coller et al . [4] . To examine the expression of SDERGS in an independent set of human prostate cancer , we used the published prostate cancer microarray data of Lapointe et al . [25] . Gene probes were matched by cDNA clone IDs as this dataset was also generated on Stanford cDNA microarrays . We computed the average expression value of the 135 SDERGs in each of the 103 samples and rank ordered the average SDERG expression values . We considered those samples with average SDERG expression value greater than the mean of all 103 samples to have induction of SDERGs and those samples with average SDERG expression value to be below the mean of all 103 samples to have repression of SDERGs . The significance of the observed grouping of over 90 percent of prostate cancers with repression of SDERGs compared to normal prostate was evaluated by a two-by-two chi-square test , yielding a p value of 10−11 . p Values were also calculated using the same procedure for the 116 quiescence maintenance genes and the 107 genes unique to Coller et al . [4] and plotted in Figure 6C as the −log10 ( p value ) . To examine the clinical significance of SDERG repression in human breast cancer , we used the published breast cancer microarray data of van de Vijver et al . [26] . Gene probes were matched by Unigene ID ( http://www . sgn . cornell . edu/bulk/input . pl ? mode=unigene ) , and the 295 breast cancer samples were organized by two-way hierarchical clustering based on the expression pattern of the SDERGs . The main bifurcation of the dendrogram separated the breast cancer samples into two groups , one group with coordinate induction of SDERGs ( termed “SDERG up” ) and one group with coordinate repression of SDERGs ( termed “SDERG down” ) . We compared the differences in overall survival and metastasis-free survival of these two groups of patients as defined by SDERGs using the Cox–Mantel test in the program Winstat ( R . Fitch Software , http://www . winstat . com ) . Data from the 295 breast cancer samples were obtained for the quiescence maintenance genes unique to Coller et al . [4] and the same procedures as described above were performed . To examine the expression pattern of SDERGs in fibroblasts undergoing cell stress , we downloaded the published microarray data of Murray et al . [27] . Gene probes were matched by Stanford cDNA clone IDs . Expression of SDERGs during SD or exposure to DTT , heat , or menadione are shown as heat maps , and the average expression values are shown across each time course as a graph in Figure 5A . Cells were transfected with 20 nM of siRNA pools corresponding to each of the target genes ( SALL2 , MXI1 , IRF1 , and TNKS1BP1 ) and a control ( siGFP ) using DharmaFECT3 according to the manufacturer's instructions ( Dharmacon , http://www . dharmacon . com ) . Fibroblasts were transfected with 20 nM of siRNAs at a density of 2 × 105 cells/well ( six-well plate ) in high serum ( 10% FBS ) media . After 24 h , the treated fibroblasts were replated at a density of 6 × 103 cells/well in four-well chamber slides and were allowed to recover in high serum for 48 h . The transfected cells were then transferred to low serum ( 0 . 1% FBS ) media for 16 h . The transfection efficiency of each siRNA was verified by qRT-PCR ( Figure S1 ) . DNA synthesis was monitored by measuring the incorporation of the thymidine nucleotide analog BrdU ( Sigma , http://www . sigmaaldrich . com ) into DNA as previously described [11] . Briefly , cells were incubated with 10 μM BrdU in the media for 6 h; then washed with PBS , fixed , and stained with an anti-BrdU monoclonal antibody ( Becton Dickinson , http://www . bd . com ) and Alexa Fluor-conjugated goat anti-mouse antibody ( Molecular Probes , http://www . invitrogen . com ) . The percentage of BrdU-positive cells among >200 DAPI-positive cells in four random fields was recorded . Propidium iodide staining of DNA content and FACS analysis were performed as described [38] , with four replicate samples for each condition . Gene expression levels for genes targeted by the siRNAs were quantitated using RNA extracted from the transfected cells by Taqman quantitative one-step RT-PCR ( Applied Biosystems , http://www . appliedbiosystems . com ) . Taqman probes to SALL2 , MXI1 , and IRF1 were used . Assays were normalized to GAPDH levels , and relative abundance was calculated using a delta–delta threshold analysis as previously described [39] . The assay identification numbers for the Taqman probes are in the Accession Numbers list in the Supporting Information section of this paper .
The LocusLink ( http://www . ncbi . nlm . nih . gov/entrez/query . fcgi ? db=gene ) and Unigene ID ( http://www . sgn . cornell . edu/bulk/input . pl ? mode=unigene ) of genes discussed in this manuscript are listed in Table S1 . The National Center for Biotechnology ( NCBI ) Probe Database ( http://www . ncbi . nlm . nih . gov/entrez/query . fcgi ? db=probe ) accession numbers for the Taqman probes discussed in this manuscript are GAPDH , Hs99999905_m1; IRF1 , Hs00233698_m1; MXI1 , Hs00365651_m1; and SALL2 , Hs00826674_m1 . | Cells constantly sense their environment to decide whether to divide . Many genes that control the entry into cell division are known , and their excessive activation may cause cancer . In contrast , the way that cells cease to divide was thought to be a passive process , where signals for cell division gradually decay . In this study , the authors show that the decision to cease cell division and enter a state termed quiescence is also an active process . By monitoring the changes in activity over all genes , the authors identify a set of genes that respond specifically to decrements of external stimuli and ensure cessation of cell division . These genes act as brakes to prevent excessive cell division , and their inactivity is characteristic in many human cancers , particular those that progress to life threatening disease . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] | [
"cell",
"biology",
"genetics",
"and",
"genomics",
"homo",
"(human)",
"computational",
"biology"
] | 2007 | A Transcriptional Program Mediating Entry into Cellular Quiescence |
HIV-1-containing internal compartments are readily detected in images of thin sections from infected cells using conventional transmission electron microscopy , but the origin , connectivity , and 3D distribution of these compartments has remained controversial . Here , we report the 3D distribution of viruses in HIV-1-infected primary human macrophages using cryo-electron tomography and ion-abrasion scanning electron microscopy ( IA-SEM ) , a recently developed approach for nanoscale 3D imaging of whole cells . Using IA-SEM , we show the presence of an extensive network of HIV-1-containing tubular compartments in infected macrophages , with diameters of ∼150–200 nm , and lengths of up to ∼5 µm that extend to the cell surface from vesicular compartments that contain assembling HIV-1 virions . These types of surface-connected tubular compartments are not observed in T cells infected with the 29/31 KE Gag-matrix mutant where the virus is targeted to multi-vesicular bodies and released into the extracellular medium . IA-SEM imaging also allows visualization of large sheet-like structures that extend outward from the surfaces of macrophages , which may bend and fold back to allow continual creation of viral compartments and virion-lined channels . This potential mechanism for efficient virus trafficking between the cell surface and interior may represent a subversion of pre-existing vesicular machinery for antigen capture , processing , sequestration , and presentation .
Conventional transmission electron microscopy is a powerful tool for investigation of HIV pathogenesis and subcellular organization of viral compartments in infected cells [1] . Insights into cellular ultrastructure using this approach have generally been obtained using thin sections derived from fixed , plastic-embedded cells or tissues . However , the use of sectioned material provides an incomplete representation of cellular architecture , since a 200 nm section typically includes <1% of the volume of the complex mammalian cell . Thus , a compartment that appears to be completely internal in one section might be revealed as an invagination of the cell membrane in another section , while what appears to be a vesicle in one section might turn out to be a tubular membrane process in another . Because of the incompleteness of data obtained from cell sections , the probability of visualizing a full and accurate representation of the cytoarchitecture is minimal . This limitation is overcome with the use of ion abrasion scanning electron microscopy ( IA-SEM ) , a newly developed approach for 3D imaging of large mammalian cells and tissues in their entirety [2] , [3] , [4] at in-plane spatial resolutions of ∼6 nm and z-axis resolutions of ∼30 nm . The functional importance of obtaining cellular images that go beyond the information obtained from a thin section is illustrated by recent controversy surrounding the biogenesis of human immunodeficiency virus type I ( HIV-1 ) in macrophages . Thus , contradictory inferences about the connectivity and surface accessibility of viral compartments have been drawn using light microscopy [5] , [6] and electron microscopy of sections from cells treated with a membrane-impermeant dye [7] . Macrophages are a persistent and long-lived reservoir of provirus and infectious virions in HIV-1 infection [8] , [9] , [10] , [11] , [12] . They can survive for weeks while being productively infected , and during this time may repeatedly mediate directed , efficient infection of T cells [11] , [13] . Previous reports have suggested that , in macrophages , HIV-1 virions assemble either in late endosomes [14] , at the plasma membrane [6] , in internal compartments that communicate with the plasma membrane and bear markers of late endosomes [7] , [15] , or at a combination of these sites [5] . Release of infectious virions is not affected when late endosome motility is disabled [6]; yet , internally sequestered HIV-1 is rapidly directed to newly-formed synapses with T cells [5] , suggesting that movement of virions from internal compartments to the plasma membrane plays a role in directed infection . Since IA-SEM should allow determination of the location and distribution of viruses and viral compartments through the depth of the cell , we applied this method to definitively resolve the controversy about the precise localization and distribution of HIV-1 in infected macrophages .
Resin blocks from standard flat embedding moulds ( EMS , Hatfield , PA ) were trimmed to a pyramidal shape using a razor blade with block faces typically of about 2 mm2 in area . The surface was smoothened by sectioning using a conventional 45° diamond knife from Diatome ( distributed by EMS , Hatfield , PA ) . The entire pyramidal block was removed and mounted with the wider base onto an SEM stub using silver paint ( SPI Supplies , West Chester , PA ) such that the ultramicrotome-prepared , flat surface of the resin block pointed upwards , perpendicular to the electron column . For room temperature experiments , images were recorded using a Nova 200 NanoLab dual beam instrument ( FEI , Hillsboro , OR ) equipped with a gallium ion source for focused ion beam milling and a field emission gun scanning electron microscope with an in-lens secondary electron detector for imaging . Prior to milling and SEM imaging , the entire sample surface was coated with a platinum/palladium layer ( ∼1 µm thickness ) using the gas injector system ( GIS ) in the main specimen chamber . The specimen stage was tilted to 52° and exposed to the focused ion beam ( such that the plane of the stage was parallel to the ion beam ) . A cross sectional cut was introduced in two stages . First , a coarse cut was made at high beam currents ( typically 7–20 nA ) and at an accelerating voltage of 30 kV to create a trench that enabled viewing of the cross-section . Usually , 50–150-µm-wide trenches were cut into the specimen . In the second step , the ion beam was scanned using a current of 3–7 nA to polish and smoothen the surface . Secondary electron SEM images were typically recorded at accelerating voltages of 3 kV , 10 , 000× magnification , and a beam current of 68–270 pA in the immersion lens mode . For slice-and-view images series , a step size of ∼15 nm was chosen for the removal of material from the specimen surface using the focused ion beam . All images are presented with inverted contrast in the figures and videos for ease of comparison to images obtained from transmission electron microscopy . Primary monocyte-derived macrophages were infected with either ( i ) vesicular stomatitis virus G glycoprotein ( VSV-G ) -pseudotyped , Env-defective HIV-1 virus stocks produced by co-transfection of 293T cells with pNL4-3/KFS/MA-TC [5] , [16] and the VSV-G expression vector pHCMV-G [17] or ( ii ) infectious HIV-1 BaL . We used Env-defective HIV-1 viruses for most of the studies to eliminate the possible formation of an apparently internal compartment resulting from fusion of two previously separate cells . The electron tomographic analyses with fixed , embedded cells were carried out with infectious HIV-1 BaL . Three days post-infection , the infected macrophages were harvested and processed for electron tomographic and IA-SEM experiments . Jurkat T cells were cultured in RPMI-1640 medium supplemented with 10% FBS and infected for 4–5 hrs with NL4-3/29/31KE virions pseudotyped with VSV-G . Infected cells were harvested 24–48 hrs post-infection and processed for electron microscopy . HIV-1-expressing macrophages were prepared for imaging by adding freshly prepared 2X fixative buffer ( 5% glutaraldehyde in 0 . 2 M sodium cacodylate ) to the cell culture medium in a 1∶1 ratio immediately after pre-warming it to the cell culture temperature , followed by incubation at this temperature for 30 min . The 2X fixative buffer was then replaced by fresh 1X fixative buffer ( 2 . 5% glutaraldehyde , 0 . 1 M sodium cacodylate ) , followed by incubation at room temperature for 15 min . The cells were centrifuged at 12 , 000×g and the pellet trimmed into small blocks ( ∼2 mm square ) with an acetone-cleaned razor blade . The samples were transferred into glass scintillation vials containing 0 . 1 M sodium cacodylate buffer , rinsed 3 times with the buffer for 10 minutes each , and post-fixed with 1% osmium tetroxide in 0 . 1 M sodium cacodylate buffer ( OsO4 ) for 1 hr , followed by 2 more washes in 0 . 1 M cacodylate buffer for 10 minutes each . The samples were washed in a cold sodium acetate buffer ( 0 . 1 M ) once and stained en bloc with 0 . 5% uranyl acetate in 0 . 1 M acetate buffer for 1 hr , followed by 3 more washes in acetate buffer . The samples were dehydrated through graded ethyl alcohol followed by propylene oxide . Samples were then infiltrated overnight at room temperature with a 1∶1 mixture of Epoxy resin: propylene oxide , embedded in Embed 812 ( Electron Microscopy Sciences , Inc . ) and cured for 48 hrs in a 55°C oven . Monocyte-derived macrophages ( MDM ) from healthy donors were grown in RPMI-1640 media supplemented with 10% fetal calf serum . The cells were plated on gold Quantifoil grids ( Quantifoil Micro Tools GmbH , Germany ) and infected with VSV-G pseudotyped HIV-NL4-3/MA-TC virus for 4 hrs ( 106 cells+105 RT cpm virus ) . The cells were washed gently and incubated in media at 37°C for a further 4 days , with a change in culture medium after 2 days . They were then fixed overnight in 2 . 5% glutaraldehyde and rinsed with PBS . After deposition of 15 nm-sized gold fiducials on the grid , the cells were rapidly frozen by plunging the grid into liquid ethane maintained at ∼−180°C using a Vitrobot device ( FEI Company , Oregon ) . The grids were imaged at liquid nitrogen temperatures on a Titan Krios electron microscope ( FEI Company , Oregon ) equipped with a Gatan 2002 energy filter and operated at 200 kV . Low dose tomographic tilt series were collected over a tilt range spanning ±65° in 1 . 5° intervals with a total dose of ∼75 e−/Å2 , with an applied defocus of −15 µm , and an effective pixel size of 1 . 9 nm at the specimen plane . Tomograms were reconstructed using the software package IMOD [18] , [19] . MDM that had been infected for 7 days with the primary isolate strain HIV-1 BaL were provided by Tracy Hartman and Robert Buckheit ( Imquest Biosciences , Frederick , MD ) , and prepared for sectioning with the same procedure used for IA-SEM above . Samples were cut into 90–100 nm-thick sections using a Leica Ultracut T Microtome ( Leica Microsystems , Vienna , Austria ) , placed on carbon-coated copper EM grids , stained with lead citrate , coated with 10 nm gold beads ( Nanoprobes , Yaphank , NY ) , and imaged in a Tecnai 12 transmission electron microscope at 120 kV , 52 , 000× magnification ( image pixel size = 0 . 44 nm ) , and applied defocus of −1 µm . Tilt series were recorded using the Xplore3D software ( FEI , Netherlands ) , and the Saxton tilt scheme with an initial tilt increment of 2° . Data were collected either as single axis tilt series and reconstructed using weighted back-projection ( for fixed , embedded sections ) , or as dual axis tilt series which were aligned , reconstructed , and merged into dual-axis tomograms ( data in Video S1 ) . Reconstructions were carried out in the environment of the software package IMOD [18] , [19] . Image stacks were aligned using the Inspect3D software package ( FEI , Netherlands ) . Individual virons were automatically identified with a spherical Hough transform . Segmentation of the cell was accomplished with a combination of region growing and level-set methods found in the Insight Toolkit ( http://www . itk . org ) . These results were then verified and refined with manual classification , specifically for the details in the virion channels , with Slicer3D ( http://www . slicer . org ) . The resulting models were then rendered with 3ds MAX software using Brazil , a rendering plug-in .
We first present cryo-electron tomographic studies of primary MDM infected with HIV-1 . Projection images of whole cells on Quantifoil grids show that useful images can only be recorded at the very outer edges of the cells ( Figure 1A ) , where long filopodial extensions can be seen to emanate from the cell surface ( Figure 1B ) . Tomograms recorded from these filopodial extensions show the remarkable proximity and association of released virions to the actin-rich filopodial structures ( Figure 1B ) even at regions many microns away from the main body of the cell . Complementary information on subcellular architecture was obtained from electron tomography of thin sections from fixed , plastic-embedded HIV-1 infected macrophages , which provides a view of the interior of the cell ( Figure 2 and Video S1 ) . Inspection of the tomographic data from the interior of the cell shows clear examples of budding , immature and mature virions inside the compartment , but does not provide the spatial connections , if any , between these internal compartments and the surface of the cell . Since cryo-electron tomography of intact cells is limited to visualizing the thin edges of the cell , and conventional electron tomography is limited to imaging thin sections , we used IA-SEM to carry out 3D imaging of the infected macrophages through large segments of the central , thicker regions of the cell ( Videos S2 , S3 , S4 , S5 , S6 ) . We used Env-defective HIV-1 virions for these experiments in order to eliminate the possible formation of an apparently internal compartment resulting from fusion of two previously separate cells . Inspection of a 3D image stack shows that regions from which viruses are released at the plasma membrane are connected to larger viral compartments deeper in the cell ( Figures 3 and 4 , Video S3 ) . These internal compartments contain immature , mature , and budding HIV-1 virions ( Figure 3B ) , establishing that they are sites of virogenesis . Sites of viral budding were also observed at the cell surface associated with electron-dense plasma membrane patches ( Video S5 ) , as were immature and mature virions , confirming that HIV-1 is not only assembled in internal compartments in primary MDM , but can be simultaneously assembled at the plasma membrane . Surprisingly , IA-SEM imaging revealed highly structured tubules in the cytoplasm ( these can be seen in Figure 4 , but are especially clear in Video S6 , in which a long HIV-1 containing tubular structure can be visualized starting from the top portion of the stack ) . These tubules , which have a roughly uniform diameter of ∼150–200 nm , and lengths of up to 5 µm , project from the virion-containing compartments , occasionally connecting them to the cell surface ( Figure 4 , Supplementary legends , Videos S2 , S3 , S6 ) . The presence of a chain of HIV-1 virions within these tubular compartments suggests that they permit movement of virions between the cell surface and the intracellular compartments , and we therefore refer to them as “virion channels” . These tubular structures differ markedly from the narrow ( <20 nm ) , opposed macrophage membrane sheets reported previously [15] which were not associated with virions . Some of the deeper compartments well inside the interior of the cell do not display detectable cell surface communication ( Figure 4E , Video S4 , with animation in Video S7 ) , potentially representing multivesicular body compartments . Since the studies we report here use cells fixed with glutaraldehyde , one possible concern is that the use of the fixative may alter the shapes of internal compartments because of local changes in osmolarity in the fixation medium . Previous electron tomographic studies of fixed , stained B-cells have suggested that the use of glutaraldehyde did not alter the overall shape of lysosomes , but led to some shrinkage and deformation of early and late endosomes [20] . Although it is difficult to imagine how an entire system of virion channels with connections to the cell surface can be created instantaneously by local deformations , we evaluated this possibility by carrying out control experiments with Jurkat T cells , taking advantage of the recent demonstration by Joshi et al [21] that the HIV-1 29/31 KE Gag matrix mutant is targeted to multivesicular bodies in Jurkat T cells . Both transmission electron microscopic ( Figure 5A ) as well as IA-SEM ( Figure 5B , Video S8 , S9 ) analyses reveal that the MVB compartments in these cells are closed and display no evidence of virion channels that are connected to the cell surface . These experiments show that the virion channels are a specialized feature of macrophages and possibly other antigen-presenting cells , not present in all HIV-infected cell types , and unlikely to be an artifact of glutaraldehyde fixation . IA-SEM studies also reveal new insights into the membrane organization on the surfaces of macrophages , which can have extended membranous features ( Figure 6A; Video S10 ) . 3D imaging through these regions shows that some membrane processes that might be identified as “filopodia” in cross section ( using transmission electron microscopy of thin sections ) actually correspond to massive wavelike projections that would be far more efficient at particle capture than filopodia ( Figures 6B , 6C ) . These membrane extensions likely correspond to the “ruffles” or “veils” previously observed by SEM imaging of the MDM cell surface [22] , [23] , [24] , [25] .
Serial sectioning is a method that has been used for decades to attempt 3D reconstruction by manually combining individual 2D images . This is a tedious approach even when it does work , because it requires identification of the same cell in each successive serial section , and it also only works when there is no distortion in any of the sections at the location of the particular cell of interest . In a few rare instances , when the same cell can be located in a series of sections it is possible to attempt tomography of each section , but it still does not solve the problem of material that is lost between successive sections . The use of IA-SEM circumvents these problems , and as demonstrated in this work and in recent publications from our laboratory , allows site-specific 3D reconstruction of cells and tissues at resolutions of ∼30 nm in the z-direction and 3–6 nm in the x-y plane . Simultaneous observation of concentrated budding sites in completely internal compartments , compartments that communicate with the plasma membrane , and the plasma membrane itself is consistent with a model wherein virogenesis can be initiated at any of these sites . The observation of structured virion channels provides an explanation for productive virion release in the absence of late endosome motility that does not necessarily invoke plasma-membrane-only budding [6] . More importantly , this virion channeling system may imply a mechanism behind the efficient , directed transmission of HIV-1 from macrophages to uninfected cells [5] . This hypothetical mechanism would allow HIV-1 to remain sequestered until such contact . The virion channeling system would also allow directed release of virions without gross movement of entire endocytic compartments . Furthermore , inherent to the channeling system is the possibility that virions from a single endocytic compartment could be simultaneously directed to multiple cells . Since the channeling model does not invoke fusion of the endocytic compartment membrane with the plasma membrane , jettison of the entire compartment of stable [11] infectious virions need not occur for each synapse . The virion channeling system described here thus suggests how controlled sequestration and directed propagation of HIV-1 from infected macrophage reservoirs may be achieved . Tubulovesicular membrane systems have been observed in numerous mammalian cell types , ranging from oxyntic cells [26] , where they function to rapidly traffic secretory proteins to the cell surface , to neurons , where specific proteins involved in synaptic vesicle trafficking have been implicated in the formation of tubules similar to those reported here [27] . The reorganization of multivesicular bodies or lysosomes into tubular structures within antigen presenting cells has been reported previously , primarily in conjunction with the traffic of MHC class II and related molecules to the plasma membrane of these cells [28] , [29] . It is possible that these are related to the tubular structures described in this study . It will also be important to compare the 3D structures of macrophage viral compartments with those in other antigen-presenting cells such as dendritic cells which are either infected or exposed to HIV to get a better appreciation of the generality of this phenomenon , which may be a viral subversion of an existing membrane system that normally would channel , sequester , or deliver cellular proteins or foreign antigens . The discovery of narrow virion channels in macrophages highlights the importance of methodologies for 3D imaging of whole cells at resolutions where all cellular components can be visualized at nanometer resolution . The combination of IA-SEM and semi-automated image segmentation such as that used in this work and elsewhere [30] provide powerful tools to explore the architectures of complex eukaryotic cells to simultaneously uncover details ranging from the internal structures of viruses to the organization of compartments inside the cell . | Current treatment regimens for HIV-infected individuals are not capable of eradicating HIV infection , even though combinations of highly potent antiviral drugs are used . Indeed , drug regimens must be periodically altered as the virus resurges from a persistent reservoir . Macrophages , which serve as “search-and-destroy” immune surveillance cells of the body , are now thought to be a key component of this reservoir . Evidence suggests that macrophages can harbor infectious HIV virions for long periods , and transmit them to bystander T cells . We have used a new technique called ion abrasion scanning electron microscopy ( IA-SEM ) to image entire HIV-infected human macrophages at a resolution high enough to see individual HIV virions and their location within the cell . This approach revealed that HIV is present in a system of nanoscale tubes , barely larger than a virus at some places , which connect internal viral reservoirs to the cell surface . These tubes could allow the macrophage to deliver HIV virions to bystander cells from its continually replenished stores of ammunition , held deep within the cell . Our work provides a glimpse of how the structure of these reservoirs allows macrophages to accomplish viral delivery . Discovery of these virion-channeling tubes provides a potential drug target to address the problem of persistent HIV infection . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"virology/vaccines",
"virology/virion",
"structure,",
"assembly,",
"and",
"egress",
"cell",
"biology/membranes",
"and",
"sorting",
"infectious",
"diseases/hiv",
"infection",
"and",
"aids",
"infectious",
"diseases/viral",
"infections"
] | 2009 | Ion-Abrasion Scanning Electron Microscopy Reveals Surface-Connected Tubular Conduits in HIV-Infected Macrophages |
After many years of neglect , schistosomiasis control is going to scale . The strategy of choice is preventive chemotherapy , that is the repeated large-scale administration of praziquantel ( a safe and highly efficacious drug ) to at-risk populations . The frequency of praziquantel administration is based on endemicity , which usually is defined by prevalence data summarized at an arbitrarily chosen administrative level . For an ensemble of 29 West and East African countries , we determined the annualized praziquantel treatment needs for the school-aged population , adhering to World Health Organization guidelines . Different administrative levels of prevalence aggregation were considered; country , province , district , and pixel level . Previously published results on spatially explicit schistosomiasis risk in the selected countries were employed to classify each area into distinct endemicity classes that govern the frequency of praziquantel administration . Estimates of infection prevalence adjusted for the school-aged population in 2010 revealed that most countries are classified as moderately endemic for schistosomiasis ( prevalence 10–50% ) , while four countries ( i . e . , Ghana , Liberia , Mozambique , and Sierra Leone ) are highly endemic ( >50% ) . Overall , 72 . 7 million annualized praziquantel treatments ( 50% confidence interval ( CI ) : 68 . 8–100 . 7 million ) are required for the school-aged population if country-level schistosomiasis prevalence estimates are considered , and 81 . 5 million treatments ( 50% CI: 67 . 3–107 . 5 million ) if estimation is based on a more refined spatial scale at the provincial level . Praziquantel treatment needs may be over- or underestimated depending on the level of spatial aggregation . The distribution of schistosomiasis in Ethiopia , Liberia , Mauritania , Uganda , and Zambia is rather uniform , and hence country-level risk estimates are sufficient to calculate treatment needs . On the other hand , countries like Burkina Faso , Mali , Mozambique , Sudan , and Tanzania show large spatial heterogeneity in schistosomiasis risk , which should be taken into account for calculating treatment requirements .
Schistosomiasis is a snail-borne parasitic disease , which affects more than 200 million people globally based on 2003 population estimates [1] . The main burden is concentrated in sub-Saharan Africa , where schistosomiasis control had been neglected for many decades , and hence country-specific infection rates above 25% are the norm rather than the exception [1]–[3] . The global strategy for the control of schistosomiasis , as recommended by the World Health Organization ( WHO ) , is preventive chemotherapy , that is the repeated , large-scale administration of the antischistosomal drug praziquantel to at-risk populations [4] . Because school-aged children are at highest risk of schistosome infection , preventive chemotherapy is primarily targeted at this population group . The frequency of treatment depends on the level of endemicity , defined as low ( prevalence <10% ) , moderate ( 10–50% ) , and high ( >50% ) based on WHO thresholds [4] , [5] . Classification into one of these categories should be done according to schistosomiasis prevalence estimates among the school-aged population . In practice , estimates of schistosomiasis and other helminth infection are often based on old survey data and/or very sparse geographical data . However , historical data are largely obsolete due to ongoing control efforts , ecological transformations , demographic changes , and improved hygiene , among other reasons , while sparse geographical data are not representative for large areas [1] , [6]–[9] . In addition , most prevalence estimates lack empirical modeling , which is mandatory to obtain accurate averages of the risk of infection over large geographical regions , taking into account spatial heterogeneity . Even though WHO treatment guidelines are valuable tools to plan and conduct preventive chemotherapy programs , it remains unclear at which geographical scale or other stratifications ( e . g . , ecological zones ) the level of endemicity should be defined to avoid unnecessary treatments , increase cost-effectiveness , and meet local needs . The aim of this paper was to assess the effect of the geographical scale of schistosomiasis risk estimates on the amount of praziquantel treatment needed in the school-aged population in selected parts of Africa . In contrast to existing treatment need calculations based on crude country prevalence estimates [10] , we used spatially explicit , model-based schistosomiasis risk estimates at high spatial resolution [2] , [3] . Our modeling framework includes uncertainty measures , and hence allows confidence intervals ( CIs ) to be determined around predicted treatment needs . These empirical estimates are able to capture potential small transmission hotspots within a country that are typical for schistosomiasis , as the disease is often focally distributed [11] , [12] , and facilitate spatial aggregation of schistosomiasis risk at different geographical scales . In our analysis , we focus on differences in the total amount of required treatments when risk estimates are known at country , regional , district , or pixel level .
In our previous work , we determined the spatial distribution of Schistosoma haematobium and S . mansoni in an ensemble of 29 West and East African countries at a spatial resolution of 5×5 km [2] , [3] . We used a geostatistical modeling approach , applied to survey data extracted from the Global Neglected Tropical Disease database ( http://www . gntd . org ) [13] . Statistical relations of Schistosoma prevalence with climatic and other environmental factors were used to obtain model-based predictions at unobserved locations using country-level joint posterior distributions . The median and 50% CIs of the joint posterior distributions were utilized for further analyses . The risk predictions of Schur and colleagues were originally summarized as country-specific Schistosoma infection prevalence estimates adjusted for individuals aged ≤20 years in 2010 [2] , [3] . For the current analysis , these data were re-adjusted to the school-aged population ( children aged between 5 and 14 years ) , in line with the WHO definition of the school-aged child [14] . Population count data at 1×1 km spatial resolution in Africa for 2008 were obtained from the LandScan global population database ( http://www . ornl . gov/landscan/ ) . Country-specific rates of annual change in population during the period of 2005–2010 were extracted from the United Nations World Population Prospects [15] . Estimates on the percentage of children aged between 5 and 14 years among the total population per country were obtained from the Population Division , provided by the international database of the United States Census Bureau [16] . Population estimates of 2008 were projected to 2010 taking into account estimates of annual population change , readily adjusted to school-aged children . These estimates were then converted to a 5×5 km resolution grid using the geographical information system ( GIS ) software ArcMap version 9 . 2 ( ESRI ) to match the spatial resolution of the geostatistical schistosomiasis risk estimates for West and East Africa . National preventive chemotherapy programs are usually implemented at specific administrative levels , for example at country , province , and district level . Shape files containing geographical information on the administrative boundaries of these levels were downloaded from the Map Library ( http://www . maplibrary . org ) for our ensemble of 29 West and East African countries . These files were linked in ArcMap with schistosomiasis risk estimates and school-aged population to estimate the number of infected children at pixel-level . Population-adjusted prevalence at a given administrative region was calculated by summing all infected school-aged children at pixel-level and dividing by the total school-aged population in that region . This calculation takes into account that pixels are not equally populated , and hence do contribute to the area-specific prevalence in the same way . The aggregated and pixel-level data on the school-aged children and population-adjusted prevalence were converted to the required amount of drugs per year ( “annualized treatment needs” ) using WHO schistosomiasis control guidelines [4] , [5] . These guidelines suggest different treatment strategies according to endemicity thresholds . High endemicity is defined by prevalence of at least 50% among school-aged children using parasitological diagnostic tests . Low and moderate endemicity settings are classified by prevalence levels of <10% and 10–50% , respectively . For high endemicity areas , annual treatment of all school-aged children ( and other high-risk groups , e . g . , fishermen ) is proposed . With regard to treatment needs in the school-aged population , we therefore estimated the amount of annualized praziquantel treatment needs in these areas to be equal to the number of school-aged children . In areas with moderate endemicity , it is recommended to treat school-aged children every other year . Hence , we consider half of the school-aged population for annual praziquantel treatment in our calculation of annualized needs . Areas of low endemicity warrant treatment of school-aged children twice during primary schooling ( on entry and just before leaving school ) . Assuming an average duration of 6 years in primary school , we estimate annualized praziquantel needs by considering one third of the school-aged population for treatment every year .
Country-specific estimates on the prevalence of schistosomiasis among the school-aged population and the number of infected school-aged children are summarized in Table 1 . The population-adjusted prevalence estimates revealed that 30 . 2 and 19 . 5 million school-aged children in West and East Africa , respectively , are infected with S . haematobium , S . mansoni , or both species concurrently . At the unit of the country , we found infection prevalence ranging between 14 . 4% ( The Gambia ) and 60 . 2% ( Liberia ) with a mean country-prevalence of 37 . 3% . Most countries were classified as moderately endemic ( prevalence 10–50% ) , whereas four countries ( i . e . , Ghana , Liberia , Mozambique , and Sierra Leone ) were considered as highly endemic ( >50% ) . Figure 1A displays the schistosomiasis endemicity aggregated at country-level and Figure 1B the spatial distribution of schistosomiasis risk over Africa based on current WHO endemicity thresholds . High- and low-risk areas are found in West and East Africa . Most low-risk areas are located in Cameroon , Saharan Mali , Mauritania , Niger , and parts of Ethiopia , Kenya , Uganda , and Zambia . Large high-risk regions are found along the Niger River , in Ghana , Liberia , Mozambique , and Sierra Leone . Table 1 summarizes the estimated amount of annualized praziquantel treatment needs , stratified by country , based on different administrative levels of risk estimates , i . e . , country , province , district , and pixel level , using WHO thresholds . Table 2 shows the respective 50% CIs . Overall , in West and East Africa , 72 . 7 million treatments ( 50% CI: 68 . 8–100 . 7 million ) are required if calculations are based on schistosomiasis risk estimates at country level . An additional 8 . 8 million treatments ( resulting in a total of 81 . 5 million treatments , 50% CI: 67 . 3–107 . 5 million ) are needed if the calculation is based on province-level risk estimates . Calculations employing risk estimates at higher level of disaggregation – at district or pixel level – result in a total amount of 82 . 7 million ( 50% CI: 66 . 0–111 . 1 million ) and 83 . 0 million ( 50% CI: 61 . 0–119 . 8 million ) treatments , respectively . Over 5 million annual praziquantel treatments for the school-aged population are required in Ethiopia , Ghana , and Nigeria , while less than half a million treatments are needed for preventive chemotherapy in Djibouti , Eritrea , The Gambia , Guinea-Bissau , and Mauritania . These calculations are based on country-wide risk estimates and they vary when risk estimates are available at higher level of disaggregation . Considering treatment in entire communities in high endemicity areas and 20% of the non-school-aged population in moderately endemic areas ( as recently considered by WHO [17] ) , the required annual amount of praziquantel increases to more than 170 million or almost 235 million treatments using country- or pixel-level risk estimates , respectively ( results not presented in Table 1 ) . Next , selected country examples are given regarding annualized praziquantel treatment needs in the school-aged population . These examples highlight that the required amount of praziquantel depends on the level of aggregation of the schistosomiasis risk estimates , and is particularly pronounced in countries where the distribution of schistosomiasis shows high focality . There are ongoing discussions within WHO and other organizations and consortia whether the threshold to separate between moderate and heavy schistosomiasis endemicity should be halved from 50% to 25% . Of note , the lower threshold is already adopted by the Schistosomiasis Consortium for Operational Research and Evaluation ( SCORE ) in the frame of large-scale programs that aim at gaining ( prevalence ≥25% ) and sustaining ( prevalence 10–24% ) schistosomiasis control ( http://score . uga . edu/Projects . html ) . This new 25% threshold would imply that more than 120 million annualized praziquantel treatments ( 50% CI: 122 . 4–131 . 5 million ) would be needed in the 29 African countries for the school-aged population alone when considering country-level prevalence estimates , and around 112 million treatments based on aggregations at other administrative levels ( see Tables 1 and 2 ) . A change in the threshold would not affect all countries in the same way , e . g . , the estimated total praziquantel treatment needs in The Gambia ( with an averaged population-adjusted prevalence of 14 . 4% ) , Ghana ( 54 . 6% ) , Mozambique ( 58 . 5% ) , Sierra Leone ( 59 . 3% ) , and Liberia ( 60 . 2% ) would remain almost the same , irrespective of the level of aggregation , while it would increases by more than 50% in Rwanda ( 33 . 5% ) , Sudan ( 34 . 7% ) , Burundi ( 38 . 1% ) , Somalia ( 38 . 7% ) , and Malawi ( 47 . 4% ) based on prevalence aggregations at district-level .
Preventive chemotherapy is the current mainstay for morbidity control due to schistosomiasis and other helmintic diseases [4] , [9] , [14] , [18]–[20] . The administration of praziquantel to high-risk communities , most importantly school-aged children , is often based on crude Schistosoma prevalence estimates that are aggregated at an arbitrarily chosen administrative level . Our analysis shows that the geographical scale at which schistosomiasis prevalence data are aggregated affects the estimated amount of treatment needs , and hence has to be considered in the calculation of the number of praziquantel treatments required for national schistosomiasis control programs . Contrary to what we expected , the total amount of praziquantel treatments is higher at a smaller administrative unit of schistosomiasis prevalence aggregation ( i . e . , district level versus province or country level ) . For example , we estimate that the annualized praziquantel treatment needs for the school-aged population are more than 10% higher in the ensemble of 29 West and East African countries considered here when calculated using district instead of country level prevalence estimates . This observation can be explained because many East and West African countries had a prevalence of schistosomiasis in school-aged children at around 40% ( moderately endemic situation ) , and the infection risk is not uniformly distributed throughout a country . Indeed , some province or district specific prevalence estimates were above the 50% threshold suggested by WHO , and hence warrant annual treatment rather than treatment every second year for moderate endemicity areas ( prevalence 10–50% ) . We rarely observed switches from moderate to low endemicity levels . Such switches are expected to occur in countries where the overall prevalence of Schistosoma infection is close to 10% . Such low prevalence levels were only found in The Gambia ( 14 . 4% ) and Senegal ( 18 . 2% ) . In countries currently classified as highly endemic , the required amount of treatment per year will be reduced if schistosomiasis prevalence is calculated at province or even district level , as some areas will fall under the 50% threshold . Utzinger et al . ( 2009 ) [10] estimated that 91 . 3 million praziquantel treatments are needed for the school-aged population based on risk estimates aggregated at country level with 2006 population data for the 29 East and West African countries considered here . This is almost 20 million more than the 72 . 7 million treatments estimated in the current study . What are the reasons for this difference ? Most importantly , there are marked differences in the country prevalence estimates of schistosomiasis leading to different classifications of endemicity ( e . g . , Burkina Faso , Mali , and Tanzania ) . Utzinger and colleagues employed prevalence estimates published in Steinmann et al . ( 2006 ) [1] , which are to a large extent based on data obtained before 1989 . However , due to major demographic and ecological transformations ( e . g . , urbanization ) , social and economic development , and implementation of schistosomiasis control programs , country prevalence estimates have changed ( mostly to lower levels ) . Moreover , different sources of population data and the estimated percentage of school-aged children have some leverage on the estimated annualized praziquantel treatment needs . In early 2012 , WHO released new estimates on the total number of individuals requiring preventive chemotherapy with praziquantel [17] . For Africa ( using WHO country classifications; e . g . Sudan being part of the Eastern Mediterranean rather than Africa ) , it is estimated that as many as 102 million treatments are needed for the school-aged population and 221 million treatments when including communities in high and moderately endemic areas . Of note , secondary administrative level schistosomiasis risk estimates were employed to define endemicity levels . The bulk of these risk estimates lacks geostatistical modeling and neglects potential spatial correlation inherent to the data . Moreover , in the absence of recent data , information was also obtained from surveys conducted before 1989 without accounting for changes in prevalence levels over time . Compared to the WHO treatment needs calculations for the school-aged population , our estimates are by approximately 30 million annual treatments smaller , but we used a different set of African countries in the current analysis . For the missing African countries we used the previously published country-level prevalence estimates [10] to calculate the total treatment needs in the school-aged population . This assumption would result in an annualized amount of 105 . 9 million praziquantel treatments per year in school-aged children . Interestingly , this estimate is very close to the 102 million put forth by WHO [17] . Several issues warrant discussion regarding our geostatistical model-based analysis and its implication for preventive chemotherapy . First , human , financial , and technical resources for the control of neglected tropical diseases , including schistosomiasis , have increased considerably in recent years [21] , [22] . New initiatives and partnerships aim to bring praziquantel distribution to scale , such as the Schistosomiasis Control Initiative ( SCI ) [19] and the Global Network for Neglected Tropical Diseases ( http://www . globalnetwork . org/ ) [10] . It is therefore conceivable that the overall schistosomiasis prevalence in some areas might have dropped from high to moderate , and from moderate to low endemicity thresholds , which would result in fewer treatments required . However , rapid re-infection after de-worming is a common phenomenon , which challenges the control of schistosomiasis [23] and other helminth infections [24] , [25] , and hence has to be considered when reducing the frequency of treatment campaigns . Some experts are currently discussing to lower the threshold distinguishing between moderate and high endemicity from 50% to 25% to further reduce morbidity and this strategy might also have an impact on transmission . However , such a move would strongly increase the required amount of annualized treatment needs for the school-aged population from 72 . 7 million to more than 120 million for the 29 West and East African countries considered here . Our praziquantel treatment needs were calculated based on schistosomiasis risk estimates summarized at different scales . Administrative boundaries were used because large-scale preventive chemotherapy programs are usually planned and conducted at specific administrative levels ( e . g . , district ) . However , given the strong focality of schistosomiasis , there is strong heterogeneity within administrative boundaries . For instance , villages in close proximity to freshwater bodies , where intermediate host snails proliferate , are often highly endemic for schistosomiasis , while villages farther away ( but in the same administrative region ) might have prevalence levels below 10% , even though the overall endemicity in the considered administrative unit is moderate . Hence , stratification of areas should preferably be based on the geography and suitable freshwater bodies ( e . g . , ecozones ) , which is the recommended strategy by WHO [4] . However , many schistosomiasis data are not ( yet ) available at geographically stratified locations , but are instead aggregated over administrative levels . This makes it difficult for local health managers to implement preventive chemotherapy programs considering the environment . In addition , stratification of the area according to the environment is more complex since multiple ecological factors may interact . Large parts of Africa are co-endemic for several neglected tropical diseases [26]–[29] . This is explained by similar pathways of infection and lack of preventive measures [30] , [31] . Hence , efforts should be made to integrate control programs so that multiple neglected tropical diseases can be addressed in order to reduce co-morbidity and costs . The morbidity of up to seven neglected tropical diseases might be addressed by preventive chemotherapy with four drugs , which can be safely co-administered based on the current state of knowledge [18] , [32] , [33] . However , with decreasing level of co-infection risk due to reduced prevalence , co-administration of multiple drugs becomes less cost-effective and many individuals will be treated unnecessarily with multiple drugs [34] , [35] . This potentially increases the risk of drug resistance development . Considering that , on average , three tablets of praziquantel are needed to treat a school-aged child for schistosomiasis , more than 210 million tablets would need to be administered every year to this population group in the 29 East and West African countries considered here . The distribution of such large amounts of praziquantel is a formidable challenge , especially in remote rural areas . It will require strong health systems and good supply strategies , which are currently lacking in large parts of Africa . Even though the price of a single tablet of praziquantel is now below US$ 0 . 10 [36] , [37] and pharmaceutical companies have pledged new donations , the costs for drug storage and distribution are considerable . Additionally , it might be hard to achieve high compliance after several rounds of preventive chemotherapy , because morbidity is likely to be reduced and individuals might refuse to take the drugs when feeling healthy . With the transition from morbidity control to transmission control of schistosomiasis and other helminthic diseases , preventive chemotherapy alone is unlikely to achieve this goal [38] , [39] . Complementary measures will be required depending on the setting and available resources . For instance , the “Tokomeza Kichocho” project is aiming to eliminate schistosomiasis from Zanzibar based on a multi-pronged approach , i . e . , preventive chemotherapy targeting school-aged children , intermediate host snail control , improved access to sanitation , water and hygiene , and formative health education to change human behavior . In conclusion , the geographical scale to estimate schistosomiasis risk should take into consideration the expected country-specific prevalence level and the diversity of schistosomiasis risk in the country in order to accurately calculate praziquantel treatment needs . In countries with rather homogeneous schistosomiasis distribution , it would be sufficient to implement control strategies based on schistosomiasis prevalence aggregated at country level . However , in countries with a highly focal distribution of schistosomiasis , province- or even district-level calculations of prevalence have to be considered in order to provide sufficient amounts of treatments to those at risk of morbidity and to reduce the costs in areas requiring less frequent rounds of preventive chemotherapy . The work reported here can be readily adapted to other neglected tropical diseases that have also endorsed preventive chemotherapy as the strategy of choice for morbidity control . In addition , it will be interesting to repeat our analysis in the frame of ongoing large-scale control programs , which are constantly changing prevalence levels and disease , and hence treatment needs . | More than 200 million people are affected by the snailborne disease schistosomiasis . The main strategy to control schistosomiasis is to regularly treat school-aged children with the drug praziquantel . The frequency of praziquantel treatment depends on the average prevalence of schistosomiasis , which can be defined as low ( prevalence <10% ) , moderate ( 10–50% ) , or high ( >50% ) . However , it remains unclear at which geographical scale these prevalence levels should be considered to avoid unnecessary treatments but still comply with local needs . We investigated the effect of the geographical scale for an ensemble of 29 West and East African countries using previously published model-based schistosomiasis risk estimates obtained at high spatial resolution . These estimates allow spatial risk aggregation at different geographical scales ( i . e . , country , region , district , or pixel level ) . More than 70 million praziquantel treatments are required every year for school-aged children if countrylevel estimates are used . On a more refined geographical scale ( i . e . , province ) , annualized praziquantel treatments increase by 12% . Depending on the averaged schistosomiasis prevalence and the spatial risk variation across a country , the difference in the estimated amount of praziquantel between country-level aggregation and other geographical scales might be very important , as for example in Burkina Faso , Ghana , and Mali . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"public",
"health",
"and",
"epidemiology",
"epidemiology",
"biology",
"population",
"biology",
"spatial",
"epidemiology"
] | 2012 | Determining Treatment Needs at Different Spatial Scales Using Geostatistical Model-Based Risk Estimates of Schistosomiasis |
The origins of crop diseases are linked to domestication of plants . Most crops were domesticated centuries – even millennia – ago , thus limiting opportunity to understand the concomitant emergence of disease . Kiwifruit ( Actinidia spp . ) is an exception: domestication began in the 1930s with outbreaks of canker disease caused by P . syringae pv . actinidiae ( Psa ) first recorded in the 1980s . Based on SNP analyses of two circularized and 34 draft genomes , we show that Psa is comprised of distinct clades exhibiting negligible within-clade diversity , consistent with disease arising by independent samplings from a source population . Three clades correspond to their geographical source of isolation; a fourth , encompassing the Psa-V lineage responsible for the 2008 outbreak , is now globally distributed . Psa has an overall clonal population structure , however , genomes carry a marked signature of within-pathovar recombination . SNP analysis of Psa-V reveals hundreds of polymorphisms; however , most reside within PPHGI-1-like conjugative elements whose evolution is unlinked to the core genome . Removal of SNPs due to recombination yields an uninformative ( star-like ) phylogeny consistent with diversification of Psa-V from a single clone within the last ten years . Growth assays provide evidence of cultivar specificity , with rapid systemic movement of Psa-V in Actinidia chinensis . Genomic comparisons show a dynamic genome with evidence of positive selection on type III effectors and other candidate virulence genes . Each clade has highly varied complements of accessory genes encoding effectors and toxins with evidence of gain and loss via multiple genetic routes . Genes with orthologs in vascular pathogens were found exclusively within Psa-V . Our analyses capture a pathogen in the early stages of emergence from a predicted source population associated with wild Actinidia species . In addition to candidate genes as targets for resistance breeding programs , our findings highlight the importance of the source population as a reservoir of new disease .
Despite considerable improvements in the management of plant diseases , modern agriculture remains vulnerable to losses caused by microbial pathogens . Plant diseases conservatively account for the loss of at least 10% of annual global food production [1] . The intensive cultivation of clonally propagated plants with low genetic diversity heightens opportunities for the emergence and rapid spread of infectious disease [1]–[3] . The origins of agricultural plant diseases are unclear [3] , [4] . The earliest pathogens are likely to have evolved from commensals or pathogens colonizing wild relatives of plants selected for human domestication [5]–[10] . Given that domestication of staple crop plants took place centuries ( and often millennia ) ago , the signal of this evolutionary past – the nature of the initial pathogen population , its relationship with commensal types , its diversity and genetic structure , plus factors and processes that might have led to the first outbreaks of disease – is obscured by the passage of time . Kiwifruit ( Actinidia spp . ) is an exception . Domestication of kiwifruit is recent and clearly documented; outbreaks of disease are recorded and the pathogens responsible have been preserved . The genus Actinidia comprises 55 species and about 76 taxa native to eastern Asia , with the greatest abundance and diversity in the southwestern provinces of China [11] . Fruit have long been collected from the wild , yet commercial cultivation only began to gain momentum in the 1980s , based on the success of A . deliciosa ‘Hayward’ developed in New Zealand during the 1930s . Subsequently cultivars from the species A . chinensis , such as ‘Hort16A’ , ‘Jin Tao’ and ‘Hongyang’ were commercialized in several kiwifruit growing regions , including New Zealand , China ( 1998 ) , Italy ( 2001 ) and Chile ( 2003 ) [12] . A small group of fungal and bacterial diseases on vines , roots and fruit of ‘Hayward’ were recognized once commercial plantings became substantial in the 1980s [13] . Kiwifruit canker disease caused by Pseudomonas syringae pv . actinidiae ( Psa ) was first reported and characterized on A . deliciosa in Shizuoka , Japan in 1984 [14]; in that same year bacterial canker disease was also reported in an orchard in Hunan , China [15] . Canker disease was subsequently observed in Korea ( 1988 ) and Italy ( 1992 ) [16] , [17] . Symptoms of infection include late winter die-back of young canes , frequently accompanied by rust red exudates from canes and trunks , and the presence of necrotic lesions with chlorotic halos on leaves during the spring [18] , [19] . In 2008 an aggressive form of Psa was reported in Italy on A . chinensis . Multi-locus sequencing showed the pathogenic strains to be divergent from earlier Italian and Japanese isolates [20]–[23] ) . The virulent form of Psa was subsequently detected in neighboring European countries [24] , China , Chile and New Zealand [23] , [25] . Additional genomic analyses confirmed the clonal nature of the disease outbreak and its distinctive genetic composition [19] , [22] , [26] , [27] . Throughout the paper we refer to this recent epidemic as the “2008 outbreak” with the strains responsible being referred to collectively as “Psa-V” . Rapid transmission and increased severity of infection arising from the 2008 outbreak has had devastating effects leading to complete destruction of orchards . In 2010 , Psa-V was detected in New Zealand , where kiwifruit is the most valuable horticultural export [12] , [25] . Psa-V spread rapidly from its initial incursion site in the Bay of Plenty . Within two years the number of infected orchards rose from three to 1232 ( 37% of New Zealand orchards ) and continues to increase [28] . Draft genome sequencing of one Italian isolate from the 2008 outbreak , plus another strain from an epidemic in 1992 , showed the earlier Italian strain to be identical to a 1984 Japanese isolate , but different to the 2008 ( Psa-V ) outbreak strains , particularly with respect to effector inventories [26] . Draft genomes of four additional isolates from the 2008 Italian outbreak plus three isolates from China and two from Chile also belong to the same Psa-V lineage although one Chinese strain ( M228 ) appears divergent [22] , [27] . Strains of the Psa-V lineage were shown to differ by as few as six SNPs . A genomic island with similarity to PPHGI from Pseudomonas syringae pv . phaseolicola ( Pph ) was characterized and shown to differ between the European , Chilean and Chinese/New Zealand isolates [22] , [27] , [29] . The full extent of the genetic distinctiveness of Psa-V relative to strains from early outbreaks remains uncertain . While work to date shows that the 2008 epidemic is caused by a strain distinct from previous outbreaks of canker disease , it is not clear whether Psa-V evolved from earlier outbreaks or whether it has independent origins . Clarification can come from studies that determine patterns of nucleotide diversity from strains sampled from different disease outbreaks , at different time points , from different geographical locations [30]–[33] . From the analysis of polymorphism data it becomes possible to infer underlying population processes and , provided recombination is not extensive , phylogeny [34] , [35] . Leaving aside opportunities for insight into the evolution of a newly emergent pathogen , knowledge of population structure also has important implications for the development of strategies for disease control , including protocols to prevent future outbreaks . Here we report an in depth analysis of Psa evolution based upon complete genome sequences of a Psa-V strain from the New Zealand outbreak and the type strain ( J-35 , ICMP9617 ) originally isolated in Japan , plus 34 additional draft genomes that encompass strains from previously known outbreaks . Our analyses provide evidence of a single source population from which outbreaks of disease have arisen via independent transmission events to different kiwifruit growing regions . Overall the population is clonal , however , approximately 10% of the genome shows evidence of homologous recombination marked by gene conversion . Comparative analyses reveal dynamic genomes with positive selection affecting type III effectors and candidate virulence genes with Psa-V containing numerous genes found in pathovars of vascular plants . In planta growth data provide a link between genomic inferences and ecological performance on host cultivars . Together these data provide understanding of the processes and factors affecting the emergence of a new plant pathogen .
The 25 strains of P . syringae pv . actinidiae were isolated from Actinidia spp . in Japan , Korea , Italy and New Zealand . Details of dates and locations of isolations are in Table 1 . Bacteria were maintained on King's B ( KB ) agar plates [36] and stored at −80°C in glycerol . Where possible , cultures have been deposited in the International Collection of Micro-organisms from Plants ( ICMP; www . landcareresearch . co . nz ) . Several isolates were sourced from overseas collections , including the Korean Agricultural Culture Collection ( KACC ) , Suwon , Republic of Korea , the National Institute of Agrobiological Sciences ( NIAS ) , 2-1-2 Kannondai , Tsukuba , Ibaraki 305-8602 , Japan , the National Collection of Plant Pathogenic Bacteria ( NCPPB ) , Food and Environment Research Agency , Sand Hutton , York , UK . Cultures were grown overnight in KB or nutrient broth . DNA for whole genome sequencing was purified using the PureGene DNA Isolation kit ( Qiagen , Hilden , Germany ) with some modifications . For each 750 µL of culture extracted , reagent volumes were doubled and the protein precipitation step was carried out twice . Purified DNA was quantified and its purity assessed using a nanodrop spectrophotometer ( NanoDrop Technologies , Rockland , DE ) . Paired-end libraries were generated from genomic DNA and sequenced using the Illumina GAII instrument at the Center for Genome Analysis and Function ( CAGEF , Toronto ) . The raw data were filtered using fastq-mcf ( https://code . google . com/p/ea-utils/wiki/FastqMcf ) and quality checked using fastqc ( http://www . bioinformatics . babraham . ac . uk/projects/fastqc/ ) . Contigs were built using the SOAP de novo assembler ( http://soap . genomics . org . cn/soapdenovo . html ) with a k-mer value of 37 . Sequencing and assembly statistics for each strain are shown in Table 1 . Additional sequencing was performed on two Psa isolates , NZ V-13 ( a virulent ( Psa-V ) strain isolated during the 2010 outbreak in New Zealand ) , and the Psa type strain ( J-35 , ICMP9617 originally isolated in 1984 from Actinidia in Japan ) using a 10 kb mate-end library on the Roche 454 platform by Macrogen , Korea ( www . macrogen . com ) . Scaffolds were generated using the Roche Newbler assembler , and resulted in 14 and 12 scaffolds greater than 2 kb , for NZ V-13 and J-35 , respectively . The quality of all assemblies was improved by iteratively filling in short tracts of Ns using GapCloser ( version 1 . 12 , http://soap . genomics . org . cn/soapdenovo . html ) and scaffolding with the mate-pair data using SSPACE [37] . The two Newbler assemblies were used as a basis for generating high quality reference sequences by manually integrating the Illumina-based contigs using Geneious ( version 5 . 6 . 3 , Biomatters , http://www . geneious . com/ ) . The remaining large scaffolds were linked by designing primers 1–2 kb from the 5′ and 3′ ends and using long-template PCR ( Takara Bio inc . , Shiga , Japan ) with amplification conditions recommended by the manufacturer . The quality of the genome assemblies was visually inspected using Hagfish ( https://github . com/mfiers/hagfish/ ) to align paired-end reads back to the draft genome . Misaligned and dubious contigs were manually filtered . Additional gap-filling was performed on Psa NZ V-13 with primers flanking each gap . Products were purified and Sanger sequenced ( Macrogen , Korea , www . macrogen . com ) . Assemblies for all strains were submitted to the Prokaryotic Genomes Automatic Annotation Pipeline ( PGAAP , https://www . ncbi . nlm . nih . gov/genomes/static/Pipeline . html ) for gene prediction and annotation: they are available from GenBank ( Table 1 ) . Plasmid DNA was extracted from 250 mL cultures using an alkaline lysis with SDS extraction method ( Protocol 3: Preparation of Plasmid DNA by Alkaline Lysis with SDS: Maxipreparation ) [38] . Nutrient Broth ( 250 mL ) was inoculated with a 2 mL overnight culture and left shaking for 48 h at 28°C . Cells were harvested by centrifugation at 12 , 000 g and either used fresh or stored at −20°C until required . Cells were resuspended in 18 mL Alkaline Lysis Solution I ( 50 mM Glucose , 25 mM Tris ( pH 8 . 0 ) , 10 mM EDTA ( pH 8 . 0 ) ) , and then 2 mL freshly prepared Lysozyme solution ( 10 mg/mL ) was added . Two volumes ( 40 mL ) of freshly prepared Alkaline Lysis Solution II ( 0 . 2 M NaOH , 1% SDS ) were added , the cells mixed by gentle inversion several times , and then incubated at room temperature ( RT ) . After 5–10 min of incubation , 20 mL ice-cold Alkaline Lysis Solution III ( 5 M KAcetate 60 mL , glacial acetic acid , 11 . 5 mL H2O , 28 . 5 mL ) was added , and contents mixed by gentle but effective swirling . Samples were incubated on ice for 10 min and then centrifuged at 20 , 000 g for 30 min at 5°C . The supernatant was separated from the pellet , the volume measured , 0 . 6 vol of isopropanol was added and samples were incubated for 10 min at RT . After centrifugation at 12 , 000 g , the supernatant was discarded , and the pellet washed in 70% ethanol , drained and left to dry ( without vacuum ) for 5–10 min at RT . The pellet was resuspended in 500 µL Tris/EDTA ( pH 8 . 0 ) . Concentration of samples was achieved by further ethanol precipitation . Plasmid digests were run using a Chef-DR III system ( Bio-Rad , Hercules CA ) . Volumes of up to 20 µL were loaded onto 1 . 0% 1×TAE agarose gels and run in 1×TAE buffer for 18 h at 15°C at 6 V/cm . The switch time was 5–15 s with an angle of 120° . Gels were stained with ethidium bromide according to the protocol of Sambrook and Russell ( 2001 ) . Single nucleotide polymorphisms ( SNPs ) were identified using REALPHY ( F Bertels , P . B . Rainey , O . K . Silander and E . van Nimwegen , unpublished ) with reference to both Pto DC3000 and Psa V-13 for phylogenetic analyses between all Psa strains and within the Psa V clade , respectively . A minimum read coverage of 10 , a minimum PHRED score of 20 and a proportion of unequal nucleotide sites of less than 5% was used . Alignments of SNPs and all conserved invariant sites were built using bowtie2 ( http://bowtie-bio . sourceforge . net/bowtie2/index . shtml ) allowing one nucleotide mismatch in each seed alignment ( bowtie2 parameter -N ) . Previously assembled contigs were employed for genomes deposited in Genbank by other groups . Phylogenetic relationships were inferred using RAxML with the GTRGAMMA model ( general time reversible substitution model , gamma distributed rate variation ) and SplitsTree with 100 bootstrap replicates [39] . Evidence of recombination was obtained initially from SNP alignments against the NZ V-13 genome , followed by Split Decomposition analysis and the Phi test as implemented in SplitsTree . Subsequent analysis used statistical approaches described by Sawyer ( 1989 ) [40] , and implemented in GENECONV ( http://www . math . wustl . edu/~sawyer/geneconv/ ) . The core genome of the virulent Psa genomes was identified using OrthoMCL , which uses a likelihood algorithm to place genes in ortholog groups [41] . The core genome was defined as those genes present amongst all isolates within the set: species-specific gene family expansion was excluded so that downstream analysis would not be affected by the inclusion of in-paralogs . The size of the clade specific genes refers to any ortholog cluster exclusively present in that clade and no other , however not all members of the clade need have a representative ortholog . Genes private to particular strains or sets of strains were identified using either ALFY [42] , or a BLAST approach in which , for a given query and a set of subject genomes , we looked for regions unique to the query by extracting the regions that had no ortholog in the subject genomes . The homology search was carried out using translated BLAST with a maximum E-value of 10−10 . Only regions of 1 kb or more were considered . This procedure was implemented in the AWK-script “findUniqueBlast . awk” , which is available in Protocol S1 . Alignments were created for each core ortholog group using PRANK ( http://code . google . com/p/prank-msa/ ) alignment software [43] , [44] . PRANK improves on classical global alignment methods by applying a phylogeny-aware approach that distinguishes between gaps created by insertions or deletions . Independent insertions are prevented from being matched during the progressive alignment process even when they occur at the same position , and the resulting gaps created by these insertions are not penalized during subsequent alignments [45] . Codon-based alignment was employed , with the guide tree inferred by PRANK from the ortholog sequences and default anchoring of pairwise alignments using Exonerate to speed up the alignment process [46] . PRANK alignments and trees of core orthologous groups were used as input for selection analyses using Codeml in PAML 4 . 0 . Codeml runs a likelihood calculation for multiple sequences on a phylogeny to estimate the ratio of non-synonymous to synonymous mutations ( ω ) at every codon site . The ω ratio is a measure of the direction and magnitude of selection on amino acid changes: ω<1 or ω = 1 indicate purifying or negative and neutral selection , respectively , and ω>1 indicates positive selection [43] . The variation in ω among sites is modeled by allowing codons to fall into site classes with ω values ranging from 0 to 1 ( null model 7 ) and comparing the results with a more general model allowing codons to have ω>1 . The likelihood ratio test ( LRT ) statistic was used to identify the most likely model of sequence evolution at a 1% significance level . Bayes Empirical Bayes estimates of model 8 were then used to identify positive selection acting at individual sites [47] , [48] . Subcellular localization was predicted using PSORTb 3 . 0 for core proteins exhibiting statistically significant signatures of positive selection [49] . The set of type 3 secreted effectors ( T3SEs ) present in each genome was identified by tBLASTx ( E value<1e−5 , 20% minimum identity ) using T3SE sequence queries obtained from the T3SE public database ( http://pseudomonas-syringae . org/ ) . The subject database included all virulent and low virulent Psa genomes , as well as P . syringae pv . theae ( Pth ) . Partial hits or T3SEs disrupted by contig breaks are recorded as truncated/disrupted effectors . Name assignment of T3SE orthologs was performed using phylogenetic analyses [50] . tBLASTx was subsequently also used to record significant sequence variation between strains including insertion , deletion , frame shifts and/or translocation events . IslandPath and IslandPick , implemented by IslandViewer [51] , were used as guides for the identification of genomic islands ( GI ) in NZ V-13 . Six genomes were selected on the basis of parameter cutoffs intended to produce a robust dataset of highly probable GIs: P . syringae pv . syringae ( Psy ) B728A ( NC_007005 . 1 ) [52]; P . fluorescens F113 ( NC_016830 . 1 ) [53]; P . fluorescens SBW25 ( NC_012660 . 1 ) [54]; P . fluorescens Pf0-1 ( NC_007492 . 2 ) [54]; P . protegens Pf-5 ( NC_004129 . 6 ) [55]; P . entomophila L48 ( NC_008027 . 1 ) [56]; P . putida BIRD-1 ( NC_017530 . 1 ) [57] . Predicted GIs were subject to manual curation and delineation of the probable GI size using the following criteria: presence of mobile elements such as transposases and integrases; over-representation of virulence-related genes , genes annotated as hypothetical proteins , and/or outbreak clade-specific genes; and the presence of adjacent tRNA genes , which are common phage integration sites [51] . A . deliciosa ‘Hayward’ and A . chinensis ‘Hort16A’ plantlets were inoculated with a representative strain from each clade of Psa or a stonefruit canker causing isolate of P . syringae pv . morsprunorum ( Pmp ) to compare bacterial growth and spread in infected tissues . Each strain was cultured overnight on solid KB medium . The next morning , an inoculation solution was prepared by diluting cells in 10 mM MgSO4 to a concentration of approximately 105 colony forming units ( cfu ) /mL . Three-month-old clonally propagated plantlets were inoculated by dipping a needle into the inoculation solution and stabbing the stem , followed by the placement of a 2 µL drop of inoculum on top of the wound site ( Figure S1A ) . Four separate plantlets were stab inoculated for each treatment . Plants were maintained at 17–20°C with a light/dark photoperiod of 14/10 hours . Humidity was constant at 70% . Bacterial growth was monitored subsequent to inoculation on day 0 and on days 4 , 8 and 14 in both the site of inoculation and the first leaf above the inoculation site . Stem segments and 1 cm2 leaf discs from the base , middle , tip and margin of the leaf ( Figure S1B ) were surface sterilized in 70% ethanol and individually ground in 200 µL 10 mM MgSO4 . Serial dilutions of the homogenate were plated on solid KB medium and incubated for two days at 28°C . The cfu calculated was used to determine the cfu/cm2 in the stem and leaf tissue . Statistical analysis involved fitting Analysis of Variance ( ANOVA ) models using Genstat 14 . 1 on log10 transformed colony count values [58] . This allowed for the model assumption of normally distributed residuals with stable variance to be met . As there were a number of zero observations present in the data , a value of 1 was added to the density data before the log10 transformation was applied . For data collected from stems , F-tests were used to assess if the isolate , time , and their interaction effects were statistically significant . Post hoc tests to compare fitted means were also calculated , with Fisher's protected Least Significant Difference ( LSD ) method used to account for multiple testing . The analyses were carried out separately for each of the two cultivars . The leaf data were analyzed similarly , but also included an area factor that indicated the section of the leaf from which the sample came . The ANOVA tested the area , treatment , and time effects along with all two-way and three-way interactions . In addition , plant ID was treated as a blocking factor because there were multiple measurements on the same plant . All tests were carried out at the 5% level of significance . The back-transformed means are presented in the text with the respective Standard Error Ratio ( S . E . R ) , which is also the back-transformed standard error .
High-quality draft sequences were assembled for 23 strains of Psa from Illumina data ( Table 1 ) . This included five strains from Japan isolated between 1984 and 1988 , three from Korea ( 1997–98 ) , one from Italy ( 2010 ) and 14 from NZ ( all from 2010 ) . The NZ strains were isolated from plants exhibiting either canker disease or foliar symptoms alone and verified as Psa based on diagnostic primers [59] . Strains causing canker disease were consistent with Psa-V , whereas the low virulent ( LV ) form produces mild leaf spotting , but no cankers [28] . In subsequent analyses these 23 draft genomes were supplemented by eleven previously sequenced isolates from China , Italy , Chile and Japan ( Table 1 ) . In addition , the genomes of NZ V-13 and the type strain ( J-35 ) were sequenced with Roche 454 resulting in circularization of each chromosome into a single megascaffold . At approximately 6 . 5 Mb , both chromosomes are larger than other P . syringae pathovars that have been fully sequenced ( Table S1 ) . Each strain possesses an autonomously replicating plasmid: NZ V-13 has a 70 . 1 kb plasmid and the Japanese pathotype strain has one of 32 kb . Restriction digests of plasmid preparations resolved using pulse field gel-electrophoresis ( PFGE ) matched the predicted pattern and confirm the size of the plasmids ( Figure S2 ) . These plasmids are half the size of those reported previously [26] . A synteny plot of the two closed genomes of NZ V-13 and J-35 shows evidence of substantial genomic rearrangement ( Figure 1 ) . This is unexpected for two strains from a single pathovar . Of particular note is the “X” like structure which indicates multiple translocations and inversions centered about the origin . Initial population analyses were based on single nucleotide polymorphisms ( SNPs ) obtained by aligning DNA sequence reads to the Pseudomonas syringae pv . tomato ( Pto ) DC3000 genome ( NC_004632 . 1 ) [60] . This resulted in 15 , 329 SNPs and 463 , 396 invariant sites . Overall diversity as determined by Watterson's theta was 0 . 008 ( per site ) [61] . Phylogenetic analyses identified four distinct clades , of which three exhibit strict phylogeographic structure ( Figure 2A ) . The low-virulent ( LV ) strains responsible for mild symptoms of foliar infection form a single cluster . Leaving aside the divergent NZ LV-14 isolate , the remaining LV strains differ by just 20 SNPs . The three Korean strains , which were isolated in 1997 and 1998 , form a clade defined by 147 SNPs . The Japanese strains isolated between 1984 and 1988 also form a monophyletic group ( 52 SNPs distinguish these strains ) and include the Psa I-1 strain , which was isolated in 1992 during an early incursion of Psa into Italy [17] . The phylogeographic signal displayed by earlier canker-causing and foliar Psa is absent from the group of strains isolated from the 2008 outbreak of Psa ( Psa-V ) : these form a single clade comprising isolates from New Zealand , China and Italy . These strains show little polymorphism , with just five segregating SNPs detected and none separating the Psa-V strains from NZ ( but see below for results of additional analysis of polymorphism within Psa-V ) . While each of the four major clades shows little within-clade diversity , the clades are distinguished by substantial SNP diversity . The LV clades differ from the other three clades by ∼4 , 000 SNPs: the Japanese , Korean and outbreak clades each differ by ∼1 , 000 SNPs . This level of between-lineage diversity is surprising . While reminiscent of differences among three known Pto lineages , contemporary Pto is comprised of just a single homogeneous ( T1 ) lineage [62] . Consistent with low diversity , Pto T1 shows evidence of repeated selective sweeps with signatures of recurrent arms races with its tomato host [62] . A similar picture exists for Pph and P . syringae pv . glycinea ( Pgy ) [63] . In contrast , different outbreaks of Psa have been caused by three related , but nonetheless genetically distinct lineages sampled from a diverse source population . Whether these distinct lineages persist into the future is a matter of interest . That there are such distinct lineages of Psa has significant implications , particularly in light of geographic origin . The first outbreak of canker disease occurred in Japan in 1984 [14] , but with the exception of a single transmission event to Italy ( recorded in 1992 [17] ) , the Japanese lineage remained restricted to Japan . Reports of outbreaks in China were also evident in 1984 , however no strains are available from these incidents [15] . The second recorded outbreak was from Korea , but our phylogenetic data clearly show that this was not the consequence of transmission from the population established in Japan , but was caused by a separate lineage distinguished by numerous SNPs and hundreds of unique genes ( described below ) . The recent ( 2008 ) outbreak of Psa-V represents a third independent lineage , again , distinguished from the earlier lineages by large numbers of SNPs and numerous unique genes ( described below ) . Similarly , the LV lineage detected in NZ in 2010 during surveillance for Psa-V represents a fourth clade , with the likelihood that Psa-LV14 is a fifth ( Psa-LV14 differs from the other LV strains by ∼1 , 400 SNPs ) [27] . The simplest explanation for this striking genetic structuring is the existence of a discrete but diverse source population from which independent transmission events have led to separate outbreaks of Psa disease . While the phylogenetic reconstruction of Psa received strong bootstrap support ( Figure 2A ) , we nonetheless explored the possibility of recombination . The Phi test [64] , based on 6 , 346 informative sites , revealed statistically significant evidence of recombination ( p<0 . 0001 ) . Split Decomposition analysis [65] suggests this is a result of recombination among the Japanese , Korean and Psa-V strains ( Figure 2B ) [39] . An analysis based on 5 , 506 informative SNPs ( from these three clades ) showed 3 , 633 are congruent with the phylogeny depicted in Figure 2A , with the remainder being homoplasies due to recombination . Homologous recombination among the Japanese , Korean and 2008 ( Psa-V ) outbreak lineages would add further weight to the notion of a source population comprised of a set of co-existing and co-evolving strains . If correct , and if recombination has taken place over sufficiently large genomics tracts , and in recent - as well as past - evolutionary time , then a signature of recombination should be evident by regions of clustered SNPs . To this end we mapped SNPs from a representative Korean ( K-26 ) and Japanese ( J-31 ) strain onto the fully sequenced NZ V-13 genome to clarify the positional distribution of polymorphisms . Visual observation showed highly distinctive patterns of clustered SNPs shared between pairs of strains ( Figure 3 ) . Such striking patterns , reminiscent of admixture in sexual populations , led to a statistical analysis of gene conversion ( homologous recombination ) events using GENECONV [40] , which identifies conversion events between pairs of strains in an alignment ( and gene conversion events from strains outside the alignment ) . The results are shown in Table 2 and confirm the visual observations made above ( data files for input into Artemis depicting positions of SNPs arising from comparison between K-26 , J-31 and NZ V-13 are available as supplementary data ( Dataset S1 ) ) . Mapping of the GENECONV-predicted gene conversion events onto the NZV-13 genome showed a close correspondence with regions of highly clustered SNPs ( Dataset S1 ) . GENECONV predicts that approximately 10% of each genome is the result of homologous recombination with gene conversion often being between strains included in the data set . For example , Korean and Japanese lineages have exchanged ∼300 kb ( 158 events of ∼1 . 9 kb in length ) with each other ( Table 2 ) . NZ V-13 shares most with J-31 ( ∼427 kb scattered across 173 regions and comprising ∼2 . 5 kb per gene conversion event ) . While the overall patterns of SNPs , combined with analysis based on Split Decomposition , analysis of homoplasies , the Phi test and GENECONV strongly support the hypothesis that clustered SNPs are a consequence of gene conversion , it is conceivable that these regions might arise through positive or diversifying selection on specific genes , such as those involved in interactions between Psa and its plant host [66] . If so , then genes containing highly clustered SNPs are likely to be those with known or suspected roles in arms races with plant resistance genes . A scan of the genome suggests this not to be the case: most clustered regions span housekeeping genes . Nonetheless , using SNAP [67] to estimate numbers of synonymous and non synonymous nucleotide substitutions [68] , we tested to see whether regions of the genome identified by GENECONV as resulting from gene conversion had more non-synonymous substitutions compared to regions of the genome identified by GENECONV as being involved in gene conversion . An analysis of the normalized ratio of the proportion of non-synonymous to synonymous substitutions ( pN/pS ) based on pairwise comparisons of ∼4000 SNPs from K-26 , J-31 and NZ V-13 showed pN/pS values less that 1 ( the neutral condition ) indicative of purifying – and not positive – selection: mean pN/pS ( ± standard deviation ) for putative non-recombinant regions is 0 . 131±0 . 004 and for recombinant regions is 0 . 088±0 . 002 ) . We therefore conclude that regions of clustered SNPs identified between pairs of strains by GENECONV are a consequence of gene conversion . Notable from observations of patterns of SNPs mapped onto the NZ V-13 genome ( Dataset S1 ) is the fact that some of the shared regions identified by highly clustered SNPs are identical , or almost identical , between pairs of strains ( Figure 3 ) . This indicates that some of the gene conversion events have taken place in recent evolutionary time . This , combined with the fact that each genome carries a signature of gene conversion events scattered around the chromosome , points to the fact that admixture among strains has been a persistent feature of the evolutionary history of these lineages . These patterns of gene conversion events substantiate the notion of a source population . It is difficult to imagine how such patterns could emerge without the sampled lineages having co-existed in close proximity for the bulk of their evolutionary history . GENECONV provides a measure of the overall contribution of recombination to SNPs , relative to mutation . This ranges from ∼8 to 21% for recombination events due to either gene conversion between known or unknown pairs of strains . Overall , ∼44% of all polymorphic sites are likely to be the result of a recombination event . This means that SNPs are approximately twice as likely to be generated by mutation as by recombination . This level of recombination is twice as high as previously reported for the species [69] , but still toward the clonal end of the spectrum of population structures [70] . The origin of the current global epidemic is the subject of interest and speculation , with reports that China is the likely source [22] , [27] , [29] . Previous studies identified six SNPs that distinguished European from Chinese isolates [22]; upon inclusion of two Chilean isolates , nine diagnostic SNPs were identified [27] . Our sequencing of multiple additional isolates of Psa-V confirms that the 2008 outbreak is from a single clone that has rapidly spread around the globe [22] . Whereas Mazzaglia et al [22] relied on read-mapping Psa-V to the partial genome sequence of Pth and Butler et al [27] adopted a similar approach using DC3000 as the reference genome , we were able to map reads to the complete genome of NZ V-13 . We also included the draft genome of the divergent Chinese strain C-9 ( M228 ) in our analysis [27] . Overall we identified 979 SNPs – many of which are unique to C-9: 515 SNPs distinguish Italian , Chinese ( excluding C-9 ) , Chilean and New Zealand Psa-V strains ( Table 3 ) . However , most of these SNPs are located within integrative conjugative elements ( see below ) and are therefore the result of lateral gene transfer ( the number of SNPs across these elements is in the tens of thousands , but are too polymorphic to be detected via read-mapping ) . Because SNPs due to recombination stand to mislead phylogenetic analyses recombinant SNPs were removed ( Table 3 ) for subsequent analyses . As evident in Figure 4 , the Chinese strain C-9 is highly divergent [27] . It shares no SNPs with any of the other Psa-V isolates . On the basis of the available data – and excluding SNPs due to recombination – C-9 differs from C-1 by 351 SNPs ( and other Psa-V strains by similar numbers ) . Leaving aside the Chinese strain C-9 , the Italian strains encompass the greatest level of polymorphism – a finding confirmed through SNP analysis of the genomes of six newly acquired ( unpublished ) Italian strains . This is consistent with the fact that Psa-V was first reported in Italy [17] . The Chilean strains are distinguished by five SNPs . No SNPs were identified among the five NZ Psa-V strains isolated immediately after the outbreak in November 2010 . Of the SNPs that distinguish the four Italian strains , two are shared . No SNPs are shared between NZ V-13 and C-1 . The lack of informative sites means the phylogeny of Psa-V is star-like ( Figure 4 ) . The contemporary isolation of the divergent C-9 isolate is consistent with the suggestion that Psa is endemic on wild Actinidia relatives [27] . However , while C-9 clearly shares a common ancestor with Psa-V , C-9 is not the source of the global outbreak . Assuming a mutation rate in the order of ∼10−7 SNPs per site per year [32] , [71] and the fact that the most divergent clones differ by 52 SNPs , establishment of the Psa-V clone is likely to have occurred less than 10 years ago . C-9 is likely to have diverged from the outbreak clade ∼100 years ago . Identification of genes in the accessory pool often provides clues as to traits of ecological significance . Although there is considerable gene conservation within each Psa clade , there is also evidence of extensive gain and loss . The core genome of the Psa-LV , Pth 2598 and canker-causing Korean , Japanese and outbreak Psa-V strains includes 4425 orthologs , and the flexible genome includes an additional 4710 orthologs ( Figure 5 ) . The P . syringae core genome encodes many proteins contributing to its success in the phyllosphere and endophytic environments , however it is likely that some of these core proteins may present microbe-associated molecular patterns ( MAMPs ) recognized by plant host receptors [72] . MAMP-triggered immunity limits bacterial growth in planta , and can result in selection for amino acid diversification in elicitor-active regions of conserved pathogen proteins [73] . Of the 313 orthologs represented once or more among Psa-V strains , a set of 137 orthologs are found in all outbreak strains , and are absent from the Korean and Japanese Psa clades ( Figure 5 ) . These genes do not share a common history . The most similar orthologs of 39% of these genes are found in different pathovars across 4 different phylogroups of P . syringae and the remainder have close orthologs in plant-growth promoting soil bacteria as well as aquatic and insect associated bacteria , reflecting P . syringae's cosmopolitan lifestyle ( Table S3 ) [72] . Almost 10% of the outbreak-specific genes are annotated as phage integrases or transposases . There are 18 genes present solely within outbreak strains which have top BLAST matches to known vascular and woody pathogens: these represent candidates most likely to confer a virulence advantage to Psa-V . These include genes from P . syringae pv . lachrymans ( Pla ) and P . syringae pv . pisi ( Ppi ) , which invade vascular tissues , and the strains responsible for bacterial canker disease in horse chestnut , plum , and citrus trees ( P . syringae pv . aesculi ( Pae ) ; Pmp and Xanthomonas axonopodis pv . citri str . 306 , respectively ) [74]–[77] . An analysis of selection based on the core genome of all three canker-causing clades of Psa revealed 17 genes that exhibit significant signatures of positive selection ( Table S2 ) . Five of these candidates have non-synonymous substitutions exclusive to the outbreak clade , representing a shortlist of genes whose signatures of diversifying selection could be due to interactions with pattern-recognition receptors of Actinidia species . Curiously , this shortlist includes the flagellar P-ring protein FlgI , which forms a ring structure of the basal body in the peptidoglycan layer . The flagellar filament , comprised of multiple subunits of FliC , is known to harbor an epitope that induces a strong innate immune response in plants [62] , but there are no reports suggesting this of FlgI [78] . Seven of the 17 candidates with high-confidence predictions of subcellular localization are targeted to different layers of the bacterial cell wall , consistent with their potential function as MAMPs [49] . The NZ V-13 70 . 1 kb plasmid is a low copy number plasmid based on the small amount of DNA extracted from large-scale plasmid preparations . Consistent with this is the presence of the stbD and stbE genes which encode a putative segregational stability mechanism probably encoding a toxin-antitoxin system [79] . Analysis of the genes in the plasmid revealed a number of novel features . It possesses three genes encoding predicted T3SEs , HopAU1 , HopAV1 and HopAA1-2 . Two of these effectors possess novel fusions or insertions . HopAV1 has a C-terminal in-frame fusion to a TrbC conjugal transfer protein; and HopAA1-2 has a 106 bp insertion causing a frame-shift . The plasmid also possesses two adjacent gene clusters involved in aromatic carbon metabolism . Both pathways are found in the woody pathogen Pae [72] , and some of the genes are present in the xylem-limited pathogen Xylella fastidiosa [80] . The first cluster consists of two genes involved in the biosynthesis of anthranilate . The second cluster has the hallmarks of a secondary metabolic pathway for a secreted compound including putative regulatory and efflux proteins plus three genes with predicted roles in secondary metabolism . Intriguingly , the first of these is a putative phenylacetate CoA ligase for which anthranilate may be a substrate . It suggests this cluster is making an active compound with a role in the infection of vascular tissue . Many outbreak-specific genes are found clustered with mobile genetic elements in genomic and pathogenicity islands ( GIs , PAIs ) , which are frequently linked to virulence and other environmental adaptations [51] , [81] . Sixteen genomic and pathogenicity islands ( PAI ) ranging in length from 10 to 100 kb were identified in NZ V-13 with IslandViewer . The PAIs are enriched with T3SEs as well as outbreak-specific genes . In addition , there are two prophages with homology to PSPPH06 and a Mu-like prophage . A total of 15 transposons harboring T3SEs or outbreak-specific genes were also identified . Mazzaglia et al [22] identified a polymorphic region with similarity to PPHGI-1 , an integrative and conjugative element ( ICE ) identified in Pph 1302A [82] . This element was further characterized and shown to be capable of excision from the genome and circularization [27] . A hagfish plot mapping the Illumina reads from this region in Psa NZ V-13 showed no significant additional coverage compared to the rest of the genome suggesting that it is present as a single copy per genome ( Figure S3 ) . Detailed analyses of the island present in Psa NZ V-13 , referred to as the Pacific Island , revealed a 100 kb region comprised of ∼100 predicted open reading frames bounded by parA ( left ) and xerC ( right ) and integrated at a 52 base pair att locus ( att-1 ) within a tRNA ( lys ) ( Figure 6 ) . Another copy of this att site ( att-2 ) was found in the Psa NZ V-13 genome 2 . 7Mb from att-1 . This site appears to be unoccupied in Psa NZ V-13 , however the ICE from Psa I-12 is located at att-2 and the corresponding att-1 site is empty ( Figure S4 ) . Comparison with the PPHGI-1 ICE shows striking similarities and marked differences . The “core” genes from the Psa NZ V-13 ICE are syntenous with PPHGI-1 , although they share only ∼75% nucleotide identity . Saturation at the third codon position indicates these elements have an ancient evolutionary past . PPHGI-1 carries a number of accessory genes that are not present in the Pacific Island , which for example lacks the hopAR1 effector . The Pacific Island also carries a novel set of genes located between the Pil locus and the conserved topoisomerase III ( purple genes in Figure 6 ) . This set of genes includes a predicted enolase and various transporters , including an ortholog of dctT , a putative di-carboxylic acid transporter and a methyl-accepting chemotaxis protein ( MCP ) predicted to be involved in taxis toward malate . Interestingly , the DctT transporter carries , at its N-terminus , a sequence predicted by SecretomeP version 2 . 0 to be used to target proteins for secretion ( SecP score = 0 . 64 [83] ) . Analysis using EffectiveT3 to predict proteins targeted to the type 3 secretion system ( T3SS ) ( trained on effector sequences from DC3000 ) returned a highly significant probability score ( 0 . 97 ) , strongly suggestive of type 3-targeting [84] . In accordance with this prediction we note that 14% of the first 50 amino acids of the protein are either Ser or Pro [85] . No other transporter in this region of accessory genes carries a similar signal . While this set of accessory genes awaits experimental analysis , putative functions of a number of the predicted proteins together suggest a role in manipulation of host cell metabolism . Pseudomonas has a preference for growth on TCA cycle intermediates and dctA1 a dicarboxylate transporter gene is important for the virulence of Pto DC3000 [86] . It is possible that the putative T3SS-targeted di-carboxylic acid transporter , enters the plant cell and incorporates into a host cell membranes to facilitate export of this sugar acid in a manner analogous to that recently suggested for á keto glutarate in Xanthomonas oryzae pv . oryzae [87] . Enolase is the penultimate step in glycolysis and the enzyme is exported as part of the RNA degradosome [88] . The role of the enolase in the ICE is not clear . One possibility is that it plays a role in the conversion of dicarboxylic acids to glucose , alternatively it might enhance activity of glycolysis in plant cells . In this regard it is of interest to note that a study of proteins differentially expressed in olive , with and without the pathogen P . savastanoi included dctT and plant enolase [89] . The NZ V-13 ICE is also present in the genome of the Chinese outbreak strain C-1 , however , as previously indicated [22] , [29] , this region is highly divergent in the four Italian and two Chilean outbreak strains [27] . A comparison of the Italian ( Mediterranean Island ) , Chilean ( Andean Island ) and the Pacific Island is depicted in Figure 6 . Several notable findings emerge from the comparison of these Islands . Firstly , the core genes of the Psa islands are syntenous with those of PPGHI-1 , but divergent at the nucleotide level . In fact the core genes of the Mediterranean Island are no more similar to the Pacific Island core genes as to PPHGI-1 . Secondly , the enolase/dctT region is present all three Psa Islands and differs by just four SNPs . This suggests that this region has a recent evolutionary history and has an important function in Psa virulence . Thirdly , while there is near identity at the enolase/dctT locus there are a number of significant differences in the accessory genes . Most notable is the presence of a putative P-loop protein encoded by a gene of ∼6 kb with highest similarity to a gene found in Bordetella ( Figure 6 ) . Finally , all Psa Islands carry an identical MCP/transposase region ( red genes in Figure 6 ) , which has recently been translocated – presumably via the associated transposon – in one or other of the ICEs . The picture of evolutionary change encompassed by these three laterally transferred ICEs is remarkable . Each displays a distinctly different evolutionary history that is independent of the host genome . Nonetheless , the presence of the conserved enolase/dctT region suggests that accessory gene content maybe determined by host-driven selection . Use of such dynamic elements as the basis upon which to infer Psa-V phylogeny is likely to generate erroneous conclusions . The sequenced Psa strains carry 51 known T3SEs ( Figures 7 and S5 ) . Seventeen T3SEs are found in all Psa genomes , while strains in the outbreak clade carry an average of 39 effectors , and strains in the low-virulent ( LV ) clade carry an average of 28 . Many Psa T3SEs are not predicted to be translocated or functional due to the presence of frameshift mutations and transposon insertions . The effector complement is thus highly dynamic even within a single pathovar . T3SE gain , loss and adaptation are likely to have contributed to the enhanced virulence of the outbreak lineage . A large proportion of the variation in effector complement has occurred in a single pathogenicity island encompassing a locus with features similar to that described by Alfano et al , which we have therefore referred to as the exchangeable effector locus ( EEL ) [90] . Table S4 presents a comparison between effectors at the EEL across several isolates and suggests multiple re-arrangements , insertions and deletions have occurred in and around this island . The outbreak EEL is the largest with 16 T3SEs and the non-ribosomal peptide synthase ( NRPS ) cluster ( Table S4 ) . The pathotype strain Psa J-35 EEL is missing the NRPS cluster and has 13 T3SEs , while the Korean and low-virulent clades have fewer T3SEs in this region . The EEL appears to be absent from Pth altogether , although hopAY1 is present . Three effectors in the Japanese and outbreak EEL ( hopBB1-1 , hopBB1-2 , hopX3 ) share both a chaperone ( shcF ) and their predicted proteins share significant N-terminal sequence similarity to HopF2 , as has been observed for AvrRpm2 . The disruption and loss of effectors may play a role in modulating virulence of the outbreak clade since this can result in loss of host recognition . HopA1 , for example , has been inactivated in multiple strains of Psa ( Figure 8 ) . A deletion of shcA and the 5′ fragment of hopA1 has occurred in Psa K-28; the fusion of shcA with an unknown gene downstream of hopA1 has resulted in the deletion of hopA1 in Psa J-35; and lastly , a transposon-mediated translocation has moved the 5′ region of hopA1 600 kb away in Psa NZ V-13 . In Psa K-26 hopA1 contains six non-synonymous mutations; five of these are also present in the fragments of hopA1 found in Psa NZ V-13 . Six codons were also deleted in this strain; surprisingly , in the same position as the sixth substitution in Psa K-28 . Only the NZ LV clade appears to have a fully functional version of HopA1 , suggesting the inactivation of this effector contributes to vascular infection , potentially by abolishing resistance gene-mediated recognition . HopA1 is known to disrupt an Arabidopsis protein complex formed by EDS1 , a central regulator of immunity , and RPS6 , a NB-LRR resistance gene , activating effector-triggered immunity ( ETI ) [91] . Interestingly , HopA1-triggered immunity has also been shown to block the HopM1-mediated degradation of an Arabidopsis protein ( AtMIN7 ) linked with vesicle transport and anti-microbial compound secretion [92] . HopM1 is broadly distributed among P . syringae pathovars and makes an important contribution to pathogen virulence by degrading AtMIN7 . Inactivation of this key host target severely compromises both PAMP and effector triggered immunity ( PTI , ETI ) [92] . The transposon-mediated disruption of HopA1 in the outbreak strain may thus abolish ETI and allow HopM1 to interact with an Actinidia ortholog of AtMIN7 . Transposon movement is doubly implicated in the enhanced virulence and transmissibility of the outbreak strain . The transposon-mediated introduction of HopH1 may have suppressed HopA1-induced ETI , as it has been shown to do in Nicotiana [93] . A small set of effectors appears to have been gained by the outbreak clade ( and are largely absent from the J , K , and LV clades ) . This includes hopH1 and hopZ5: both effectors are adjacent to each other on a small transposon , suggesting introduction by a lateral transfer event . hopZ5 is a new member of the YopJ acetyltransferase family and shares greatest similarity to homologs from Acidovorax and Xanthomonas , although both these lack the putative myristoylation site ( Figure S6 ) . Members of the YopJ acetyltransferase family are believed to impede protein kinases by trans-acetylating key serine or threonine residues in the kinase activation loop [94] . Recently HopZ1a has also been shown to interfere with microtubule assembly in the host , perhaps by binding to and acetylating tubulin – or indirectly via its effect on kinases as detailed above [95] . Another HopZ1 family member has been shown to interact with a protein involved in isoflavone biosynthesis , resulting in increased host susceptibility [96] . The identification of the effector repertoires and outbreak-specific genes may be linked with differences in host-specificity demonstrated by the growth of Psa strains on different kiwifruit cultivars . Growth assays of representative isolates from the low-virulent clade ( NZ LV-5 ) and canker-causing Korean ( K-26 ) , Japanese ( J-35 ) , and outbreak ( NZ V-13 ) were performed in A . deliciosa cv . ‘Hayward’ and A . chinensis ‘Hort16A’ in order to understand dynamics of in planta growth . Assays involved regular quantification of bacterial density at both the site of initial inoculation ( stem ) and at three spatially distinct regions of the first leaf above the inoculation point ( Figure S1B ) . This latter measure provides a quantitative indication of the capacity of a given strain to spread in planta . Pmp , which causes canker in Prunus spp . , was included as a negative control . Differences among treatments were tested using a three-way ANOVA within each cultivar , testing for effects of isolate ( NZ V-13 , NZ LV-5 , J-35 , K-26 and Pmp ) , time ( days 0 , 4 , 8 and 14 ) and sample location ( stem ( point of initial inoculation ) , base of first leaf , middle of first leaf and periphery of first leaf ) , with Fisher's protected LSD post-hoc test at α = 0 . 05 ( Supplementary Table S5 ) . On ‘Hort16A’ J-35 reached highest cell densities in the leaf at day 14 , however , NZ V-13 spread more rapidly reaching the leaf tip by day eight ( Figure 9A ) . K-26 showed reduced capacity for both growth and in planta spread . This capacity was further reduced in NZ LV-5 . Pmp failed to spread from the point of inoculation , or persist within stem tissue , providing a compelling example of the effectiveness of non-host resistance in vascular plant tissue . On cultivar ‘Hayward’ all strains showed a reduction in both growth and in planta spread ( relative to ‘Hort16A’ ) . This was particularly striking for NZ V-13 which attained moderate leaf population densities by day eight , failed to spread beyond the leaf base , and was barely detectable by day 14 ( Figure 9B ) . Overall these data provide evidence of cultivar-specific differences in the growth dynamics of Psa from different clades . Noticeable is the enhanced growth and spread of NZ V-13 on Hort16A relative to its performance on ‘Hayward’ . This apparent tradeoff is suggestive of adaptation of this Psa lineage to A . chinensis . J-35 , however , was equally capable of colonizing both ‘Hort16A’ and ‘Hayward’ , although showed a reduced capacity for in planta spread . This suggests that differences in the gene content between NZ V-13 and J-35 may be central to understanding the enhanced transmission of 2008 outbreak strains both within and between plants [28] . While differences in the growth of all three canker-causing lineages were evident , similarities were also apparent . For example , all three , despite substantial difference in accessory gene context , were capable of spread and persistence within both kiwifruit cultivars . In light of the predicted existence of an Asian source population this finding has special relevance . The fact that three strains with such different genomes can each grow in planta raises the possibility that the capacity to cause disease in kiwifruit resides primarily in the core – as opposed to accessory – genome . If so , then it is possible that numerous different strains from the source population may be able to cause disease in kiwifruit . This highlights the importance of understanding the source population and minimizing the chance of future transmission events from this reservoir . It also suggests that strategies to develop durable resistance should not focus solely on effectors or other typically targeted components of the accessory genome . The origins of crop diseases are linked to domestication of plants [3] . Since most crops were brought under domestication many centuries ago , opportunities to understand the emergence of disease are limited . Kiwifruit is an exception: it is one of the few plants to have been domesticated in the 20th Century [12] . Our in-depth genomic and population analysis of more than 30 Psa strains from diseased kiwifruit vines – strains representing each of three major disease outbreaks obtained from different geographic regions over the course of three decades – has captured initial stages in the emergence of a pathogen population concomitant with domestication of its host . Particularly striking is the genetic structure of the Psa population . Rather than a single genetically homogeneous lineage that establishes in one region before transmission to another , each outbreak of Psa represents a distinct lineage with its own unique repertoire of accessory genes including effectors and toxins . There can be little doubt that each lineage represents an independent sampling from a single source population . This is evident in the shared ancestry ( they form a single monophyletic group ) , the marked signatures of within pathovar gene conversion , and lack of diversity within each disease-causing lineage . The geographic location of the predicted source population and the extent of diversity encompassed awaits elucidation , nonetheless , given the recent timing of kiwifruit domestication it is likely that the source population is associated with wild Actinidia species in Asia – a finding supported by the divergence of strain C-9 [27] . The diversity of genotype and virulence is likely substantive given knowledge of the heterogeneous characteristics of the lineages studied thus far . For example , the LV isolates from New Zealand cause leaf spot but minimal damage to vines , whereas Psa-V isolates cause cankers and stem wilt , are highly transmissible , and have devastating effects on vine health . Interestingly , the source population encompasses Pth , the tea pathogen . Given that both hosts , tea and kiwifruit , have their origins in Asia , this finding is perhaps not surprising . Whether existence of two different pathovars within a single lineage represents an instance of host-shift or divergence from within the source population is unclear , but nonetheless highlights the importance of defining the nature of the source population . The concept of a source population from which disease arises by sampling events has implications for both disease control and plant breeding . From the perspective of disease control there are two interrelated issues . The first is opportunity for transmission from the source population , the second is opportunity for a given transmission event to result in establishment of a new population . Every effort needs to be made to ensure that routes of transmission from the source population to commercial orchards – and between commercial orchards – are minimized . This includes protocols to prevent transmission of Japanese and Korean lineages , which our growth data show could be equally problematic , were they to arrive in countries such as New Zealand from which these lineages are currently absent . At the same time , there is need to move from orchards comprised of single clone varieties in order to reduce opportunity for plant-to-plant transmission and the establishment of high pathogen titers . It is possible that densely planted orchards of A . chinensis in Italy were central to the establishment and global dissemination of Psa-V . Challenges to plant breeding are potentially more significant . Producing kiwifruit vines resistant to Psa-V is an important goal , however , resistance is unlikely to be durable if breeding programs exclusively target this lineage . The extensive diversity of effectors and associated virulence genes evident among the currently known Psa lineages , combined with capacity for lateral transmission , means that new Psa variants – either from the source population or produced de novo by recombination and horizontal gene transfer – are to be expected . This study has identified 17 effectors conserved across all Psa lineages that might usefully be the focus of a breeding program , while metabolic and physiological factors should not be overlooked . In this regard it maybe of more than passing interest that the only significant genomic change that has thus far occurred among the extant Psa-V lineage involves three ICEs , which despite highly diverse sets of core genes contain an identical cargo of accessory genes predicted to manipulate host metabolism . Human activities have long shaped the evolution of microbes [97] Ensuing feedback effects on the evolution of human populations – particularly in the context of human disease – are widely recognized [71] , [98] , [99] . Less well understood are anthropomorphic impacts on the evolution of plant pathogens [3] . Although our study provides just the first glimpse of a plant disease as it emerges in the face of domestication , continued analysis of the population processes and genetic phenomena underpinning evolution of Psa stand to enlighten both critical events driving the evolution of virulence and broader issues associated with the evolutionary interaction between microbes , their hosts , and the communities of which they are part . | Despite considerable scientific advances in plant protection during the last century , agricultural crops remain vulnerable to infection by pathogens . The intensive cultivation particularly of clonally propagated crop plants increases the potential for the emergence and rapid spread of new diseases . Pseudomonas syringae pv . actinidiae was first reported as a canker-causing pathogen of kiwifruit in the mid-1980s . However , a new outbreak of the disease occurred in 2008 and this strain has spread rapidly throughout growing regions of the world . In order to determine the origin , population structure and defining features of this pathogen , a large-scale sequencing project was established . This clarified the phylogenetic relationships between the different Psa isolates and identified the outbreak-specific gene sets associated with the aggressive systemic infection strategy exhibited by the virulent strain . This information is invaluable in developing robust long-term solutions for this serious disease . Given that kiwifruit production on a commercial scale is a relatively recent event , this analysis provides a unique insight into the evolution of this pathogen with its host , from its first emergence to the latest global outbreak . This understanding should aid in the mitigation of devastating outbreaks in the future . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"and",
"Discussion"
] | [
"evolutionary",
"ecology",
"genome",
"evolution",
"comparative",
"genomics",
"biology",
"genomics",
"evolutionary",
"biology",
"genomic",
"evolution",
"evolutionary",
"genetics"
] | 2013 | Genomic Analysis of the Kiwifruit Pathogen Pseudomonas syringae pv. actinidiae Provides Insight into the Origins of an Emergent Plant Disease |
Plasmodium sporozoites , the causative agent of malaria , are injected into their vertebrate host through the bite of an infected Anopheles mosquito , homing to the liver where they invade hepatocytes to proliferate and develop into merozoites that , upon reaching the bloodstream , give rise to the clinical phase of infection . To investigate how host cell signal transduction pathways affect hepatocyte infection , we used RNAi to systematically test the entire kinome and associated genes in human Huh7 hepatoma cells for their potential roles during infection by P . berghei sporozoites . The three-phase screen covered 727 genes , which were tested with a total of 2 , 307 individual siRNAs using an automated microscopy assay to quantify infection rates and qRT-PCR to assess silencing levels . Five protein kinases thereby emerged as top hits , all of which caused significant reductions in infection when silenced by RNAi . Follow-up validation experiments on one of these hits , PKCς ( PKCzeta ) , confirmed the physiological relevance of our findings by reproducing the inhibitory effect on P . berghei infection in adult mice treated systemically with liposome-formulated PKCς-targeting siRNAs . Additional cell-based analyses using a pseudo-substrate inhibitor of PKCς added further RNAi-independent support , indicating a role for host PKCς on the invasion of hepatocytes by sporozoites . This study represents the first comprehensive , functional genomics-driven identification of novel host factors involved in Plasmodium sporozoite infection .
Although malaria has long been a devastating killer for the most vulnerable populations in countries of sub-Saharan Africa and other developing nations , our understanding of the early host-parasite interactions underlying this infectious disease remains far from complete . In fact , the first stage of a malaria infection , which occurs in the liver once the Plasmodium parasite has been delivered through the bite of an infected female Anopheles mosquito , is still clearly under-studied today . Once inside the mammalian host , Plasmodium sporozoites , the motile form that is delivered in the mosquito's saliva , display a marked tropism for hepatocytes , the cells that enable the remarkable replication process that will give rise to thousands of merozoites from each invading parasite ( reviewed in [1] ) . As a first step towards infection , several hepatocytes are transiently traversed by the sporozoite before one cell is productively invaded , leading to the formation of a parasitophorous vacuole [2] . Within this cytosolic vacuole , the subsequent development and asexual replication of Plasmodium , constituting so-called exoerythrocytic forms ( EEFs ) , achieve one of the fastest growth rates among all eukaryotic cells . The invaded hepatocyte eventually releases thousands of mature merozoites into the bloodstream [3] , where these then invade erythrocytes , thereby initiating the so-called blood stage of infection and triggering the well-known symptoms of malaria . Both the strong tropism and obligate nature of the events that take place during liver infection suggest an essential requirement for hepatocyte-specific factors in enabling this complex lead-up to the blood stage . It is therefore of primary interest to identify and characterize the role of such host factors , as these may contribute to the design of rational interventional strategies for the development of novel prophylactic agents . To this end , we have used a cultured cell-based assay to study the process of liver infection by Plasmodium parasites at the cellular and molecular level . Using human Huh7 hepatoma cells and sporozoites of the rodent parasite P . berghei freshly isolated from infected Anopheles mosquitoes , we have established a high throughput assay system ( Figure 1A ) that , combined with high content readouts using automated microscopy , and quantitative RT-PCR ( qRT-PCR ) , can be used for RNA interference ( RNAi ) and/or drug screening experiments . Intracellular phosphorylation and dephosphorylation events are enzymatically catalysed by kinases and phosphatases , respectively , and constitute the most important signalling mechanisms known in eukaryotic cells [4] . The phosphorylation state of a protein can determine its activity and , thereby , regulate the pathway ( s ) in which it is involved . Thus , the present study probed the potential role of key components of host hepatocyte signal transduction pathways , focusing on kinases as key regulators for a wide range of cellular functions .
We have used systematic RNAi screening to selectively silence the expression of 727 genes encoding proteins with known or putative kinase activity , as well as kinase-interacting proteins , thereby covering the entire annotated kinome ( Table S1 ) . The effect of each gene-specific knock-down on the infection of Huh7 cells by Plasmodium sporozoites was then monitored using the high-throughput , high-content immunofluorescence microscopy-based assay mentioned above ( Figure 1A ) . Briefly , short interfering RNA duplexes ( siRNAs ) targeting each of the chosen genes were transfected into Huh7 cells 24 h after seeding in 96-well plates . Forty-eight h later , cells were infected with P . berghei sporozoites . Cells were fixed 24 h after infection and immuno-stained to detect intracellular parasites ( EEFs ) , as well as host cell nuclei and F-actin to estimate cell numbers and confluency , respectively . Following image acquisition , customized image analysis algorithms were used to automatically quantify infection rates , normalizing the number of EEFs against the cell confluency in each well . A plate-wise normalization was also used to facilitate comparisons between plates in the first pass of the screen , where the low rate of positive hits yields minimal expectation of variability in the mean infection values between different plates . To this end , the infection rate in each experimental well was calculated as a percentage of the mean infection rate from all experimental wells on that plate . In order to assess possible siRNA effects on cell proliferation , infection rate data were plotted against the number of nuclei , also expressed as a percentage of the mean number of nuclei for that plate . The RNAi strategy employed was validated by targeting 53 randomly chosen genes with 3 siRNAs each and performing quantitative real-time PCR ( qRT-PCR ) analysis to determine the level of knock-down achieved in each case . For 13 of these genes either expression was too low to be correctly assessed or primer specificity was insufficient . Most importantly , for 85% of the genes whose expression could be determined , at least 1 of the siRNAs led to an expression knock-down greater than 70% ( Figure 1B ) . Sporozoite infection assays inevitably have considerable levels of variation , a problem that cannot currently be overcome and which exacerbates difficulties generally associated with siRNA screens . In order to reduce the risk of reporting false positives , a multi step screening system was devised in which candidate genes were subjected to three screening passes with increasingly stringent selection criteria ( Figure 2A ) . In the first pass , the 727 selected genes were screened by targeting each with three distinct siRNAs used individually ( Table S1 ) . In order to minimize the number of false negative results , candidate gene hits were selected for follow-up in pass 2 if any single one of the three siRNAs yielded an increase or decrease on infection greater than 2 standard deviations ( s . d . ) of the average of the infection of the whole data set , within a defined range of nuclei number ( ±40% of the average number of nuclei in each experimental plate ) ( Figure 2B ) . The latter precaution , while relatively inclusive , allowed us to exclude from further analysis those siRNAs yielding strong effects on cell proliferation or survival . As a result , 73 genes were selected to undergo a second pass of confirmation screening , in which up to 2 additional siRNAs were included to maximize the detection sensitivity for those genes that had yielded only a single siRNA hit in pass 1 ( Table S1 ) . In this round of analysis , siRNAs were noted as “positive candidates” if they yielded infection rates more than 2 s . d . above or below the mean of all the negative controls in this pass . Negative controls replaced whole data set mean for s . d . calculation , since the selected subset of genes in this pass 2 was expected to have a significantly higher hit rate than in pass 1 . To minimize the risk of false positives due to siRNA sequence-dependent off-target effects , the selection of candidate genes for follow-up beyond pass 2 required that at least two independent siRNAs targeting the same gene be “positive candidates” according to the above selection criteria ( Figure 2C ) . Furthermore , genes for which different siRNAs yielded conflicting phenotypic results were also excluded from further analysis . In order to further minimize any bias due to experimental variability , all pass 2 siRNAs were assayed in two independent experiments , and were selected for follow-up only if the criteria were met in both experiments ( Figure 2D ) . It is worthwhile noting that while in Pass 1 only 3 . 6% of the siRNAs met the selection threshold , 18 . 4% of the siRNAs tested for the first time in Pass 2 met similar criteria , while the distribution of infection levels in controls is not statistically different between pass 1 and pass 2 experiments , showing that a 5-fold enrichment has taken place from Pass 1 to Pass 2 . The 16 genes thus selected for further verification in pass 3 were targeted with the siRNAs yielding the strongest phenotypes in the second pass . This third pass was used to further restrict our selection to those genes showing clearest functionality , i . e . those with at least two siRNAs yielding infection rates more than 3 s . d . above or below the mean of all the negative controls in the assay , respectively ( Figure 2E ) . Secondly , target mRNA knock-down levels attained for these genes were also assessed in this pass by qRT-PCR . This allowed the selection of positive hit candidates to be refined further yet by excluding genes for which a correlation between phenotypic severity and decreased mRNA levels could not be confirmed ( Figure 2F , Table 1 ) . Based on these data , the following 5 genes have emerged from our screen as the clearest and strongest positive hits , showing RNAi-induced loss-of-function phenotypes with specific , reproducible and marked effects on P . berghei infection rates in our Huh7-based assay: MET , PKCζ ( PKCzeta ) , PRKWNK1 , SGK2 and STK35 . As illustrated in Figure S1 , knock-down of the expression of these genes did not lead to any significant effects in terms of cell proliferation or morphology ( see also Table 1 ) . It should also be noted that the present data do not rule out the possible involvement of other genes among those tested here , since negative results in RNAi screens are generally inconclusive [5] , and certain genes showing phenotypes with lower than 3 s . d . from mean levels in our assays may provide real , though perhaps more subtle , functionalities in this context . For this reason , the reader is referred to Table S1 for a comprehensive list of siRNAs employed throughout the screen and their corresponding z scores ( which measure the number of standard deviations away from the mean for the whole normalized data set ) , obtained throughout the screen . It may be noted that all five genes we identified as top hits encode protein kinases belonging to 3 different classes , according to the kinome classification [6]: “AGC” , “other” and “TK” ( Figure S2A ) . While it is tempting to draw conclusions from this , we would advise against it since , as aforementioned , the experimental methods and prioritization strategies used here cannot conclusively rule out the involvement of other tested genes which did not make the final selection . The range of cellular processes implicated by the genes identified as top hits in the present screen include cell cycle control , cytoskeleton regulation , osmotic balance and stress/immune responses ( Figure S2B ) . This is consistent with the broad range of cellular functions similarly implicated by Agaisse et al . [7] in the infection of Drosophila cells by intracellular bacterial pathogens . However , when we performed an hypergeometric test to identify which Gene Ontology ( GO ) terms were significantly enriched in the analysis ( p<0 . 05 ) , not all categories were equally represented in passes 2 and 3 with a major proportion of genes involved in cytoskeleton regulation . Interestingly , amongst the five final hits , we could observe a complete shift of representation , as a 50/50 segregation was observed for genes related to stress/immune responses and to cell cycle control ( Figure S2C ) . In order to further characterize the functionalities identified in our cell-based infection model and to validate their relevance both from a physiological point of view and in terms of human malaria , we have initiated detailed follow-up studies for all 5 of the top hits from our RNAi screen , and present herein results for PKCζ , the first of these to be prioritized due to its role in several liver pathological processes [8] , [9] . PKCζ is part of the large family of PKCs , which has been implicated in a wide range of cellular processes . PKC isotypes include 10–15 members , divided into 4 groups [10] . One of these groups , known as the atypical PKCs ( aPKCs ) [11] , comprises the PKCζ [12] and PKCλ/ι ( PKC lambda/iota ) [13] isoforms . The aPKCs have been implicated in numerous processes , including cell growth and survival , regulation of NF-κB ( NF-kappaB ) activation and polarity ( reviewed in [11] , [14] , [15] ) . All PKC isoenzymes have an autoinhibitory pseudosubstrate domain sequence that can bind to the substrate-binding cavity and prevent catalysis [16] . This inhibtory effect can be mimicked in vitro by addition of a corresponding synthetic peptide [17] . Thus , we used the cell-based assay described above for our screen to test the effects of a myristoylated PKCζ pseudosubstrate ( myr-SIYRRGARRWRKLYRAN ) , previously characterized as a specific PKCζ inhibitor ( PKCζInh ) [18] , [19] , on P . berghei infection . Further data on the specificity of PKCζInh is shown in Figure S3 . A scrambled myristolated peptide was used as control in all PKCζ inhibition experiments [18] . Treatment of cells with PKCζInh had no obvious effects on nuclear or cell morphology and as well as on the cell number and confluency ( Figure 3A–C ) , as previously observed for cells transfected with siRNA oligonucleotides targeting PKCζ ( Figure S1 , Table 1 ) . Still , treatment of cells with PKCζInh had a significant effect in the level of cell infection by P . berghei sporozoites , as quantified using qRT-PCR-based measurements of Plasmodium 18S rRNA levels found within Huh7 cells ( Figure 3D ) and mouse primary hepatocyte extracts ( Figure 3E ) harvested 24 and 48 h after P . berghei sporozoite addition , respectively . Our results show that a 20 µM concentration of PKCζInh leads to a ∼80% and ∼60% reduction in P . berghei infection rates in Huh7 hepatoma cells and primary hepatocytes , respectively ( Figure 3D and E; p<0 . 01 and p<0 . 05 ) , offering a RNAi-independent confirmation of our present findings on the role of PKCζ in P . berghei infection . In order to gain a better insight on the possible role of PKCζ in the infection process , the effects of PKCζInh on different periods of hepatocyte infection were examined by flourescence activated cell sorting ( FACS ) analysis of host cells infected with GFP-expressing P . berghei parasites , measuring the proportion of GFP+ cells [20] . Indeed , FACS analysis of cells infected with GFP-expressing parasites enables discerning whether the observed effect on infection is due to a decrease in the number of infected cells or to an impairment of Plasmodium development inside host cells [20] . Treatment of Huh7 cells with PKCζInh 1 h prior to addition of GFP-expressing P . berghei sporozoites led to a marked , dose-dependent decrease in infection rate , as measured by the proportion of infected cells relative to control samples 24 h after sporozoite addition ( Figure 4A; p<0 . 05 for PKCζInh≥5 µM ) . Treatment with PKCζInh did not affect Plasmodium development , as shown by the similar GFP intensities of treated and control cells ( Figure 4B ) . Next , we sought to determine whether the decrease in the number of infected cells observed at 24 h after sporozoite addition was due to a decrease in invasion rate or to the disappearance of infected cells throughout infection . Since , in the infection assay employed , >95% of invasion events are known to take place within the first 2 h after sporozoite addition [20] , any effects on invasion can be quantified by analyzing cells at this timepoint . As shown in Figure 4C , the effect of PKCζInh in cells analyzed 2 h after sporozoite addition is closely comparable to that seen with the full 24 h treatment , indicating that PKCζ likely plays a role during host cell invasion by P . berghei sporozoites ( p<0 . 05 for PKCζInh≥5 µM ) . In addition , when PKCζInh was added 2 h after sporozoite addition , no significant effect was observed in infection rate measured at 24 h ( Figure 4D ) , not only showing that the effect observed on the early steps of infection is not due to PKCζInh toxicity to host cells but also strengthening the notion that PKCζ influences Plasmodium infection by playing a role during cell invasion . Importantly , infection rates were not affected by pre-incubation of Plasmodium sporozoites with PKCζInh for 1 hour prior to their addition to hepatoma cells , showing that PKCζInh has no direct effect on sporozoite viability ( Figure 4E ) . Further confirmation of the involvement of PKCζ in sporozoite invasion of Huh7 cells , but not on the parasite's intracellular development was obtained by employing qRT-PCR to quantify infection 24 h after infection of cells incubated with PKCζInh either during the invasion or the development periods , exclusively ( Figure 4F ) . These results show that a marked decrease in intracellular parasite numbers is observed when cells are incubated with the inhibitor during the first 2 hours after sporozoite addition ( p<0 . 001 ) , whereas no effect is observed when the compound is added after invasion is completed . Together , these data confirm the physiological relevance of PKCζ , identified in the RNAi screen , and suggest that the latter plays a role during the invasion of hepatoma cells by P . berghei sporozoites . Finally , we tested the in vivo physiological relevance of our cell-based findings more thoroughly by using systemically-delivered , liposome-formulated siRNAs designed to specifically silence PKCζ expression in adult mice , and infecting these with P . berghei sporozoites . In vivo RNAi treatments using the same systemic administration of siRNAs including the same formulation used here have previously been shown to yield potent gene-specific knock-downs in adult mice without major toxicity , nor any detectable disruption of the endogenous microRNA pathway [21]–[23] . In our present experiments , mice from the same litter were given an initial intravenous ( i . v . ) injection of either test or control siRNAs and , infection was initiated 36 h later by i . v . injection of freshly isolated P . berghei sporozoites . Mice were sacrificed 40 h after infection to permit parallel analyses of gene silencing and infection load . In order to address the risk of sequence-dependent off-target effects , three distinct siRNA sequences targeting PKCζ were tested individually , while a siRNA targeting luciferase , a transcript known to be absent in these mice , was used to address sequence-independent off-target effects that may arise from these treatments . Under these conditions , no toxicity was observed ( Figure S4 ) and PKCζ expression was reduced in adult mouse livers when using each of the 3 distinct PKCζ-specific siRNAs , yielding an average of ∼56–73% remaining PKCζ mRNA , as measured by qRT-PCR of liver extracts taken 76 h after siRNA treatment , relative to the controls ( Figure 5A; p<0 . 05 ) . This silencing was accompanied , for all three PKCζ-specific siRNAs , by significant reductions in liver infection , yielding an average per siRNA of ∼9–40% of control infection loads , as measured by qRT-PCR of P . berghei 18S rRNA in liver extracts taken 76 h after siRNA treatment , as described above ( Figure 5A , p<0 . 05 ) . The reductions in liver infection load showed a broad correlation with the level of PKCζ silencing achieved by the siRNAs with siRNA #1 , which leads to the most significant reduction in PKCζ , showing the most striking difference in infection ( Figure 5A; p<0 . 01 ) . In a further , parallel experiment , semi-quantitative Western blotting analysis of liver extracts taken 76 h after siRNA treatment , from mice injected with the PKCζ siRNA yielding the strongest reduction in liver infection , confirmed that PKCζ expression was significantly reduced at the protein level in these mice ( ∼55%; Figure S5 , p<0 . 01 ) . Additionally , another 3 independent groups of mice treated with the same 3 distinct PKCζ siRNAs showed a decrease in blood parasitaemia ( percentage of infected erythrocytes ) , relative to control mice ( Figure 5B ) . In fact , while by day 4 after sporozoite injection all 5 mice in the control group were positive for blood stages , none of the 6 mice in the group pre-treated with the strongest PKCζ-specific siRNA were ( Figure 5B; p<0 . 001 ) . Although less striking , both other siRNAs also led to a delay in the appearance of parasites in the blood and siRNA #2 led to a significant reduction in average blood parasitaemia ( Figure 5B; p<0 . 05 ) . Together , these data strongly support the conclusion that PKCζ is a physiologically important host factor needed for the liver stage of Plasmodium infection both in cultured cells in vitro and in animals in vivo .
The approach described here constitutes , to our knowledge , the first report of a genome-scale RNAi screen for key host factors for infection of human cells by a parasite . Plasmodium sporozoite infection in vitro assays are fraught with specific biological variability issues that enhance the difficulties inherent to any high-throughput RNAi screening assay . Thus , in order to minimize the chance of excluding false negative and identifying false positive results , we have devised a multi-step strategy employing increasingly stringent selection criteria throughout the screen . Nevertheless , we cannot definitively rule out the possibility of , in this process , having discarded candidates that may indeed play a role during infection , but which did not “survive” the criteria employed throughout the three stages of selection . For this reason , we encourage the reader to consult Table S1 , where details of the screening process are presented and potentially relevant genes can be identified . Our systematic analysis of the human kinome in the context of the liver stage of malaria infection has directly implicated at least five host kinases in this process: MET , PRKWNK1 , SGK2 , STK35 and PKCζ . By doing so through direct functional tests for each of the genes assayed , this dataset establishes clear causal roles in the processes examined and reveals novel key host molecules in these pathways that significantly affect Plasmodium's success in infecting hepatocytes . Among our top hits , the MET gene , which encodes the hepatocyte growth factor ( HGF ) receptor , is the only one to have been previously shown to influence Plasmodium infection of hepatocytes [24] , [25] , an effect that has been proposed to occur through inhibition of apoptosis [25] . It has been demonstrated that transfection of hepatoma cells with a dominant-negative form of MET leads to a reduction in Plasmodium infection while transfection with a constitutively active form results in an infection increase [24] . Thus , the emergence of MET as one of our top hits whose knock-down consistently led to a decrease in infection represents a validation of the screening methodology used here , and strengthens the predictive value of the other hits . Among those , both SGK2 and PRKWNK1 are serine/threonine kinases that have been implicated in osmotic control through the regulation of Na+ and K+ transport channels [26]–[29] . Down-modulation of both of these osmotic and oxidative stress-responsive proteins also led to a reduced infection in the present screen . Although their role in Plasmodium infection remains unclear , the present data may be highlighting the importance of maintaining an optimal osmotic balance in the host cell to permit successful infection . In addition , it has recently been shown that exposure of sporozoites to the intracellular K+ concentration enhances sporozoite infectivity [30] . Whether or not SGK2 or PRKWNK1 act on infection through the control of K+ concentration will require further investigation . Concerning STK35 , it is known to interact with CLP-36 , a PDZ-LIM protein , and re-localize from the nucleus to actin stress fibres [31] . This has led to the suggestion that STK35 may act as a regulator of the actin-myosin cytoskeleton in non-muscle cells [31] . Indeed , there are indications that the reorganization of the host cell actin cytoskeleton may be important for Plasmodium infection [24] . Thus , it is appealing to consider the hypothesis that recruitment of STK35 may influence Plasmodium infection by playing a role in this process . Finally , the gene which we have characterized in most detail here , PKCζ , is part of the large family of PKCs that has been implicated in numerous cellular processes . PKC isotypes include 10–15 members , divided into 4 groups [10] , [16] . One of these groups , known as the atypical PKCs ( aPKCs ) [11] , comprises the PKCζ [12] and PKCλ/ι ( PKCiota/lambda ) [13] isoforms . The aPKCs have been implicated in numerous processes , including cell growth and survival , regulation of NF-κB activation and polarity ( reviewed in [11] , [14] , [15] ) . In the present study , loss of PKCζ function both in vitro and in vivo , whether by RNAi silencing or by pseudo-substrate inhibition , led to decreased infection rates . This unequivocally establishes a key role of PKCζ in host liver malarial infection , thereby , giving confidence that the other genes identified in the RNAi screen also play a relevant role during infection . Furthermore , as we gain additional understanding of the pathway defined by PKCζ in this context , its potential biomedical value may develop not merely as one , but rather also as the founding member in an entirely novel class of anti-malarial drug target . In addition , our data reveals a role for PKCζ signaling in host cell invasion by Plasmodium sporozoites . To our knowledge , this is the first host cell signaling molecule to be identified as an important player in Plasmodium invasion and paves the way to a better understanding of this essential host-Plasmodium interaction in the establishment of a malaria infection . The interaction of a pathogen with its host cell activates intracellular signaling cascades that regulate innate immune responses and govern the outcome of infection ( see [32] for a recent review on this topic ) . These signaling pathways may also be exploited by the pathogen to its own benefit as was recently suggested for the interaction of hepatitis C virus , another major liver pathogen , with its target cells [33] . While our present study does not demonstrate any active exploitation or modulation of host signal transduction pathways by Plasmodium , it does reveal novel key host molecules in these pathways that significantly affect the parasite's success in infecting hepatocytes . Our efforts are thus starting to yield crucial molecular details needed to build a coherent picture of the key cellular events taking place during the liver phase of malaria infection . Altogether , these results contribute toward a better understanding of host-pathogen interactions , which may help in accelerating the rational design of prophylactic , therapeutic and/or diagnostic strategies aimed to control malaria .
Huh7 cells , a human hepatoma cell line , were cultured in RPMI medium supplemented with 10% fetal calf serum ( FCS , Gibco/Invitrogen ) , 1% non-essential amino acid ( Gibco/Invitrogen ) , 1% penicillin/streptomycin ( pen/strep , Gibco/Invitrogen ) , 1% glutamine ( Gibco/Invitrogen ) and 1% HEPES , pH 7 ( Gibco/Invitrogen ) and maintained at 37°C with 5% CO2 . Mouse primary hepatocytes were obtained as previously described [34] . Briefly , they were isolated by perfusion of mouse liver lobule with liver perfusion medium ( Gibco/Invitrogen ) and purified using a 1 . 12 g/ml; 1 . 08 g/ml and 1 . 06 g/ml Percoll gradient . Cells were cultured in William's E medium containing 4% FCS , 1% pen/strep , 50 mg/ml epidermal growth factor ( EGF ) , 10 µg/ml transferrin , 1 µg/ml insulin and 3 . 5 µM hydrocortisone in 24 well plates coated with 0 , 2% Gelatine in PBS . Cells were maintained in culture at 37°C and 5% CO2 . C57BL/6 mice , were bred in the pathogen-free facilities of the Instituto de Gulbenkian de Ciência ( IGC ) and housed in the pathogen-free facilities of the Instituto de Medicina Molecular ( IMM ) . All protocols were approved by the Animal Care Committees of both Institutes . Green fluorescent protein ( GFP ) expressing P . berghei ( parasite line 259cl2 ) sporozoites [35] were obtained from dissection of infected female Anopheles stephensi mosquito salivary glands . All siRNAs were purchased from Ambion's Silencer genome wide library ( Ambion/Applied Biosystems , Austin USA ) . Each gene was targeted by using distinct siRNAs used individually in all cases . Negative control samples included untransfected cells , and cells transfected with a negative control siRNA not targeting any annotated genes in the human genome . A full list of gene names , siRNA ID numbers and sequences , and associated screening data are shown in Table S1 . Huh7 cells ( 4500 per well ) were seeded in 100 µl complete RPMI medium in optical 96-well plates ( Costar ) and incubated at 37°C in 5% CO2 . Twenty-four h after seeding , cells were transfected with individual siRNAs in a final concentration of 100 nM per lipofection . Each siRNA was transfected in triplicate . Briefly , for each well , cell supernatant was replaced by 80 µl of serum-free culture medium without antibiotics . One µl of 10 µM siRNA diluted in 16 µl of Opti-MEM ( Invitrogen ) was complexed with 0 . 4 µl Oligofectamine ( Invitrogen ) diluted with 2 . 6 µl Opti-MEM and added onto the cells following the manufacturer′s protocol . Four h after addition of the complex , 50 µl of fresh RPMI medium , supplemented with 30% FCS , 3% pen/strep , 3% non-essential amino acid , 3% glutamine and 3% HEPES were added to the cells . Two d after siRNA transfection , cells were infected with 104 P . berghei sporozoites/well . Twenty-four h after infection , cells were fixed with 4% paraformaldehyde ( PFA ) in PBS and permeabilized with 0 . 2% saponin in PBS . Cell nuclei were stained with Hoechst-33342 ( Molecular Probes/Invitrogen ) , filamentous actin was stained with Phalloidin AlexaFluor488 ( Molecular Probes/Invitrogen ) , EEFs were detected using the mouse monoclonal antibody 2E6 and an AlexaFluor555 labeled goat anti-mouse secondary antibody ( Molecular Probes/Invitrogen ) . Plates were acquired with a Discovery1 automated fluorescence microscope ( Molecular Devices Corporation , CA , USA ) using a 10× lens . In each well , cell nuclei , actin and EEFs were imaged in 9 fields covering a total area of 2 . 7×2 . 0 mm . Image data was analyzed using a custom MetaMorph ( Molecular Devices Corporation , CA , USA ) based algorithm extracting the following values for each imaged field: cell proliferation as measured by the number of nuclei per imaged field ( Hoechst staining ) , cell confluency as measured by the percentage of the imaged field covered by actin staining and number of EEFs as number of compact , high contrast objects in a size range from 16 to 150 µm2 . Within each field , the number of EEFs was normalized to the cell confluency . Normalized EEF numbers and number of nuclei were averaged between the 9 imaged fields within each well . Mean and standard deviations were calculated for each experimental triplicate . For gene-specific expression in vitro , total RNA was isolated from Huh7 cells 48 h post-transfection ( Invitek Invisorb 96-well plate kit ) and converted into cDNA ( ABI's HighCapacity cDNA reagents ) with random hexamers , following the manufacturer recommendations . qRT-PCR used the SybrGreen method with Quantace qPCR mastermix at 11 µl total reaction volume , containing 500 nM of the target-specific primers , and primers that were designed to specifically amplify a fragment of the selected genes . Real-time PCR reactions were performed on an ABI Prism 7900HT system . Relative amounts of remaining mRNA levels of RNAi targets were calculated against the level of RPL13A or 18S rRNA , as housekeeping genes . Remaining mRNA levels of RNAi-treated samples were compared with those of samples transfected with Negative unspecific siRNA . RPL13A-specific primer sequences were: 5′-CCT GGA GGA GAA GAG GAA AGA GA-3′ and 5′-TTG AGG ACC TCT GTG TAT TTG TCA A-3′ . 18S rRNA-specific primer sequences were: 5′-CGG CTT AAT TTG ACT CAA CAC G-3′ and 5′-TTA GCA TGC CAG AGT CTC GTT C-3′ . For infection determination in vivo or ex vivo , total RNA was isolated from livers or primary hepatocytes using Qiagen's RNeasy Mini or Micro kits , respectively , following the manufacturer's instructions . The determination of liver parasite load in vivo , was performed according to the method developed for P . yoelii infections [36] . Livers were collected and homogenized in denaturing solution ( 4 M guanidine thiocyanate; 25 mM sodium citrate pH 7 , 0 . 5% sarcosyl and 0 . 7% β-Mercaptoethanol in DEPC-treated water ) , 40 h after sporozoite injection . Total RNA was extracted using Qiagen's RNeasy Mini kit , following the manufacturer's instructions . RNA for infection measurements was converted into cDNA using Roche's Transcriptor First Strand cDNA Synthesis kit , according to the manufacturer's protocol . The qRT-PCR reactions used Applied Biosystems' Power SYBR Green PCR Master Mix and were performed according to the maunufacturer's instructions on an ABI Prism 7000 system ( Applied Biosystems ) . Amplification reactions were carried out in a total reaction volume of 25 µl , containing 0 , 8 pmoles/µl or 0 , 16 pmoles/µl of the PbA 18 S- or housekeeping gene-specific primers , respectively . Relative amounts of PbA mRNA were calculated against the Hypoxanthine Guanine Phosphoribosyltransferase ( HPRT ) housekeeping gene . PbA 18 S- , mouse and human HPRT-specific primer sequences were 5′- AAG CAT TAA ATA AAG CGA ATA CAT CCT TAC – 3′ and 5′ - GGA GAT TGG TTT TGA CGT TTA TGT G – 3′ and 5′ – TGC TCG AGA TGT GAT GAA GG – 3′ and 5′ – TCC CCT GTT GAC TGG TCA TT – 3′ and 5′ – TGC TCG AGA TGT GAT GAA GG – 3′ and 5′ – TCC CCT GTT GAC TGG TCA TT – 3′ , respectively . For PKCζ mRNA level determination by qRT-PCT , PKCζ-specific primers were used ( RT2 qPCR Primer Assay for Mouse Prkcz , SuperArray Bioscience Corporation ) . Inhibition of PKCζ was carried out by incubation of the cells with a myristoylated PKCζ peptide ( myr-SIYRRGARRWRKLYRAN ) , whose sequence corresponds to that of a pseudosubstrate inhibitor of the enzyme . A myristoylated scrambled peptide ( myr- RLRYRNKRIWRSAYAGR ) was used as a control in these experiments . In order to determine the specificity of the pseudosubstrate inhibitor , Huh7 cells were incubated overnight with either scrambled or pseudosubstrate peptides and then harvested in modified RIPA buffer ( 150 mM NaCl; 50 mM Tris , pH 7 . 5; 1% Triton X100; 50 mM NaF; 1 mM Na3VO4; complete EDTA-free protease inhibitor cocktail ) . After migration on a 10% Tris-glycine gel , proteins were transferred to a nitrocellulose membrane ( BIO-RAD ) , which was probed with anti-phospho-PKC ( pan ) ( βII Ser660 ) ( Cell Signaling Technology ) or anti-phospho-aPKC ( Thr555/PKCι; Thr560/PKCζ ) ( Upstate ) plus HRP-conjugated anti-rabbit ( Amersham ) . The membrane was developed with the SuperSignal West Pico Chemiluminescent Subtrate ( Pierce ) . To further evaluate the specificity of the pseudosubstrate inhibitor towards PKCζ versus PKCι , Huh7 cells were transfected ( Lipofectamine 2000 , Invitrogen ) with plasmids encoding GFP-PKCζ or GST-PKCι . Forty eight hours after transfection the cells were incubated with either scrambled or pseudosubstrate peptides for 1 hour and then harvested as before . The relative expression levels of GFP-PKCζ and GST-PKCι were determined by probing the membrane with anti-aPKCζ ( C20 , Santa Cruz Biotechnology ) , which recognizes the two isoenzymes . The % of inhibition of PKCζ versus PKCι was calculated from the anti-phospho-aPKC signals . All signals were normalized to those of actin . FACS analysis at 2 and 24 after sporozoite addition was performed to determine the percentage of parasite-containing cells and parasite-GFP intensity within infected cells . For infection level measurement at 2 h , 1 mg/ml Dextran tetramethylrhodamine 10 , 000 MW , lysine fixable ( fluoro-ruby ) ( Molecular Probes/ Invitrogen ) was added to the cells immediately prior to sporozoite addition . Cell samples for FACS analysis were processed as previously described [20] . C57Bl/6 mice ( male , 6–8 weeks ) were treated with a single intravenous ( i . v . ) administration of 5 mg/kg of siRNA formulated in liposomal nanoparticles ( Alnylam ) . Three different modified siRNAs targeting PKCζ were used: siRNA#1 – 5′-GGGAcAGcAAcAAcuGcuudTsdT-3′; siRNA#2 – 5′-GGccucAcAcGucuuAAAAdTsdT-3′; siRNA#3 – 5′-cccuuAAcuAcAGcAuAuGdTsdT-3 . A modified siRNA targeting luciferase was used as control ( 5′- cuuAcGcuGAGuAcuucGATsT-3′ ) . Lower case letters represent 2'OMe nucleotides and “s” represents phosphorothioate linkage . Thirty-six h after siRNA administration mice were infected by i . v . injection of 2×104 P . berghei sporozoites . Remaining PKCζ mRNA levels , parasite load in the livers of infected mice were determined by qRT-PCR 40 h after sporozoite injection , 76 h after siRNA administration . Infection of mice treated with one PKCζ siRNA was allowed to proceed onto the blood stage and parasitemia ( % of infected red blood cells ) was measured daily . The PKCζ protein level in the liver of siRNA-treated mice was determined by Western blot . PKCζ protein level in the liver of mice treated with a PKCζ siRNA was quantified by Western blot using the primary antibody ( rabbit anti-PKCζ ( C20 ) : sc-216 , Santa Cruz Biotechnology ) and normalised against actin level detected using rabbit anti-actin ( A2066 , Sigma ) . Anti-rabbit horseradish peroxidase-conjugated ( NA934V , GE Healthcare , UK Ltd . ) was used as secondary antibody . The membrane was developed using the ECL Western Blotting Analysis System , according to the manufacturer's instructions ( Amersham Bioscience , Germany ) . Signal quantification was performed using the ImageJ software package ( NIH , USA ) . For samples in which n>5 , statistical analyses were performed using unpaired Student t or ANOVA parametric tests . Normal distributions were confirmed using the Kolmogorov-Smirnov test . For samples in which n<5 , statistical analyses were performed using Kruskall-Wallis or Wilcoxon non-parametric tests . p<0 . 05 was considered significant , p<0 . 001 was considered highly significant . | During a mammalian malaria infection , Plasmodium sporozoites injected by an infected mosquito travel to the liver where they invade hepatocytes and multiply into thousands of new parasites . These newly formed merozoites are then released into the bloodstream where they infect red blood cells and cause the symptoms of the disease . Although asymptomatic , the liver stage of malaria is an obligatory step in the parasite's lifecycle and constitutes an appealing target for prophylatic intervention . The marked tropism of sporozoites for hepatocytes suggests the latter may provide the parasite with a molecular environment that it can exploit to its own benefit . The identification of host factors that influence hepatic infection can thus provide clues for potential anti-malarial strategies . To this end , we carried out an RNA interference screen of the entire human kinome and associated signaling molecules and assessed the effect of knockdown of their expression in the infection of a human hepatoma cell line by Plasmodium . This strategy identified at least 5 kinases whose down-regulation leads to a marked decrease in infection . Further characterisation of one of these proteins , PKCζ , confirmed that it plays a role in infection by influencing the parasite's invasion of the host liver cells . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"infectious",
"diseases/neglected",
"tropical",
"diseases",
"microbiology/parasitology",
"cell",
"biology/cell",
"signaling"
] | 2008 | Kinome-Wide RNAi Screen Implicates at Least 5 Host Hepatocyte Kinases in Plasmodium Sporozoite Infection |
Although numerous quantitative trait loci ( QTL ) influencing disease-related phenotypes have been detected through gene mapping and positional cloning , identification of the individual gene ( s ) and molecular pathways leading to those phenotypes is often elusive . One way to improve understanding of genetic architecture is to classify phenotypes in greater depth by including transcriptional and metabolic profiling . In the current study , we have generated and analyzed mRNA expression and metabolic profiles in liver samples obtained in an F2 intercross between the diabetes-resistant C57BL/6 leptinob/ob and the diabetes-susceptible BTBR leptinob/ob mouse strains . This cross , which segregates for genotype and physiological traits , was previously used to identify several diabetes-related QTL . Our current investigation includes microarray analysis of over 40 , 000 probe sets , plus quantitative mass spectrometry-based measurements of sixty-seven intermediary metabolites in three different classes ( amino acids , organic acids , and acyl-carnitines ) . We show that liver metabolites map to distinct genetic regions , thereby indicating that tissue metabolites are heritable . We also demonstrate that genomic analysis can be integrated with liver mRNA expression and metabolite profiling data to construct causal networks for control of specific metabolic processes in liver . As a proof of principle of the practical significance of this integrative approach , we illustrate the construction of a specific causal network that links gene expression and metabolic changes in the context of glutamate metabolism , and demonstrate its validity by showing that genes in the network respond to changes in glutamine and glutamate availability . Thus , the methods described here have the potential to reveal regulatory networks that contribute to chronic , complex , and highly prevalent diseases and conditions such as obesity and diabetes .
Genetic linkage and association studies have the power to establish a causal link between gene loci and physiological traits . These studies can make novel connections between biological processes that would not otherwise be predictable based on current knowledge . The pace of gene discovery has greatly accelerated in recent years , and numerous quantitative trait loci ( QTL ) influencing disease-related phenotypes have been identified through gene mapping and positional cloning . While it has become relatively straightforward to map a phenotype to a broad genomic region , identification of the individual gene ( s ) responsible for the phenotype remains difficult . Consequently , only a few percent of the many QTL that have been mapped have had their underlying gene ( s ) identified [1]–[7] . Another limitation of traditional QTL mapping is that it is based on association with a physiological phenotype , but often does not reveal the molecular pathways leading to that phenotype . One way to uncover molecular mechanisms of disease states is to broadly expand the types of phenotypes analyzed in genetic screens . For example , with microarray technology , one can measure the abundance of virtually all mRNAs in a segregating sample . Importantly , mRNA abundance shows sufficient heritability in outbred populations and experimental crosses to allow mapping of gene loci that control gene expression , termed expression QTL ( eQTL ) [8] , [9] . When eQTL co-localize with a physiological QTL , one can hypothesize a shared regulator and offer a potential pathway leading to the physiological trait [9] , [10] . The pathway between a QTL and a physiological trait often involves changes in the steady-state levels of metabolic intermediates , in addition to changes in mRNA abundance . These metabolites can correlate with the genetic , transcriptional , translational , post-translational , and environmental influences on phenotype [7] , [11] . Moreover , metabolites are intermediates in signaling pathways that can regulate gene expression . For example , fatty acids act as ligands for several of the PPAR nuclear hormone receptors , bile acids activate FXR in liver , and diacylglycerol regulates protein kinase C [12]–[14] . Metabolite abundance reflects a biological response to exogenous and endogenous inputs , and when investigating pathways from genotype to phenotype , metabolites can provide a powerful complement to gene expression data and give novel insights into disease pathogenesis mechanisms [7] , [11] , [15]–[25] . Our laboratories have begun to apply targeted metabolic profiling to study mechanisms underlying obesity-induced diabetes [15]–[20] , but have not yet attempted to integrate these methods with genotyping and transcriptional profiling . This has included the application of gas chromatography/mass spectrometry ( GC/MS ) and tandem mass spectrometry ( MS/MS ) for measurements of acyl-carnitine , organic acid , amino acid , free fatty acid , and long and medium-chain acyl-CoA metabolites in tissue extracts and bodily fluids . Herein , we have applied these methods to measure various metabolites in liver samples from mouse strains that differ in susceptibility to obesity-induced diabetes . C57BL/6 ( B6 ) leptinob/ob mice are obese but essentially resistant to diabetes , whereas BTBR leptinob/ob mice are severely diabetic [22] . In an F2 cohort derived from these parental strains , we have shown that the range of blood glucose , insulin levels , and body weight exceeds that of either the C57BL/6 ( B6 ) leptinob/ob or BTBR leptinob/ob parental strains . We went on to identify several diabetes-related QTL in this F2 sample [21] , [22] . In the current study , we focused on a subset of 60 F2 mice that have previously been evaluated in detail with regard to liver gene expression profiles [24] to ask if the abundances of hepatic metabolic intermediates would show sufficient heritability to enable us to map metabolic QTL ( mQTL ) . Because we previously performed mRNA expression profiling on liver samples from this F2 sample , we were also able to investigate the potential for integrative analysis of the expression profiling and metabolite data sets . We show that liver metabolites do map to distinct genetic regions , thereby demonstrating that tissue metabolite profiles are heritable . In addition , we show that mQTL co-localize with eQTL , suggesting common genetic regulators . Finally , as a proof of principle of the practical significance of this multi-disciplinary approach , we illustrate the construction of a specific causal network that links gene expression and metabolic changes , and demonstrate its validity by targeted gene expression analysis .
We determined the concentration of 67 liver metabolites , comprised of 15 amino acids and urea cycle intermediates , 45 acyl-carnitines , and 7 organic acids ( TCA cycle intermediates and related metabolites ) in the F2 sample . The specific analytes are summarized in Table S1 . We created a correlation matrix of all pairwise comparisons among individual metabolites . Unsupervised hierarchical clustering revealed several “hot spots” of highly correlated metabolites ( Figure 1 ) . It is striking that several hot spots correspond to the biochemical pathway to which the metabolites belong . For example , 12 of the 15 amino acids cluster in this matrix . Moreover , when we consider pairwise correlations between all amino acids , 75% had absolute correlation coefficients greater than 0 . 5 ( p<0 . 01 ) ( Table S2 ) . Permutation analysis of these pairwise correlations confirm that the 15 amino acids correlate as a functional group ( p<0 . 001 ) . Several specific acyl-carnitine derivatives are also clustered , such as hexadecadienoyl carnitine ( C16∶2 ) , 3-hydroxy-tetradecanoyl carnitine or dodecenedioyl carnitine ( C14∶1-OH/C12∶1-DC ) , and 3-hydroxy-palmitoleoyl carnitine or cis-5-tetradecenedioyl carnitine ( C16∶1-OH/C14∶1-DC ) . The fact that metabolites of a common functional group are highly correlated suggests that there are potential regulators of these biochemical pathways segregating in this F2 sample . In another cluster , pyruvate correlates most highly with alanine ( r = 0 . 53 , p<0 . 01 ) , and also with lactate and tiglyl carnitine ( C5∶1 ) ( p<0 . 01 ) . Alanine and short-chain acyl-carnitines are products of peripheral protein and fatty acid catabolism , respectively , and are delivered to the liver . The liver uses alanine , along with pyruvate and lactate , as gluconeogenic substrates and rapidly interconverts these metabolites through transamination and oxidation/reduction . The clustering of these metabolites based on their relative concentration in F2 animals suggests that static metabolic profiling can be used as a marker for changes in flux through certain metabolic pathways . All metabolite-metabolite correlation coefficients are listed in Table S2 . It has been demonstrated that mRNA abundance , as determined with microarray technology , is sufficiently heritable to map QTL [7] , [8] , [10] , [23]–[27] . Lan et . al . showed that using expression mapping , specifically in this F2 intercross , can uncover mechanisms that explain correlations between specific transcripts [8] . We therefore sought to determine if metabolite abundance , as measured in F2 liver samples by mass spectrometry , was similarly heritable . If so , resulting metabolic QTL ( mQTL ) could be integrated with expression QTL ( eQTL ) to form network models of gene expression that might ultimately help to explain diabetes susceptibility and resistance in the BTBR leptinob/ob and B6 leptinob/ob strains , respectively [28] , [29] . We found that individual metabolites mapped to specific regions of the genome . By permutation analysis , 21% of the metabolites map significantly to genomic regions ( LOD>5 . 0 , p<0 . 05 ) , indicating those genomic regions could potentially influence ( either directly or indirectly ) the abundance of these metabolites . We used LOD threshold of 3 . 0 to investigate both major and minor putative mQTL where groups of metabolites map . Figure 2 displays a heat map , with metabolites organized by hierarchical clustering as in Figure 1 . The twelve amino acids that clustered based on correlation ( citrulline , tyrosine , and alanine are the exceptions ) map to common mQTL , e . g . , an overlapping region of chromosome 9 . Amino acids that act together in specific pathways show additional common mQTL . For example glx ( glutamine+glutamate ) and urea cycle intermediates arginine , asx ( asparagine+aspartate ) , and ornithine , map to a common region of chromosome 7 . The gluconeogenic substrates alanine and pyruvate have a mapping profile distinct from the majority of amino acids in that they lack the prominent mQTL on chromosome 9 ( Figure 2 ) . This unique alanine/pyruvate mQTL may explain why alanine clusters with pyruvate rather than the amino acids in the correlation matrix ( Figure 1 ) . The foregoing results demonstrate that metabolites of a functional class often are correlated with one another and have common mQTL . To better understand how gene expression and metabolites are related , we adopted the approach used by Carrari [30] and created a correlation matrix between liver metabolites and selected liver transcripts of our 60 F2 mice . Three categories of transcripts were chosen , based on gene ontology terms relating to the biological process in which they play a role: 1 ) carbohydrate metabolism ( glucose metabolism , gluconeogenesis , glycolysis , carbohydrate biosynthesis , TCA cycle , glucose transport , and glycogen metabolism ) ; 2 ) lipid metabolism ( fatty acid biosynthesis , fatty acid oxidation , steroid metabolism , cholesterol metabolism and biosynthesis , and lipid biosynthesis ) ; and 3 ) protein metabolism ( urea cycle , amino acid biosynthesis , protein catabolism , and amino acid transport ) . We organized the metabolites into functional classes to reveal whether biochemical groups of metabolites correlated in a specific pattern with transcripts of a particular pathway ( Figure 3 ) . We found evidence for correlations among functionally similar metabolites and transcripts when organized by biological process . For example , several long-chain acyl-carnitine species show a positive correlation with groups of transcripts involved in glycolysis , fatty acid biosynthesis , steroid metabolism , cholesterol metabolism , and lipid biosynthesis . In contrast , a subset of medium-chain acyl-carnitines and short chain acyl-carnitines exhibit a negative correlation to these same individual transcripts . These findings are consistent with recent studies from our laboratories showing that long-chain acyl-carnitines accumulate in muscle of animals with diet-induced obesity at the expense of short-chain acyl-carnitines , and that this abnormality is resolved when obese animals are exercised [17] . The 15 amino acids displayed a common correlation pattern with mRNA transcripts in pathways of protein metabolism , as well as glycolysis , the TCA cycle , and several lipid metabolism transcripts . These amino acids are very tightly correlated with one another , leading us to investigate the role played by individual transcripts in control of amino acid abundance . Our data show that two very highly correlated metabolites often correlate with the same set of individual transcripts . However , we also see that within this metabolite group , subsets of amino acids will have a unique transcript correlation pattern ( Table S3 , Table S4 ) . For example , thirteen of fifteen amino acids correlate ( r>0 . 35 , p<0 . 01 ) with Slc38a3 , a sodium-dependent transporter that mediates entry of a select group of amino acids across the plasma membrane . There are pathways by which the few known Slc38a3 amino acid substrates ( alanine , asparagine , histidine , and glutamine ) could serve as precursors for biosynthesis of non-substrate amino acids that also correlate with this transporter [31] , [32] . In contrast , only valine and leucine+isoleucine correlate as highly ( r>0 . 35 , p<0 . 01 ) with Ppargc1a mRNA , and could represent a unique metabolic pathway involving the branched-chain amino acids . One hypothesis that follows from our results is that unique genetic regulators could affect the abundance of clusters of metabolites . Unlike mRNA transcripts , metabolites can be interconverted with other metabolites , generating a cluster to which the precursor metabolite will be highly correlated [33] . The downstream product metabolites will also be correlated with the regulatory transcript and co-map with the eQTL of the regulatory transcript [7] , [34] . Glutamate is a substrate and product in amino acid catabolic and biosynthetic pathways . Glutamate can act either as an ammonium donor or acceptor in transamination reactions ( via α-ketoglutarate ) and the glutamate dehydrogenase reaction , and can also be rapidly synthesized from glutamine via glutaminase , thus providing precursor metabolites for the generation of other organic acids and amino acids . Glutamine can also act as a signaling molecule to alter expression of urea cycle and gluconeogenic enzymes [35]–[39] . Given that glutamine and glutamate ( glx ) can generate a network of related metabolites and can also change gene expression , we focused on glx as the start-point for building a proof-of-principle causal network from the F2 liver expression and metabolite profiling data sets . We generated a network featuring glx and a limited number of transcripts that passed multiple , stringent selection filters ( see materials and methods ) . This provided a testable network that would enable us to gain insights into metabolite-transcript relationships . Transcript nodes of the network are highly correlated to glx ( p<0 . 05 by 10 , 000 permutations ) as well as other amino acids ( Table 1 , Table S4 ) . Table 1 depicts the overlap of the glx mQTL interval and the physical location of the transcripts or their eQTL encompassing a 1 . 5 LOD support interval around LOD peaks that are at least 3 . 0 [40] , [41] . We note that glx is correlated with mRNA of two transporters: sodium-dependent amino acid transporter Slc38a3 and glutamate transporter Slc1a2 , whose genes are located on chromosomes 9 ( 102 . 5 Mb ) and 2 ( 107 . 5 Mb ) , respectively . Additionally , the glx mQTL on chromosome 9 spans a region containing Slc38a3 and the mQTL on chromosome 2 and 9 overlaps with the eQTL of Slc1a2 ( Table 1 ) . We hypothesize that both Slc1a2 and Slc38a3 could mediate the entry of glx into liver cells , but that Slc1a2 may also have expression regulated by glx abundance . Table 1 also shows that glx is significantly correlated to argininosuccinate synthetase 1 ( Ass1 ) , arginase 1 ( Arg1 ) , phosphoenolpyruvate carboxykinase 1 ( Pck1 ) , isovaleryl coenzyme A dehydrogenase ( Ivd ) and alanine∶glyoxylate aminotransferase ( Agxt ) mRNAs . The physical location and/or mapping location of these transcripts with respect to the glx mQTL indicates that the metabolite-transcript relationship may go beyond correlation . For example , on chromosome 2 , we see that the glx mQTL co-maps with the eQTL for Agxt , Arg1 , Ass1 , and Ivd [41] . This is consistent with network models in which the QTL regulates glx , which then regulates gene expression or conversely , the QTL regulates mRNA abunance of the four transcripts , which then regulate glx [9] . Using the method described by Chaibub et al . ( in review ) , we generated a causal network consisting of glx and these highly correlated transcripts ( FDR = 0 . 014 ) , incorporating mQTL and eQTL to determine directionality between the nodes ( Figure 4 ) . This network model predicts that modulation of glutamine and/or glutamate levels should lead to a change in the expression of Agxt , Arg1 , and Pck1 . To test this prediction , we isolated hepatocytes from lean B6 and BTBR parental strains and measured changes in gene expression as a result of addition of 10 mM glutamine to the cultured cells . Glutamine exposure changed transcript abundance , and no transcript-specific strain differences in glutamine effect on gene expression were found ( p = 0 . 53 ) ( Figure 5 ) . Glutamine significantly increased expression of Agxt , Arg1 , Pck1 , and Ass1 in both strains ( p<0 . 05 for both strains ) ; the increases in Pck1 and Ass1 confirm prior studies [36]–[39] . Given its role as a glutamate transporter , it is not surprising that Slc1a2 is upstream of glx in the best proposed causal network ( BF = 163 ) ( Figure 4 , solid lines ) . However , glutamine exposure in vitro reduced Slc1a2 expression in isolated hepatocytes from either mouse strain , supporting the second-best causal network solution ( Figure 4 , dotted lines ) . Glutamine also reduced Ivd expression in the B6 strain but showed no effect in the BTBR strain , despite Ivd being upstream of glx in our best causal network . Our causal network predicts Slc38a3 should be unchanged by glutamine treatment . Our hepatocyte experiments confirm this prediction ( Figure 5 ) . Argininosuccinate lyase ( Asl ) , which is neither correlated nor co-maps with glx , served as a negative control and indeed was not altered by glutamine treatment .
Genomics , transcriptomics , proteomics , and metabolomics have delivered large arrays of data , allowing one to correlate physiological states with patterns of gene expression , protein levels , and metabolite abundance . A major challenge in the analysis and interpretation of this data is delivering models of causation from correlations [9] , [42] . Mouse models of diabetes provide a unique method for exploring correlation structure since metabolic dysregulation creates a window for simultaneous application of multiple “omic” technologies . We have previously shown that diabetes traits show strong heritability in an F2 intercross between the diabetes-resistant C57BL/6 leptinob/ob and the diabetes-susceptible BTBR leptinob/ob mouse strains . We assume that the disease phenotype is brought about by a complex pattern of gene expression changes in key tissues [21] , [22] . However , we also recognize the complexity inherent in discriminating the gene expression changes that cause diabetes from those that occur as a consequence of the disease . For example , many genes are known to be responsive to elevated blood glucose levels [43] . Through correlation alone , it is difficult to distinguish these “reactive” genes from ones that are “causal” for the disease . We have taken advantage of the high heritability of mRNA abundance phenotypes , and via microarray technology , have mapped gene loci controlling gene expression at the genome-wide level [8] . This establishes at least one node in a network simply because genetic variation leads to changes in gene expression and not vice versa . However , it does not establish whether the link between a locus and a phenotype is direct or via multiple steps and pathways [27] , [44] . The purpose of the current study was to explore the possibility that the levels of metabolites in tissues are sufficiently heritable in an F2 intercross to provide significant linkage signals , leading to metabolic QTL . Given that many pathways converge upon common metabolites and that these pathways have multiple controllers , any one genetic locus may not alter metabolite levels significantly , and therefore may not be identified as a metabolite QTL . Nonetheless , in our F2 sample , we found significant linkage signals , including some that are quite strong ( e . g . tyrosine: LOD>7 , p<0 . 005; chromosome 2 ) . Our results reveal that metabolites can be mapped to distinct genetic regions , much like mRNA transcripts . Although QTL mapping in an F2 sample does not provide sufficient resolution to identify individual genes with high certainty , it can yield novel information about regulatory networks . Phenotypes mapping to the same locus can be hypothesized to be co-regulated by that locus . With our definition of “phenotype” now including transcripts , metabolites , and physiological traits , we can begin to devise relationships between these phenotypes and genetic regions . This F2 study provides evidence of co-regulation of biologically related pathways . An example is the correlations we found between amino acids and short-chain acyl-carnitine derivatives . These findings are consistent with our understanding of metabolic physiology . In a catabolic , “glucose starved” state , muscle degrades proteins and delivers amino acids to the liver for glucose production . The liver transaminates amino acids to corresponding α-keto acid gluconeogenic substrates . Alpha-ketoglutarate is often the α-keto acid acceptor for these transaminase reactions , generating glutamate as a product . Glutamate , which can also be generated from glutamine in the glutaminase reaction , is then deaminated to produce ammonia by glutamate dehydrogenase , to be fixed through the urea cycle . Additionally , hepatic fatty acid oxidation and amino acid catabolism yield even and odd-numbered short-chain acyl CoAs , which can be used for fuel and for production of ketone bodies . These short-chain acyl-CoA species are readily converted to the cognate carnitine esters , which we have profiled by MS/MS in this study . The amino acid metabolites provide the most striking evidence of functional clustering . We see in both the correlation matrix ( Figure 1 ) and the genetic linkage data ( Figure 2 ) that the majority of amino acids group together . However , a subset of the amino acids , asx , glx , arginine , and ornithine uniquely map to chromosome 7 . Our data predict that these metabolites are driven by different genetic regulators , leading to a unique mapping signature , even within a group of highly correlated metabolites . The C/EBP transcription factors have been shown to alter expression of enzymes acting in the urea cycle and gluconeogenic pathway [45]–[51] , and the C/EBPα isoform is encoded on chromosome 7 . Although we cannot determine that metabolites are mapping to the same individual genes , we can identify genetic regions that coordinate groups of metabolites and transcripts and contain plausible candidate genes . The relationship between mRNA transcripts and metabolites , however , can be bi-directional . Our network identifies a specific metabolite , glx that regulates gene expression . This is consistent with previous studies where glutamine alone increases hepatic expression of argininosuccinate synthetase and phosphoenolpyruvate carboxykinase , but when combined with other essential amino acids , alters additional transcripts of urea cycle and gluconeogenic pathways [36]–[38] , [52] . Our work extends these prior observations by showing that glutamine also changes expression of Agxt , Arg1 , Ivd , and Slc1a2 , but does not alter Slc38a3 , despite the positive correlation with this transcript . The combination of pathway construction based on transcriptional and metabolic profiling and direct model testing in living cells provides evidence for a new pathway by which glx can regulate a key gluconeogenic enzyme . Future studies will be needed to investigate if this pathway is perturbed in development of diabetes . The glutamine induced reduction in Slc1a2 expression was unexpected given that this glutamate transporter is upstream of glx in the best-proposed causal network ( Figure 4 , solid lines ) . Slc1a2 mRNA abundance , however , maps in trans ( to a locus distinct from the physical location of the gene ) to chromosome 9 , its eQTL overlapping with the glx mQTL . It is therefore possible that glutamine could regulate Slc1a2 , as indicated by the second causal network ( Figure 4 , dotted lines ) . Several studies have shown that Slc1a2 expression in astrocytes is reduced by increased ammonia [45]–[47] , [51] , [53]–[55] . Despite the positive correlation between Slc1a2 and glx in vivo , the glutamine-treated hepatocytes produce ammonia via glutaminase , and could decrease expression of hepatic Slc1a2 in vitro . We also did not predict altered expression of Ivd , an enzyme of leucine oxidation . It is interesting to note that Ivd is a case where a gene maps both in cis ( to the locus containing the Ivd gene ) and in trans , here overlapping with the glx mQTL on chromosomes 2 and 13 . Studies have shown that glutamine has an inverse relationship with leucine oxidation , and this could be mediated by glutamine-induced decreased Ivd expression [48] , [50] . We show that the combined use of eQTL and mQTL , with correlations allows one to derive a network and establish data-driven hypotheses about metabolite and gene expression relationships . For example , glycine and serine are the two amino acids most highly correlated with glx , and the transcript most highly correlated with glx is Agxt ( Table 1 , Table S2 ) . Indeed , in our experiments , Agxt was upregulated by glutamine . We hypothesize that the upregulation by glx of Agxt is one mechanism by which glx is correlated with glycine and serine since Agxt catalyzes the transamination of glyoxalate to form glycine , which can then be converted to serine . In further support of this hypothesis , in the F2 sample , serine and glycine correlate ( r>0 . 5 , p<0 . 01 ) to Agxt . The concurrent use of transcriptomics and metabolomics is not limited to one biochemical pathway . For example , the correlation between amino acids and transcripts of carbohydrate and lipid metabolism might reflect a broader signaling function of amino acids beyond pathways of protein metabolism . Furthermore , this correlation , co-mapping , and causal network analysis can uncover roles for transcripts of unknown function . We note Riken clones and ESTs are among the transcripts highly correlated to individual metabolites ( Table S3 ) . By incorporating these transcripts of unknown function as nodes into causal networks , along with transcripts from known pathways , we may infer the functions of these previously unidentified mRNA species . In conclusion , this study shows that metabolites , in addition to transcripts and physiological traits , can be mapped to genetic regions , providing a powerful tool to establish connections between genetic loci and physiological traits . The groups of metabolites and transcripts that are correlated or co-map to physiological traits in our F2 sample may offer insight into metabolic pathways that are causal or reactive to diabetes pathology .
BTBR , B6 , and B6-ob/+ mice were purchased from The Jackson Laboratory ( Bar Harbor , ME ) and bred at the University of Wisconsin . The lineage and characteristics of the BTBR strain have been reviewed by Ranheim et al . Mice were housed in an environmentally controlled facility ( 12-hour light and dark cycles ) and were weaned at 3 weeks of age onto a 6% fat diet ( Purina; #5008 ) . Mice had ad libitum access to food and water , except for 4 hour fasting periods before blood draws and killing ( by CO2 asphyxiation ) . Plasma glucose levels were measured using a commercially available kit ( 994-90902; Wako Chemicals ) . Plasma insulin levels were measured by radioimmunoassay ( RI-13K; Linco Research ) . The facilities and research protocols were approved by the University of Wisconsin Institutional Animal Care and Use Committee . Sixty F2 leptinob/ob mice ranging in age from 13 to 26 weeks were genotyped as previously described [22] . Mapmaker/EXP was used to compile genotype data into framework map . Liver RNA was arrayed as described in Lan et . al [8] . Ten to 12 week old male and female F2 leptinob/ob mice were killed by CO2 asphyxiation after a 4-h fast . Total RNA from sixty F2 mice using RNAzol reagent ( Tel-Test ) and was further purified using an RNeasy kit ( Qiagen ) . The sample labeling , microarray hybridization , washing , and scanning were performed according to the manufacturer's protocols ( Affymetrix ) . Labeled cRNA was prepared and hybridization assay procedures including preparation of solutions were carried out as described in the Affymetrix GeneChip Expression Analysis Technical Manual . A total of 60 MOE430A and MOE430B arrays were used to monitor the expression levels of approximately 45 , 000 genes or ESTs . The distribution of fluorescent material on the array was obtained using G2500A GeneArray Scanner ( Affymetrix ) . Microarray Suite ( MAS ) version 5 . 0 and GeneChip Operating Software ( GCOS ) supplied by Affymetrix was used to perform gene expression analysis . Expression levels of all the transcripts were estimated using the RMA algorithm [49] . Amino acids , acyl-carnitines and organic acids were measured using stable isotope dilution techniques [15] , [18] , [56] . Amino acids and acyl-carnitine species were measured using flow injection tandem mass spectrometry and sample preparation methods described previously [15] , [56] . Briefly , samples were equilibrated with a cocktail of internal standards , de-proteinated by precipitation with methanol , aliquoted supernatants were dried , and then esterified with hot , acidic methanol ( acyl-carnitines ) or n-butanol ( amino acids ) . The data were acquired using a Micromass Quattro micro TM system equipped with a model 2777 autosampler , a model 1525 µ HPLC solvent delivery system and a data system controlled by MassLynx 4 . 0 operating system ( Waters , Milford , MA ) [15] , [56] . Organic acids were quantified using a previously described method that utilizes Trace GC Ultra coupled to a Trace DSQ MS operating under Excalibur 1 . 4 ( Thermo Fisher Scientific , Austin , TX ) [18] . Sixty-seven liver metabolites were measured , comprised of 15 amino acids and urea cycle intermediates , 45 acyl-carnitine derivatives , and 7 organic acids ( TCA cycle intermediates and related analytes ) . The specific metabolites are listed in Table S1 . All MS analyses employed stable-isotope-dilution . The standards serve both to help identify each of the analyte peaks and provide the reference for quantifying their levels . Quantification was facilitated by addition of mixtures of known quantities of stable-isotope internal standards from Isotec ( St . Louis , MO ) , Cambridge Isotope Laboratories ( Andover , MA ) , and CDN Isotopes ( Pointe-Claire , Quebec , CN ) to samples , as follows: Acyl-carnitine assays–D3-acetyl , D3-propionyl , D3-butyryl , D9-isovaleryl , D3-octanoyl , and D3-palmitoyl carnitines; Amino acid assays–15N1 , 13C1-glycine , D4-alanine , D8-valine , D7-proline , D3-serine , D3-leucine , D3-methionine , D5-phenylalanine , D4-tyrosine , D3-aspartate , D3-glutamate , D2-ornithine , D2-citrulline , and D5-arginine; Organic acid assays–D3-lactate , D3-pyruvate , 13C4-succinate , D2-fumarate , D4-glutarate , 13C1-malate , D6-alpha-ketoglutarate , and D3-citrate . In addition to mass , analytes are identified on the basis of the particular MS/MS transitions that we monitor for each class of metabolites . For example , all acyl-carnitine methyl esters produce a fragment m/z 99 . We make the assumption that all even mass precursors ions of m/z 99 are acyl-carnitines to which we assign plausible molecular structures . We differentiate isobaric structures e . g . , dicarboxylic and hydroxylated acyl-carnitines , by comparing of MS/MS spectra for precursors of m/z 85 butylated acyl-carnitine species . We can infer whether the original compound had one or two carboxyl groups on the basis of the mass change from methyl to butyl esters . Given our sample size , we initially analyzed metabolite abundance by hierarchical clustering using the distance function 1-correlation [40] , [57]–[60] . Pairwise Spearman correlation coefficients of r>0 . 254 and r>0 . 330 reflected p-values p<0 . 05 and p<0 . 01 , respectively . To test whether the 15 amino acids are significantly correlated as a group , groups of 15 metabolites were permuted 1 , 000 times and the percentage of pairwise correlations exceeding 0 . 5 was recorded for each group . The fifteen amino acids cluster significantly as a group based on 1 , 000 permutations ( p<0 . 001 ) . Detection and mapping of QTL was performed as previously described [8] , [22] . Briefly , genotypes of 512 F2 mice at 293 markers were assembled using MAPMAKER/EXP [61] . A previously established subset of 60 mice with transcript data was used for expression QTL analysis [24] . Interval mapping methods adjusted for sex as implemented in R/qtl [62] were used to compute linkage to the traits of interest and to investigate mode of inheritance . The traits included the 45 , 265 probe sets surveyed by microarray analysis , and the 67 liver metabolites assayed by MS methods . We used standard interval mapping implemented in R/qtl to map each of the transcripts and liver metabolites at 1-cM resolution with age as additive covariates and sex as both additive and interactive covariates [62] . A LOD threshold of 5 . 0 is required to reach a level of p<0 . 05 in this data set with sample size 60 based on 10 , 000 permutations . We used threshold of 3 . 0 in order to highlight genetic regions to which groups of metabolites map . To visualize regions of mQTL co-localization in highly correlated metabolites ( Figure 2 ) , we constructed heat maps where metabolites are ordered as in hierarchical clustering using 1-correlation , as in Figure 1 . When mice with the B6 allele at a marker have greater levels of metabolites on average than mice with the BTBR allele at that marker , the LOD score at that marker is multiplied by −1 . This adjustment allows us to visualize whether the B6 or BTBR allele results in elevated metabolite abundance . Hepatocytes from 10-week lean male and female BTBR and B6 parental strain mice ( n = 5 for each genotype ) were isolated by liver perfusion [63] . Hepatocytes were seeded at subconfluency ( 3 . 5 × 106 cells/6 well plate ) in low glucose DMEM ( GIBCO ) supplemented with FBS ( 10% vol/vol; GIBCO ) , pen/strep antibiotic ( 1% , GIBCO ) , glutamine ( 2 mM; GIBCO ) , and pyruvate ( 1 mM; GIBCO ) . Cells were left to attach for 3 hours in an incubator at 37°C , 5% CO2 . After a wash with PBS , the cells were treated with unsupplemented DMEM ( Sigma ) with 1 g/L glucose , pen/strep ( 1% ) , and +/− 10 mM glutamine . Cells were treated for 24 hours . RNA was extracted from hepatocytes using RNeasy kits ( Qiagen ) after treatment described above . Hepatocytes in 6-well plates were homogenized in 0 . 35 ml of RLT buffer and stored at −80 C . RNA was purified using RNeasy-mini columns ( Qiagen ) according to the manufacturer's directions . The ratio of the optical densities from RNA samples measured at 260 and 280 nm was used to evaluate nucleic acid purity and total RNA concentrations were determined by the absorbance at 260 nm . The quality of total RNA was estimated based on the integrity of 28S and 18S rRNA separated using 1% agarose gel electrophoresis . Gene expression was measured using a 7500 fast real-time PCR system ( Applied Biosystems ) . cDNA was synthesized from 1 ug of total RNA using the SuperScriptIII first-strand cDNA synthesis kit ( Invitrogen ) primed with a mixture of oligo-dT and random hexamers . Primers were obtained from Integrated DNA Technologies and MWG Biotechnology . The SYBR Green PCR core reagent kit ( Applied Biosystems ) was used to determine relative expression . The housekeeping gene Actb was used as a normalization control . Primer sequences and gene accession codes for transcripts of the glx network are provided in Table S5 . Causal networks were constructed using the methods of Chaibub , et al . ( in review ) . Although the network has the ability to accommodate 100 or more transcripts , we chose a limited number of transcripts passing several selection filters . The transcripts for the glx network were derived from the top 250 most correlated transcripts ( p<0 . 002 ) according to the WebQTL software ( www . genenetwork . org ) . A hypergeometric test was performed and identified the GO term category “metabolism” as one of the two processes significantly enriched by these correlates ( p<0 . 004 ) . Transcripts were chosen from this category , with an additional requirement being that they have at least one eQTL overlapping with the glx mQTL ( Table 1 ) . QTL in the genetic region encompassing a 1 . 5 LOD support interval around LOD peaks that are at least 3 . 0 are also included [41] , [42] . Based on 10 , 000 permutations for each of the transcripts , the LOD threshold is significantly higher to reach significance ( LOD>5 . 0 is required for p<0 . 05 ) , but the 3 . 0 threshold was used include major and minor putative QTL [8] , [24] . If more than one probe set was used to identify a transcript of interest , only probe sets with a grade A annotation on Affymetrix were considered . For these probe sets , only those with all eleven oligonucleotides aligning ( via BLAST ) to their appropriate target sequence provided by the National Center for Biotechnology Information ( www . ncbi . nlm . nih . gov ) were considered acceptable . If more than one primer set still identified the transcript , an average of the probe sets in the network . We built an undirected dependency graph ( UDG ) of order 6 with glx and these transcripts as nodes with a two-tailed significance level of 0 . 05 [64] . We remove edges that are based on spurious or partial correlations , and then orient causal edges between all pairs of connected phenotypes using associated multiple QTLs to break likelihood equivalence . Quantitative trait loci for glx and the selected transcripts were identified with R/qtl [62] using a 3 . 0 LOD cutoff; the marker closest to each peak provided key information for inferring causal direction . We oriented phenotype edges using our QTL-directed dependency graph ( QDG ) approach . For any two phenotypes connected by an edge , the direction LOD score was computed by regressing these phenotypes on each other and on their respective multiple QTLs , adjusting for age and for QTL-sex interactions , and by other phenotypes that might be directly connected to either phenotype by an UDG edge . For each edge , we evaluate a LOD score comparing the two possible orientations and we orient the edge in favor of the direction with the higher likelihood in the ratio . P-values for the direction of the edges were computed using 10 , 000 permutations . Our QDG algorithm used random starts to converge to possible solutions . The best solution was determined by an approximate Bayes factor ( BF ) [65] , [66] . A detailed materials and methods section describing the construction of causal networks is provided in Supporting Protocols ( Protocol S1 ) . We estimated network parameters from the true data and simulated synthetic data according to the causal network in Figure 4 . We simulated 1 , 000 realizations from the causal network and for each edge , we recorded the percentage of undirected edges recovered by the UDG algorithm and the percentage correctly inferred direction by the orientation steps of the QDG algorithm . Overall , the average percentage of true recovered edges was 75% and the average percentage of correctly inferred direction was 83% . False edges were detected at a rate below 2% . To calculate the false discovery rate for the network , we simulated 1 , 000 data sets from the true network . For each data set , the UDG algorithm was used to infer the network topology , and computed the fraction of false edges ( those detected that do not exist in the true network ) relative to the total number edges detected by the UDG algorithm . The FDR for the network topology , computed as the average fraction for these 1 , 000 simulations , is 0 . 014 . The fold changes relative to the untreated hepatocytes for each animal were calculated . An overall ANOVA analysis was performed with gene transcripts nested within subject; interest focused on gene transcript effects and possible gene transcript differences between strains . This analysis showed that glutamine-induced expression change differed by gene ( p<0 . 0001 ) . Significant overall gene transcript effects allows separate transcript-specific paired t-tests between the difference in delta CT values of untreated and glutamine induced gene expression ( relative to Actb ) in each strain separately . Statistics on these data were analyzed with Prism software version 4 . 02 ( Graph Pad Software ) and the aov command in R ( www . r-project . org ) . | Although numerous quantitative trait loci ( QTL ) influencing disease-related phenotypes have been detected through gene mapping and positional cloning , identifying individual genes and their potential roles in molecular pathways leading to disease remains a challenge . In this study , we include transcriptional and metabolic profiling in genomic analyses to address this limitation . We investigated an F2 intercross between the diabetes-resistant C57BL/6 leptinob/ob and the diabetes-susceptible BTBR leptinob/ob mouse strains that segregates for genotype and diabetes-related physiological traits; blood glucose , plasma insulin and body weight . Our study shows that liver metabolites ( comprised of amino acids , organic acids , and acyl-carnitines ) map to distinct genetic regions , thereby indicating that tissue metabolites are heritable . We also demonstrate that genomic analysis can be integrated with liver mRNA expression and metabolite profiling data to construct causal , testable networks for control of specific metabolic processes in liver . We apply an in vitro study to confirm the validity of this integrative method , and thus provide a novel approach to reveal regulatory networks that contribute to chronic , complex , and highly prevalent diseases and conditions such as obesity and diabetes . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genetics",
"and",
"genomics/genomics",
"genetics",
"and",
"genomics/gene",
"expression",
"genetics",
"and",
"genomics/complex",
"traits",
"genetics",
"and",
"genomics/bioinformatics",
"genetics",
"and",
"genomics/disease",
"models",
"diabetes",
"and",
"endocrinology/obesity",
"genetics",
"and",
"genomics/genetics",
"of",
"disease",
"biochemistry/bioinformatics",
"molecular",
"biology/bioinformatics",
"diabetes",
"and",
"endocrinology",
"genetics",
"and",
"genomics",
"cell",
"biology/gene",
"expression",
"diabetes",
"and",
"endocrinology/type",
"2",
"diabetes"
] | 2008 | Genetic Networks of Liver Metabolism Revealed by Integration of Metabolic and Transcriptional Profiling |
When mitochondrial respiration or ubiquinone production is inhibited in Caenorhabditis elegans , behavioral rates are slowed and lifespan is extended . Here , we show that these perturbations increase the expression of cell-protective and metabolic genes and the abundance of mitochondrial DNA . This response is similar to the response triggered by inhibiting respiration in yeast and mammalian cells , termed the “retrograde response” . As in yeast , genes switched on in C . elegans mitochondrial mutants extend lifespan , suggesting an underlying evolutionary conservation of mechanism . Inhibition of fstr-1 , a potential signaling gene that is up-regulated in clk-1 ( ubiquinone-defective ) mutants , and its close homolog fstr-2 prevents the expression of many retrograde-response genes and accelerates clk-1 behavioral and aging rates . Thus , clk-1 mutants live in “slow motion” because of a fstr-1/2–dependent pathway that responds to ubiquinone . Loss of fstr-1/2 does not suppress the phenotypes of all long-lived mitochondrial mutants . Thus , although different mitochondrial perturbations activate similar transcriptional and physiological responses , they do so in different ways .
Mitochondria generate most of the cell's energy as well as its reactive oxygen species ( ROS ) , and mitochondrial dysfunction can cause disease and accelerate aging . Paradoxically , mitochondrial dysfunction can also increase longevity . Yeast petite mutants , which lack mitochondrial DNA and do not carry out respiration , have an increased replicative lifespan [1] . In C . elegans , two types of mutations that affect mitochondrial function also increase lifespan . The first type reduces respiration substantially . One such mutant , isp-1 ( qm150 ) , was identified in an EMS screen for mutants with delayed development and defecation rates . These animals harbor a mutation in an iron-sulfur protein in complex III of the electron transport chain and have reduced rates of oxygen consumption [2] . In addition , two independent RNA interference ( RNAi ) longevity screens revealed that knock-down of genes encoding components of the respiratory chain or ATP synthase decreased ATP production and rates of respiration , reduced behavioral rates and increased lifespan [3] , [4] . Respiratory-chain RNAi-treated animals are smaller than isp-1 mutants [3] , [5] , implying either a more severe reduction in respiration or , conceivably , a qualitatively different response . Interestingly , a mutation that reduces the level of the respiratory-chain component cytochrome c oxidase extends the lifespan of mice [6] , suggesting that the underlying mechanism may be conserved in higher animals . The second type of mitochondrial mutant is exemplified by clk-1 mutants , which are also long lived and have reduced behavioral rates [7] . clk-1 mutants lack a mitochondrial hydroxylase necessary for synthesis of ubiquinone , a prenylated benzoquinone required for shuttling electrons from complexes I and II to complex III during respiration [8] . Oxidative phosphorylation measurements in isolated mitochondria have shown that clk-1 mutations reduce electron transport between complex I and III , but not between complex II and III [9] . In yeast , the clk-1 homologue COQ7 is necessary for respiration , and coq7 mutants are unable to grow on non-fermentable carbon sources . In contrast , C . elegans clk-1 mutants are not only viable , but they have nearly normal levels of respiration and ATP [10] , [11] . clk-1 mutants compensate for the lack of endogenous ubiquinone , Q9 ( the subscript refers to the number of isoprene units ) with bacterial Q8 , provided in their diet [12] , [13] . In the absence of clk-1 , the animals accumulate the Q9 precursor demethoxyubiquinone ( DMQ9 ) . There is some debate over what role DMQ9 plays in the clk-1 phenotypes , but the data suggest that they are caused by the absence of Q9 [13] , [14] , [15] . Mice with reduced levels of Mclk-1 are also long lived [16]; though curiously , these mice do not have reduced Q9 levels [17] . How do these mitochondrial mutations extend lifespan ? Because respiration is the major source of ROS , which could potentially accelerate aging , a simple explanation for the increased longevity of animals with reduced respiration is that they generate less ROS as the animal ages . However , timed RNAi experiments indicate that respiratory-chain activity must be reduced during development for lifespan extension . Adult-only RNAi treatments reduce ATP levels and slow behavioral rates but do not extend lifespan [3] , [18] . If reducing mitochondrial respiration extended lifespan by reducing the level of ROS produced during the aging process itself , then one might expect reducing respiration at any time would extend lifespan . In addition , lifespan extension does not correlate with resistance to the oxidative stressor paraquat [4] or levels of protein carbonylation [18] . Likewise , little is known about the mechanism by which clk-1 mutations , which have relatively small effects on respiration , extend lifespan . Overexpression of clk-1 shortens lifespan and increases movement rates in C . elegans [11] suggesting that whatever the mechanism , its influence on longevity and rates of living is rate limiting in the animal . In previous studies , mutations that suppress the slowed defecation phenotype ( but not other defects ) of clk-1 mutants have been found [19] . One of these suppressor genes , dsc-1 , has been cloned and found to a encode homeodomain protein [20]; however , in general their mode of action is not well understood . Several interesting studies have associated a clk-1 germ-line phenotype with altered ROS signaling , and showed that sod-1 mutations can partially suppress clk-1 mutant's developmental delays [21] , [22] . This is suggestive that clk-1 mutations may alter ROS signaling in cells . How might mutations affecting respiration and ubiquinone biosynthesis slow behaviors and extend the lifespan of C . elegans ? One possibility is that these perturbations trigger a transcriptional response that alters the animal's physiology and lifespan . In yeast , loss of mitochondrial DNA is known to induce a robust transcriptional response . This change in gene expression has been called the “retrograde response” , because it implies a reversal in the normal direction of information flow between the mitochondria and nucleus [1] , [23] , [24] . The genes expressed during the yeast retrograde response lead to a metabolic remodeling of the cell , heat-shock resistance , and increased mitochondrial biogenesis [25] , [26] , [27] , [28] . The retrograde response has been shown to be required for the increased longevity of these so-called yeast “petites” . Thus , lifespan extension in these yeast cells is actively regulated , and not simply a passive consequence of decreased respiration . A gene expression profile similar to the yeast retrograde response has been observed in cultured mammalian cells when mitochondrial DNA is depleted using ethidium bromide , suggesting that this transcriptional response has been conserved evolutionarily [28] . The retrograde response may be a compensatory reaction to the normal decline in mitochondrial function seen with age , since it is observed in older cells [1] . Whether it could potentially play a role in longevity determination in multicellular organisms is not known . Consistent with this possibility , C . elegans isp-1 mutants have increased levels of expression of at least one protective gene , the superoxide dismutase sod-3 [2] . In this study , we carried out microarray analysis of C . elegans mitochondrial mutants to test the hypothesis that a transcriptional response to mitochondrial perturbation slows the animal's rates of behavior and aging .
clk-1 mutants are enigmatic because they exhibit a respiration-defective behavioral and longevity phenotype without having major changes in respiration [16] . For gene expression profiling , we grew synchronized populations of clk-1 ( qm30 ) mutants and wild-type ( N2 ) animals and collected them as pre-fertile adults . One thousand seven genes ( listed in Supplementary Table 5 ) were found to be differentially expressed and were ranked using the SAM ( Significance Analysis of Microarrays ) tool [29] using a False Discovery Rate ( FDR ) of ∼0 . 1 as cut-off . Interestingly , the majority of the genes in this group ( 99% ) were up-regulated relative to wild type , as was also the case in yeast petites [25] , [26] . The genes were assigned to Gene-Ontology ( GO ) categories using the software BiNGO [30] , and several GO categories were found to be overrepresented ( Figure 1 ) . The nature of these categories suggested that clk-1 mutants undergo significant metabolic reorganization and , in addition , activate a stress response similar to that elicited by xenobiotics . For example , GO categories 6006 , 6007 , and 6096 include genes involved in glycolysis; GO categories 16835 and 44275 include genes potentially involved in glycolysis , gluconeogenesis or anaplerotic pathways; GO category 32787 contains genes involved in anaplerotic reactions ( which generate Krebs cycle intermediates ) ; GO categories 9072 , 9074 and 30170 encompass genes involved in amino acid metabolism; GO categories 6825 and 5375 include genes involved in Cu transport; GO category 6629 includes genes involved in lipid metabolism; GO category 46040 includes genes involved in nucleotide metabolism; and GO categories 4499 and 16758 include genes involved in xenobiotic response and maintenance of cellular redox state . When we looked at individual genes , we observed up-regulation of genes encoding enzymes required for glycolysis , such as GPD-2 , GPD-3 , ( gliceraldehyde 3-phosphate dehydrogenase ) , T05D4 . 1 ( aldolase A homologue ) , and LDH-1 ( lactate dehydrogenase ) . GEI-7 , which is an enzyme necessary for the glyoxylate cycle in worms ( see Discussion ) , was also up-regulated . We also observed increased expression of an isocitrate dehydrogenase , C30G12 . 2 , likely involved in the Krebs cycle , and other alcohol dehydrogenases ( dhs-29 , dhs-3 ) that could potentially act in anaplerotic pathways . We found increased expression of proteins involved in oxidative phosphorylation , such as asg-2 ( subunit of ATP synthase complex ) and F17A9 . 4 ( NADH oxidoreductase ) . Also , several genes coding for enzymes involved in amino acid and nucleotide metabolism were up-regulated in clk-1 mutants . There was also a significant increase in enzymes involved in cellular detoxification , including UDP-glycosyl transferases ( UGT-53 , UGT-13 , UGT-43 , UGT-6 , UGT-39 ) , gluthathione S-transferases ( GST-4 , GST-13 , GST-36 ) , superoxide dismutase ( SOD-3 ) , flavin-containing monooxygenases ( FMO-1 , FMO-3 ) and other gene classes potentially involved in xenobiotic metabolism ( cytochrome P450s , alcohol dehydrogenases and ABC transporters ) . To compare the transcriptional profile of clk-1 mutants to the yeast retrograde-response genes , we referred to two previous publications studying the effects of inhibiting mitochondrial respiration in yeast petites [25] , [26] . We applied BiNGO software to the most differentially expressed genes reported in those studies and compared the yeast-petite GO categories to those of clk-1 mutants . Among the top ranking GO categories ( p<0 . 1 ) , we observed a remarkable degree of similarity ( p<0 . 001 ) between the clk-1 and yeast petite BiNGO categories ( Figure 1 ) . One of the hallmarks of the yeast response to respiration inhibition is an increase in mitochondrial biogenesis [31] . To test whether there might be an increase in mitochondrial biogenesis in clk-1 mutants , we used Real-Time qPCR to quantify mitochondrial DNA . Total mtDNA was measured relative to genomic DNA , which provided the average number of mitochondrial genomes per cell in a population of worms ( see Methods ) . We observed a significant increase in mitochondrial DNA levels ( Figure 2 ) . Together , these data suggest that clk-1 mutants exhibit a response to mitochondrial dysfunction that is similar to the yeast retrograde response . To identify genes necessary for the increased longevity of clk-1 ( − ) animals , we compiled a list of differentially expressed genes from populations of L4 ( late larval stage ) as well as prefertile adults ( see Methods ) . We included L4 animals in this set because previous work has shown that the L4 period is the critical period for lifespan determination in at least some mitochondrial mutants [18] ( and data not shown ) . We picked the 75 top differentially expressed genes ranked using the SAM software package , inhibited their functions using RNAi and measured lifespan . Out of the initial list of 75 , we collected lifespan data for 63 genes over two independent trials ( Table S1 ) . We established a significance cut-off of p<0 . 05 and selected RNAi clones that decreased clk-1 longevity significantly in both trials or were statistically significant in one experiment and showed a decrease of at least 5% in the other ( a 5% decrease in overall lifespan corresponds to an ∼25% decrease in the lifespan extension produced by clk-1 mutation ) . We retested the positive clones in clk-1 ( − ) and wild-type animals for effects on longevity ( Table S2 ) . Out of 63 RNAi clones tested , only two decreased clk-1 mutant longevity in all three trials ( Figure 3A and 3B ) . Neither of these clones significantly shortened wild-type lifespan , suggesting that they may play a role specifically in clk-1 mutant lifespan ( Figure S1 ) . One of these clones corresponded to aqp-1 , which encodes a glycerol channel [32] . aqp-1 RNAi decreased the lifespan extension that would normally be produced by clk-1 mutations from 26% to 7% ( p<0 . 05 ) and from 17% to 0% ( p<0 . 0001 ) , and did not affect wild-type longevity in two separate experiments ( Figure S1 ) . Interestingly , aqp-1 ( also called dod-4 ) has already been shown to contribute to the long lifespan of daf-2 insulin/IGF-1-receptor mutants [33] . The other RNAi clone , corresponding to a gene we call fstr-1 ( for “faster” , also known as gfi-1 ) decreased the lifespan extension produced by clk-1 mutations from 26% to 3% ( p<0 . 01 ) and from 17% to 0% ( p<0 . 0001 ) while not affecting wild-type lifespan in two separate experiments . The effects of aqp-1 RNAi and fstr-1 RNAi on the longevity of clk-1 mutants were tested three times , with consistent results , although the extent of suppression varied between experiments ( Tables S1 and S2 , Figure S1 ) . We examined the genome for the possibility that fstr-1 RNAi might cross-inhibit another gene , and found that the RNAi clone was likely to knock down a close homolog ( with 96% protein sequence identity ) located next to fstr-1 that did not exhibit clk-1-dependent regulation in our microarray analysis . We call this gene fstr-2 , and henceforth we refer to their combined functions , as inferred from RNAi , as fstr-1/2 function . In addition to increased longevity , the most striking phenotypes of C . elegans mitochondrial mutants are their decreased behavioral rates . In principle , these rates could decrease as a direct consequence of impaired mitochondrial function . Alternatively , it is possible that their slowed behavioral rates reflect a regulated response to mitochondrial perturbation; simply speaking , they slow down to conserve energy . To look for genes that slow the behaviors of clk-1 mutants , we inhibited the top 100 up-regulated genes from microarrays of L4 and pre-fertile adults using RNAi and measured time it took for L1 larvae to develop to adulthood . We found that knockdown of fstr-1/2 in clk-1 mutants consistently increased the rate of growth to adulthood ( Figure 4A ) . This phenotype was most striking when the animals were examined 75–80 hours after hatching . At this time , no control clk-1 ( − ) mutants had reached adulthood , whereas 95–100% of the fstr-1/2 RNAi treated animals were adults . fstr-1/2 RNAi treatment also increased the behavioral rates of clk-1 ( − ) animals , as measured by thrashing and pumping ( Fig 4B and 4C ) . These effects were not observed in wild type; in fact , in wild type , knock-down of these genes had the opposite effect , slowing development and decreasing rates of thrashing and pumping . To test whether fstr-1/2 RNAi somehow restored wild-type clk-1 function , we examined ubiquinone profiles . Using HPLC , we observed the expected decrease in UQ9 and increase in DMQ9 in clk-1 mutants , and we observed the same mutant pattern of ubiquinone species in clk-1 mutants subjected to fstr-1/2 RNAi ( Figure 4D ) . Thus , fstr-1/2 RNAi suppresses the phenotypes of animals that still have an altered , Clk-1 ( − ) , pattern of ubiquinone species . This suggests that the wild-type fstr-1/2 gene slows behavior and extends lifespan in response to the changes in ubiquinone produced by clk-1 mutations . Next we asked whether FSTR-1/2 modulates the clk-1 mutant phenotype by influencing the retrograde response . Using real-time qPCR , we looked at the effects of fstr-1/2 RNAi on the expression levels of five of the genes whose expression was most significantly up-regulated in the microarrays: gpd-2 , a glyceraldehyde 3-phosphate dehydrogenase involved in glycolysis; T22B7 . 7 , an Acyl-CoA thioesterase , involved in anaplerotic reactions; dhs-26 , an alcohol dehydrogenase; ugt-43 , an UDP-glucoronosyl transferase; and the aquaporin gene aqp-1 . The qPCR data confirmed the microarray studies , in that all of these genes were up-regulated in clk-1 mutants . We found that fstr-1/2 RNAi significantly and consistently decreased expression of these genes in a clk-1 ( − ) background ( Figure 5 ) but not in wild type ( Figure S2 ) . Thus , the gene expression changes observed in clk-1 mutants are at least partially dependent on fstr-1/2 . Together these findings suggest that in clk-1 mutants , fstr-1/2 ( + ) decreases rates of behavior and extends lifespan by triggering downstream changes in gene expression . Because fstr-1 is up-regulated in clk-1 mutants , we were particularly interested to learn where in the animal it was expressed . To investigate this , we generated transgenic animals expressing the fluorescent protein mCherry under the control of the fstr-1 promoter . We observed strong expression in three neurons located in the head and throughout the intestine , particularly in the anterior intestinal cells ( Figure 6 ) . We identified the three neurons as RIH and I1L/R . RIH is a nerve-ring interneuron of unknown function and I1L/R are pharyngeal interneurons that regulate pharyngeal pumping rates in response to touch and removal of bacteria . We saw the same pattern of expression in the clk-1 ( − ) mutant and wild-type , but the intensity of expression was increased in the mutant , consistent with our qRT-PCR and microarray data . Together these findings suggest that fstr-1 acts in the intestine and/or in specific neurons to slow the rates of aging and behavior in clk-1 mutants . To compare the pattern of gene expression in clk-1 mutants to that of respiration mutants that have more strongly reduced levels of oxygen consumption and ATP , we performed microarray analysis of isp-1 ( qm150 ) and cyc-1 ( RNAi ) animals . ( cyc-1 encodes cytochrome c reductase , which is a component of complex III of the electron transport chain . ) We grew synchronized populations of isp-1 ( qm150 ) mutants and cyc-1 RNAi-treated animals in parallel with wild-type control animals , collected them as young , pre-fertile adults and analyzed the resulting microarray data as described above for clk-1 mutants . The SAM algorithm with a false discovery rate of ∼0 . 1 yielded 814 significant genes for isp-1 ( − ) mutants and 7662 significant genes for cyc-1 ( RNAi ) animals . ( For the complete list , please see Tables S6 and S7 . ) Thus , it seems that cyc-1 RNAi induces a broader transcriptional response than do clk-1 and isp-1 mutations , which correlates with the increased severity of the Cyc-1 ( RNAi ) phenotype . In addition , cyc-1 RNAi-treated animals , when compared to isp-1 and clk-1 mutants , showed increased expression of additional cell-protective genes , including chaperones ( hsp-6 , hsp-70 ) , superoxide dismutases ( sod-4 , sod-3 ) and xenobiotic detoxification enzymes ( ugt-2 , ugt-47 , ugt-36 , gst-8 , gst-22 , gst-24 , dhs-5 , dhs-28 ) . Interestingly , other genes encoding detoxification enzymes were down regulated , possibly implying the deployment of a specific detoxification program . Using BiNGO analysis , we identified several GO categories that were overrepresented in each mutant ( Figure 1 ) . A highly significant fraction of the top GO categories ( p<0 . 1 ) was shared between either two , or all three , of the mutant strains ( p<0 . 001; Figure 1 ) . By chance , one would expect 2 of the 10 GO categories shared by all three mitochondrial mutants and annotated in yeast to also be significant in the yeast petite cells . In contrast , we observed 8 categories in common ( p<0 . 001 ) . In addition to identifying GO categories , we looked for individual genes that were expressed in a similar way in the three C . elegans mitochondrial mutants . We found a highly significant ( p = 7 . 19E-15 ) overlap set of 73 differentially-expressed genes ( Table S3 ) . In addition , we observed an increase in mitochondrial DNA levels in isp-1 mutants and a smaller increase in cyc-1 ( RNAi ) animals that did not reach statistical significance ( p = 0 . 354 ) ( Figure 2 ) . Taken together , these data suggest that the gene expression profiles of different C . elegans mitochondrial mutants are similar to one another and to the yeast retrograde response . Since isp-1 ( − ) , clk-1 ( − ) and cyc-1 ( RNAi ) animals are all long-lived , gene expression patterns that are shared between all three might be particularly likely to contribute to lifespan extension . Double RNAi experiments in C . elegans can be difficult to interpret , so we did not attempt RNAi knockdowns in cyc-1 ( RNAi ) animals . However , we did attempt to knock down the top thirty statistically-significant shared genes , individually , in an isp-1 background . ( Of these , 21 were not present in the clk-1 set we described above , which contained only the top 75 differentially expressed genes ) . We obtained data for all of these genes ( Table S4 ) . Of these , only one RNAi clone , cdr-2 , consistently made our significance cut-off [p<0 . 05 and a 10% decrease in isp-1 mutant longevity , which corresponds to a 50% decrease in the lifespan extension produced by the isp-1 mutation . ] cdr-2 RNAi reduced the lifespan extension produced by the isp-1 mutation from 43% ( control RNAi ) to 26% ( p<0 . 001 ) , while not affecting wild-type longevity ( Figure 3C ) . cdr-2 encodes a member of the glutathione S-transferase family . These enzymes catalyze the conjugation of reduced glutathione to electrophilic centers on different substrates . This activity contributes to detoxification of both endogenous toxins and xenobiotics , suggesting that the increased longevity of isp-1 ( − ) mutants is at least partially dependent on a cellular detoxification response . We note that in one of our three trials , cdr-2 RNAi significantly shortened the lifespan extension produced by clk-1 mutation ( from 21% to 14% ) . This finding suggests that cdr-2 may be involved more generally for lifespan extension in mitochondrial mutants . Given the remarkable reversal of the clk-1 mutant phenotype by fstr-1/2 RNAi , we were interested in examining its function in the respiratory-chain mutants . We found that fstr-1 was significantly up-regulated in isp-1 mutants but , unexpectedly , not in cyc-1 ( RNAi ) animals . Using RNAi , we asked whether fstr-1/2 might influence the behavioral phenotypes of isp-1 ( qm150 ) mutants . We restricted our analysis to lifespan and time to adulthood because isp-1 mutants did not move often enough to provide consistent behavioral rates . We found that the developmental rates of isp-1 mutants were severely decreased in the presence of fstr-1/2 RNAi , leading to developmental arrest of many animals . We found that fstr-1/2 RNAi had no effect on the lifespan of isp-1 mutants that reached adulthood ( Figure 3D ) . Thus the effect of fstr-1/2 RNAi on isp-1 mutants was more similar to the effect of fstr-1/2 RNAi on wild type than to its effect on clk-1 mutants . We wanted to know whether fstr-1/2 was necessary for gene expression changes in an isp-1 ( − ) background , but because isp-1 mutants subjected to fstr-1/2 RNAi grew very slowly and asynchronously , we could not use qRT-PCR . Instead , we assayed gene expression in vivo by introducing the isp-1 mutation into a strain expressing GFP under the control of the gpd-2 promoter , which drives expression of a glycolysis gene that is up-regulated by these mitochondrial mutations ( Figure 7 ) . We found that fstr-1/2 RNAi prevented the up-regulation of this reporter in a clk-1 background but not in an isp-1 background . Together , these data suggest that the Isp-1 ( − ) and Clk-1 ( − ) behavioral and longevity phenotypes are established by distinct mechanisms .
Gene expression profiling of these mutants was quite revealing , because each of their expression profiles exhibited striking similarity to the yeast retrograde response . The yeast retrograde response , which also lengthens lifespan , appears to remodel the cell's metabolism . Without respiration , the Krebs cycle cannot be completed , as succinate cannot be oxidized to fumarate ( Figure 8 ) . This prevents the formation of oxaloacetate ( OAA ) , which in turn decreases the availability alpha-ketoglutarate , which is the precursor of glutamate , an essential metabolite in amino acid metabolism . In order to generate precursors of glutamate , respiration-deficient cells must activate alternative ( anaplerotic ) pathways that supply the mitochondria with OAA and acetyl-CoA [27] . Activation of anaplerotic pathways was observed in respiration-defective yeast [25] , [26] and human cells [28] . Our microarrays reveal transcriptional activation of genes encoding several metabolic enzymes that have roles in anaplerotic reactions , such as the glyoxylate cycle and fatty acid oxidation . The glyoxylate cycle occurs mainly in the peroxisomes and bypasses the succinate-to-fumarate step of the Krebs cycle through the formation of glyoxylate , eventually leading to the formation of succinate , which can be fed back into the Krebs cycle ( Figure 8 ) . We observed increased expression of the gene encoding the major C . elegans glyoxylate-cycle enzyme , GEI-7 , in the three mitochondrial mutants we examined . We did not observe an RNAi phenotype for this clone in the clk-1 mutant; however , because a gei-7 mutant was available , we examined cyc-1 ( RNAi ) ; gei-7 animals and found a large suppression of the cyc-1 ( RNAi ) longevity phenotype , decreasing lifespan extension from 80% to 15% with little effect on wild type ( Figure S4 ) . Malate dehydrogrenases catalyze synthesis of OAA from malate , which is also an important step in recycling Krebs cycle intermediates . F46E10 . 10 encodes a malate dehydrogenase and is significantly up-regulated in all three long-lived mitochondrial mutants we examined . Fatty-acid oxidation provides acetyl-CoA , which feeds into the Krebs cycle by reacting with OAA to form citrate . This pathway is activated in long-lived yeast lacking mitochondrial DNA [27] . We also detected increased expression of several genes that are involved in fatty acid oxidation . clk-1 ( − ) animals exhibited increased expression of acs-2 ( acetyl-CoA synthetase ) and fat-6 ( fatty acid desaturase ) ; and isp-1 mutants exhibited increased expression of T02G5 . 4 ( acetyl-CoA thiolase ) and T05G5 . 6 ( enoyl-CoA hydratase ) . The expression profiles of cyc-1 ( RNAi ) animals , however , contained fewer significant genes involved in fatty acid oxidation , suggesting there may be some differences in metabolic adjustments between different mitochondrial mutants . Interestingly , recent work has shown that long-lived mice with decreased levels of Mclk-1 have an increase in α-ketoglutarate dehydrogenase activity , consistent with an up-regulation of the Krebs cycle [17] . Furthermore , the animals have a decrease in overall NAD levels , which in itself could hamper normal Krebs cycle activity and further decrease glutamate synthesis [26] . In all three long-lived mitochondrial mutants , we observed a significant increase in expression of genes involved in glycolysis . This was expected , since glycolysis becomes a more important source of ATP when oxidative phosphorylation is inhibited . In addition to these metabolic shifts , we also observed increased expression of a significant number of stress response genes in all three mitochondrial mutants , ranging from genes increasing xenobiotic drug resistance to protein chaperones . This is consistent with previous in vitro observations that impairment of electron flow during oxidative phosphorylation is actually likely to generate more ROS [34] , and suggests these animals may be responding to this additional cellular insult . Finally , in all three strains , genes involved in the oxidative phosphorylation process itself were up-regulated , and we observed increased levels of mitochondrial DNA in clk-1 and isp-1 mutants . Thus , apparently C . elegans mitochondrial mutants , like yeast petites [26] , attempt to compensate for reduced levels of respiration . Together these findings indicate that the transcriptional response triggered by conditions that inhibit respiration in C . elegans is similar to that triggered in yeast , and suggest the presence of a conserved underlying mechanism for lifespan extension . While this manuscript was in preparation , Falk et al . reported the gene expression pattern of a mixture of long-lived and short-lived respiration-defective mutants compared to wild type [35] . In the future , it will be interesting to learn whether the expression patterns of short-lived mitochondrial mutants differ from those of the long-lived mutants . The lifespan of one such short-lived mutant , mev-1 ( kn1 ) was increased when respiration was lowered further using respiratory-chain RNAi ( Figure S3 ) , arguing that their short lifespans are not due simply to insufficient respiratory-chain activity . It was interesting to find that clk-1 mutations trigger a conserved transcriptional response even though they only have a mild effect on respiration . There are at least two possible interpretations for this finding . The first is that the transcriptional response to clk-1 mutation need not be triggered by reduced respiration itself , but instead can be triggered by signals that are generally associated with reduced respiration , such as fluctuations in ubiquinone levels . Such fluctuations could have acquired the ability to induce the retrograde response during evolution because they allowed the animal to conserve energy in the face of a perceived energy shortage . Alternatively , perhaps the clk-1 mutation does inhibit respiration more severely initially , but the physiological changes elicited by the retrograde response restore the steady-state level of respiration closer to normal . In yeast , the retrograde response is induced via helix-loop-helix transcription factors that do not appear to be present in C . elegans . When the genes encoding these transcription factors are deleted in yeast , inhibiting respiration does not induce the retrograde response , and lifespan is not extended [1] . Thus , in yeast , the retrograde response likely increases lifespan . Our data suggest that this is the case in this multicellular animal as well . First , inhibiting the activity of at least some of the genes up-regulated in mitochondrial mutants was sufficient to shorten their lifespan without obviously affecting the lifespan of wild type . In particular , the glutathione S-transferase gene cdr-2 was up-regulated in all of the long lived mutants and it contributed to lifespan extension consistently in isp-1 respiration-defective mutants and , at least in some trials , in clk-1 mutants as well . This finding suggests that the prominent cell-protective gene expression response that we observe contributes to longevity . In addition , the metabolic shifts we observed are also likely to influence lifespan , as the longevity of cyc-1 ( RNAi ) animals required the glyoxylate-cycle gene gei-7 , and the longevity of clk-1 mutants was promoted by the glycerol channel aqp-1 . Our failure to observe effects on lifespan with most of the RNAi clones we tested does not necessarily mean that they do not influence lifespan ( though this may be the case ) . It seems possible that many of these genes could act cumulatively to influence lifespan . The lifespan extension of clk-1 and isp-1 mutants was only ∼20% in most experiments , so only perturbations that are fairly strong would be visible in our assays . In general , although we have no reason to discount the importance of genes whose knockdowns produced statistically significant effects on lifespan , because the magnitude of the effects were small , we remain cautious in our interpretation . The second argument for the importance of the C . elegans transcriptional response in the longevity of mitochondrial mutants comes from our studies of fstr-1/2 . fstr-1 was up-regulated in clk-1 mutants , and this gene , and/or its constitutively-expressed homolog fstr-2 , is required , in turn , for a robust transcriptional retrograde response . Knocking down fstr-1/2 activity with RNAi did not suppress the primary ubiquinone defect . However , none of the five up-regulated genes we tested was up-regulated in the presence of fstr-1/2 RNAi . These genes included metabolic as well as cell-protective genes , arguing that fstr-1/2 may be a major regulator of the retrograde response in clk-1 mutants . This restoration of a normal transcriptional profile correlated with a suppression of the behavioral , growth and longevity phenotypes of clk-1 mutants . Together all of these findings support the hypothesis that a conserved mitochondrial retrograde response extends lifespan in metazoans as well as in yeast . We note , however , that fstr-1/2 RNAi had stronger effects on the induction of the retrograde-response genes we tested than it had on the clk-1 behavioral phenotypes . This suggests either that part of the retrograde response is expressed independently of fstr-1/2 , or that mechanisms that do not involve transcription also influence the clk-1 phenotype . How does FSTR-1/2 regulate gene expression ? Little is known about the molecular function of FSTR-1/2 . The predicted FSTR-1 and FSTR-2 proteins contain 21 ET modules , which are domains of unknown function containing 8–10 conserved cysteines predicted to form 4–5 disulphide bridges and a C-terminal putative transmembrane domain . Sequence alignment studies using the BLAST algorithm showed weak similarities to a secreted yeast protein ( AGA1 ) and a predicted mouse membrane protein ( Zonadhesin ) . We also looked for structural homologues of fstr-1/2 using the software package PHYRE [36] and found highly significant predicted structural similarities to portions of ErbB1/2/3/4 . ErbB proteins belong to a highly conserved family of receptor tyrosine kinases that play many roles in cell biology and disease . Members of the ErbB family usually contain an extracellular region ( ∼620 amino acids ) that recognizes and binds ligands , a single membrane spanning region and an intracellular tyrosine kinase domain . However , we did not find a predicted tyrosine kinase domain or a secretion signal in fstr-1/2's sequence , which makes a potential connection to the ErbB family unclear . Finally , FSTR-1 ( originally called GFI-1 , for GEX-interacting protein ) was identified in a two-hybrid screen as a potential binding partner of UNC-68 , a muscle-specific C . elegans ryanodine receptor . Our attempts to find phenotypic similarities or functional interactions between these two genes were unsuccessful ( data not shown ) , so their potential relationship remains unclear . Our finding that FSTR-1 is expressed in the intestine ( which is C . elegans entire endoderm , including its site of fat storage ) , as well as a small number of neurons , raises the possibility that these two tissues are particularly important for the response to mitochondrial perturbation . Interestingly , the glycerol channel aqp-1 is expressed exclusively in the intestine and pharynx [32] , further implicating the intestine in the clk-1 longevity pathway . In the future , it will be very interesting to learn how FSTR-1 and FSTR-2 function at the molecular level to initiate a response to altered ubiquinone levels . Long-lived C . elegans mitochondrial mutants share many phenotypes; however , there are also some significant differences between them . As mentioned above , they differ in their respiration rates , ATP levels and body size . In addition , respiratory-chain RNAi and clk-1 mutations both extend the lifespans of daf-2 insulin/IGF-1 receptor mutants [3] , [7] , whereas isp-1 mutations do not [2] ( DC and CK , unpublished ) . These differences have prompted the question of how similar the C . elegans mitochondrial mutants really are to each other [5] . Our microarray observations suggest that the overall nature of the response is similar , though the specific genes affected and the extent to which they are activated varies between mitochondrial mutants . Perhaps these differences are phenotypically significant; for example , clk-1 mutants may have near normal respiration rates because they can compensate more fully than the other mutants to a primary respiratory-chain defect . On the other hand , there is a clear difference in the regulation of the clk-1 and isp-1 mutant phenotypes , since fstr-1/2 is necessary for the clk-1 mutant phenotypes but not for the isp-1 mutant phenotypes ( Figure 9 ) . It is possible that there exist yet-unidentified regulators that control the transcriptional response to mitochondrial perturbation in all of these mutants , thus unifying their phenotypes . In any case , it will be interesting to learn how the isp-1 and cyc-1 retrograde responses are regulated and at what point these pathways converge to control the same downstream genes .
The strains used in this study were: N2-Bristol ( WT ) , fer-15 ( b26 ) ; fem-1 ( hc17 ) , clk-1 ( qm30 ) [37] , isp-1 ( qm150 ) [2] , gei-7 ( ok531 ) , mev-1 ( kn1 ) [38] , muEx491[pfstr-1::mCherry+podr-1::cfp] , clk-1 ( qm30 ) ; muEx491 , sEx11128[pgpd-2::gfp] , isp-1 ( qm150 ) ; sEx11128[pgpd-2::gfp] . All strains used except for mev-1 ( kn1 ) were outcrossed to our laboratory's wild-type N2 strain 4 times . We constructed microarrays using single-strand DNA oligos representing 20 , 374 unique C . elegans genes . These were purchased from Illumina [39] . Populations were starvation-synchronized as L1s overnight and collected at two different times: as L4s staged based on vulval morphology and as pre-fertile adults soon after the L4-to-adult molt , to guarantee maximum synchronicity between animals that grew to adulthood at different rates . Hybridizations were performed using standard techniques described in [33] . Total RNA was purified using TriZol reagent , mRNA was purified using Oligotex ( Qiagen ) and cDNA was labeled using Cy-dyes prior to hybridization . The chips described are direct comparisons between N2 and either clk-1 ( qm30 ) or isp-1 ( qm150 ) animals; or between fer-15 ( b26 ) ; fem-1 ( hc17 ) animals subjected to control ( vector only ) RNAi versus cyc-1 RNAi . The animals were harvested as L4 larvae or young ( pre-fertile ) adults . We performed four independent biological repeats for each condition with the exception of clk-1 ( − ) and isp-1 ( − ) L4-staged populations , where we only collected data for two biological repeats . Dye-swaps and technical repeats were averaged and analyzed as one biological repeat . Scanning was done using a GenePix 4000B scanner , and initial spot quality check was done using Genepix 6 . 0 software . During the analysis we used two different sets of chips for each mutant: the “combined set” includes a combination of all L4 and adult chips , and the “adult-only set” only includes chips from populations collected as adults . The microarray data were analyzed twice over the several-year period spanned by these studies . The initial analysis , used to generate candidate genes that may be functional in the extended longevity of mitochondrial mutants , was performed on the combined set ( L4 and adults ) of microarrays . In this analysis , standard ratio-based normalization and default program settings for flagging missing or “bad” spots were used in the Acuity 4 . 0 software package . Gene significance was calculated using the SAM software package . All of the lifespan analysis described herein was based on genes at the top of the list when the data were analyzed in this way . The second analysis , used to generate genes for use in the comparative GO analysis , was performed on the adult-only set of arrays . These data were renormalized using lowess as well as ratio-based normalizations . On the assumption that genes within the same operon should in general have similar expression patterns , flagging parameters were adjusted to those that maximized the expression correlation of genes within the same operon on a “training” subset of the arrays . Genes not present in at least three of the arrays were not considered . These data were analyzed using the SAM software package and genes were considered significant below a FDR ( false discovery rate ) of 0 . 1 . An algorithm , which we named the “p-q algorithm” , was designed and implemented in the Python programming language to determine the set of genes that are differentially regulated in all three combined ( L4 and adult ) microarray data sets . The algorithm takes as input a set of microarray hybridization data and estimates the q-value for each gene using the method described in [40] and p-values estimated using Student's t-test . It then iterates through each q-value and calculates the probability of seeing the observed number of genes that would be significant in all three data sets , should that q-value be used as the threshold for significance . The algorithm reports the set of genes that overlap between the three data sets at the q-value cut-off that achieved maximum overlap significance , as well as the probability of seeing such a degree of overlap by random chance . Probabilities are calculated using the hypergeometric distribution when possible , or the Poisson approximation when necessary . GO categories were found using the BINGO software starting from a list of differentially expressed genes obtained from running SAM on the set of adult-only microarrays , with a significance cut-off of FDR< = 0 . 1 for each C . elegans mutant . Yeast GO categories were obtained by analyzing a dataset that we constructed pooling the differentially-expressed genes from two different publications [25] , [26] . Bacterial feeding RNAi experiments were performed as described previously [33] . Clones were picked from Julie Ahringer RNAi library and were all verified by DNA sequencing . Lifespan analysis was conducted as previously described [3] . All assays were done at 25°C unless otherwise stated . The lifespan measurements depicted in Supplementary Table 1 were done in the presence of 20 mM FUDR ( fluorodeoxyuridine ) to inhibit progeny growth . The Stata 8 . 0 software package ( Stata Corporation ) was used for statistical analysis and to calculate means and percentiles . In all cases p-values were calculated using the logrank ( Mantel-Cox ) method . Mitochondrial DNA was quantified using Real Time-qPCR . We used two primer sets for mitochondria DNA graciously provided by Dana Miller from the Roth lab at the Fred Hutchinson Cancer Research Center: Mito1 Forward: GTTTATGCTGCTGTAGCGTG , Reverse: CTGTTAAAGCAAGTGGACGAG; Mito2- Forward: CTAGGTTATATTGCCACGGTG , Reverse: CAATAAACATCTCT-GCATCACC . The results were normalized to genomic DNA using a primer pairs specific for ama-1 and nhr-23: ama-1- Forward: TGGAACTCTGGAGTCACACC , Reverse: CATCCTCCTT-CATTGAACGG; nhr-23 – Forward: CAGAAACACTGAAGAACGCG , Reverse: CGATCTGCAGTGAATAGCTC . Animals were grown and collected as described above for microarray studies and lysed in a standard buffer containing proteinase K for 1 hour at 65°C . qPCR was performed using SYBR GREEN PCR Master Mix ( Applied Biosystems ) . Each comparison pools 5 biological repeats . Results were normalized to wild type using 7300 System SDS Software . Time to adulthood was measured as time ( ±2 hours ) at which 95% of animals reached adulthood . Measurements shown represent pooled data from five independent experiments , error bars represent SEM . Pumping rate was measured as the average number of pharyngeal pumps per minute ( n = 10 ) over three independent trials . Thrashing rate was measured as the average number of body thrashes in M9 buffer in one minute ( n = 10 ) over three independent trials . All measurements were conducted on day three of adulthood . Real-time RT-PCR was carried out using the 7300 Real Time PCR System ( Applied Biosystems , Foster City , CA , USA ) . Primers and probes were designed specifically for each gene using Primer3 software . To generate pfstr-1::mCherry-expressing animals , a pfstr-1::mCherry construct was made using the Invitrogen Gateway Cloning technology . The promoter was amplified from genomic DNA using a primer set obtained from Mark Vidal's online promoterome database ( Forward: ggggacaactttgtatagaaaagttgaggccagctttagataat; Reverse: ggggactgcttttttgtacaaacttgtcatctgaaatttgaatgtgttagt ) . The construct obtained was sequenced and injected as described ( Mello and Fire , 1995 ) at 10 ng/µl into N2 animals to generate a transgenic line ( indicated by muEx491 designation ) . The coinjection marker Podr-1::gfp was injected at 50 ng/µl . | Mitochondrial respiration generates energy in the form of adenosine triphospate ( ATP ) , a molecule that powers many cellular processes . When respiration is inhibited in C . elegans , rates of behavior and growth are slowed and , interestingly , lifespan is extended . In this study , we investigated the mechanism of this response . We find that inhibiting respiration increases the expression of genes predicted to protect and metabolically remodel the animal . This pattern of gene expression is reminiscent of the expression profile of long-lived respiration-defective yeast , suggesting ancient evolutionary conservation . Mutations in clk-1 , which inhibit the synthesis of the respiratory-chain factor ubiquinone , produce gene expression , longevity , and behavioral phenotypes similar to those produced by inhibiting components of the respiratory chain . We find that knocking down the activities of two similar genes—fsrt-1 and fstr-2—accelerates the behaviors and aging rates of clk-1 mutants and inhibits the clk-1 ( − ) transcriptional response . Thus , fstr-1/2 , which encode potential signaling proteins , appear to be part of a mechanism that actively slows rates of growth , behavior , and aging in response to altered ubiquinone synthesis . Unexpectedly , fsrt-1/2 are not required for the longevity and behavioral phenotypes produced by inhibiting the gene isp-1 , which encodes a different component of the respiratory chain . Our findings suggest that different types of mitochondrial perturbations activate distinct pathways that converge on similar downstream processes to slow behavioral rates and extend lifespan . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"genetics",
"and",
"genomics/gene",
"function",
"genetics",
"and",
"genomics/gene",
"expression"
] | 2009 | A Regulated Response to Impaired Respiration Slows Behavioral Rates and Increases Lifespan in Caenorhabditis elegans |
Relationships between spiking-neuron and rate-based approaches to the dynamics of neural assemblies are explored by analyzing a model system that can be treated by both methods , with the rate-based method further averaged over multiple neurons to give a neural-field approach . The system consists of a chain of neurons , each with simple spiking dynamics that has a known rate-based equivalent . The neurons are linked by propagating activity that is described in terms of a spatial interaction strength with temporal delays that reflect distances between neurons; feedback via a separate delay loop is also included because such loops also exist in real brains . These interactions are described using a spatiotemporal coupling function that can carry either spikes or rates to provide coupling between neurons . Numerical simulation of corresponding spike- and rate-based methods with these compatible couplings then allows direct comparison between the dynamics arising from these approaches . The rate-based dynamics can reproduce two different forms of oscillation that are present in the spike-based model: spiking rates of individual neurons and network-induced modulations of spiking rate that occur if network interactions are sufficiently strong . Depending on conditions either mode of oscillation can dominate the spike-based dynamics and in some situations , particularly when the ratio of the frequencies of these two modes is integer or half-integer , the two can both be present and interact with each other .
The brain is a multiscale system , whose dynamics spans from microscale structures , such as ion-channels and synapses , to emergent behavior , such as oscillations at the whole-brain scale . The problem then arises of how to simultaneously incorporate these diverse scales to make predictions about brain dynamics . Neuronal dynamics has most often been studied by starting from single-neuron perspective via Hodgkin-Huxley equations [1] and their many variants for different neural types ( e . g . , [2] , [3] ) , or via idealized models such as integrate-and-fire and binary neurons . Strong nonlinearities are responsible for spiking , with the spike cycle often described in terms of a nonlinear oscillator [4] , [5] . Such approaches have been extremely successful in accounting for neural dynamics at the single- or few-neuron level . Single-neuron approaches can also be applied to networks of many neurons by incorporating their synaptic interconnections . While very large networks can be simulated if sufficient computer power is available [3] , [6] , [7] , the results of brute-force simulations can be difficult to interpret , especially when emergent network-level phenomena are involved . Moreover , common misconceptions that arise from the single-neuron viewpoint sometimes impede understanding of large-scale dynamics . For example , the starting-point picture of spiking being due to a nonlinear oscillator often leads to a focus on coupled-oscillator descriptions of neural interactions . If overemphasized , this can obscure the existence of ( often linear , or near-linear ) collective modes of oscillation in the network , which modulate spike rates at frequencies that are not related to the spike rate itself [8]–[10] — in general , both nonlinear-spiking and collective-oscillation phenomena exist . Some widespread errors in the literature that stem from this standpoint ( when adopted naively ) are: ( i ) that large-scale brain rhythms and electroencephalographic ( EEG ) oscillation frequencies must correspond to spike rates of specific neural “generators” or “pacemakers” , whereas they are quite different from spike rates in general , and ( ii ) that brain rhythms and EEG oscillations must be highly nonlinear because spikes are , whereas collective oscillations that modulate firing rates can actually be linear , or very nearly so [10] . Of course , collective oscillations can also have their own large-amplitude nonlinearities that survive averaging over spike generation , or arise through other effects [11]–[13] . An alternative starting point is to average over neural properties at the outset to obtain a neural field theory ( NFT ) [10] , [14] in which the average dynamics of large numbers of neurons are modeled . In this case , instantaneous local firing rates are tracked , but individual neuronal spike dynamics are not . Such approaches are well suited to studying large-scale phenomena and bridging across scales and are much less computationally intensive than corresponding studies based on direct computation of single-neuron dynamics . However , as noted , they do not directly incorporate spiking dynamics of individual neurons . Two aspects are of particular significance here . One is the internal dynamics of neurons . In this study , this is discussed in terms of a comparison between spike events described by changes in membrane potential ( in the spike-based approach ) and spike rates ( in the rate-based approach ) . Communication between neurons is also critical . In the spike-based approach spikes travel between neurons that are coupled pairwise or via a field that carries spike profiles [14]; in the neural field theory communication is through propagation of fields that carry the spike rate only . In this work , we examine two limiting cases , one in which spiking neurons communicate via spikes , and one in which populations of neurons with rate-based internal dynamics communicate via rates — and make the dynamics as similar as possible in all other respects by having the same type of field carry either spike profiles or spike rates in the respective cases . In other words , one case involves spiking dynamics of neurons coupled by spikes carried by fields , and the second involves rate dynamics of continuous neural matter coupled by rates carried by fields; the fields obey the same propagation equations in both cases . It is important to understand the relationships between the two limiting approaches , especially because they are complementary , not mutually exclusive . It is thus essential to understand when each is appropriate to be used , whether there are phenomena to which both can be applied , and which is the more convenient and tractable in given cases . Moreover , there can be situations where a fuller understanding requires an application of both approaches . This is analogous to situations arising in many other branches of science . For example , the properties of materials can be studied from a molecular viewpoint but , when dealing with large numbers of molecules , statistical approaches or continuum approximations are more convenient and appropriate starting points for obtaining understanding at the scales of most relevance — hydrodynamics is usually studied in terms of fluids , not molecules , for example . Likewise , statistical mechanics of particles passes over into thermodynamics for many applications as the number of particles becomes large , and there are intermediate regimes that can be addressed using either formalism , or variants such as nonequilibrium thermodynamics . Some work toward understanding the complementarity of spiking and mean-field approaches has been done , in part by developing hybrid models that preserve aspects of both single-neuron and mean-field approaches . For example , Robinson et al . [15] and Wu et al . [16] showed how to write the spike rate of Wilson neurons [2] in terms of the spike rate itself ( rather than instantaneous cellular voltages ) , thereby eliminating the need to track individual spikes if rate is all that is desired . This work put the Wilson model of spiking and bursting neurons [2] in a form suitable for incorporation into NFT and allowed top-down systems-level influences on single neurons to be analyzed tractably . The predictions of this NFT were subsequently investigated for a model system incorporating a simple delayed feedback loop whose resonances could interact with natural neural spiking and bursting frequencies [16] . Robinson and Kim have very recently developed a series of hybrid methods of treating neural interactions that combine various aspects of spike- and rate-based neural dynamics and of the discrete vs . mean-field features of spatial coupling [14] . Bressloff and Coombes have shown how fluctuations in firing rates consistent with a neural field model can be produced by a network of integrate-and-fire neurons particularly when slow interactions are present [17] . There are other approaches to neuronal modeling which we mention for completeness . The population density approach ( e . g . [18] ) moves beyond a model based on mean firing rates alone by considering the changes in the distribution of neuron properties . In the population density approach , individual neurons or groups of neurons are not modeled explicitly , rather the change in the probability density function of the state of the neurons is modeled . This approach can be many times faster than a direct Monte Carlo simulation of neurons or groups of neurons . One can also focus on the correlations and higher-order moments of a distribution . This is of significance since correlations in activity may form a significant part of the mechanism through which neuronal signals carry information . Recently , Touboul and Ermentrout [19] have studied the correlation approaches of Bressloff [20] and Buice et al . [21] and shown them to be equivalent when applied to a system of infinite size . This allows large networks of neurons to be analyzed . Significantly , by considering correlations rather that just mean firing rates , dynamical behaviors can appear that cannot be accounted for with a lowest order mean field approach alone . However , the specific aim of this paper is to investigate and elucidate the complementarity between spike- and rate-based approaches to neural dynamics by use of an overarching approach that can accommodate both pictures in the analysis of a test system that is suited to exposing the key phenomena . Although other approaches may also be informative , we focus on the complementarity between spike- and rate-based simulations in the current work . We begin by reviewing the theoretical background and developing our model . We then present the numerical methods , and give the results of our analyses . We analyze , compare and contrast the dynamics of the spike-based and rate-based approaches . Finally , we interpret the results and discuss their applicability and significance . For simplicity , homogeneous models are used; however , the methods discussed are generalizable to inhomogeneous situations .
In this section we briefly outline the NFT equations required , specializing the treatment to a specific , idealized test system . The model we use is that of a single cortical population , driven by an external drive and incorporating direct interactions between neurons and indirect ones via a delayed feedback loop . We consider the system of interconnected neurons which includes synaptic input to a set of neurons ( labeled by a suffix ) from an external set of neurons ( suffix ) . The former set consists of a one-dimensional chain of neurons with periodic boundary conditions , and has a feedback both directly and via a loop , where and are the rates of incoming spikes at each synapse ( i . e . have dimensions of inverse time ) , and is time . The loop features a feedback delay time , and the feedback is assumed to be topographically organized ( i . e . , each point in space feeds back most strongly to itself ) . This idealized system is sufficiently general to study complementarity between rate- and spike-based treatments; it is also easily generalized to include more types of neurons and higher dimensionality [22] , [23] . Biologically , such topographical feedback is found in the thalamocortical loop . Excitatory neurons in the cortex drive the coupled thalamocortical and thalamic reticular neurons of the thalamus; in turn the thalamocortical neurons project back to the cortex in a manner such that a signal returns very close to where it originated [24] . The spiking model is summarized in graphical form in Fig . 1 . When applied to real brain tissue , NFT averages neural properties over linear scales of a few tenths of a millimeter , sufficient to embrace many neurons [10] , [22] . The soma potential , measured relative to its resting potential , responds to spikes via synaptic dynamics , dendritic signal dispersion , and soma capacitance . The resulting response to synaptic input approximately obeys [10] , [22] , [25] , [26] ( 1 ) where ( 2 ) is the mean response rate of to synaptic input , is the mean connectivity strength to neurons of type from those of type , is the corresponding mean number of synaptic connections , and ( with dimensions voltage times time ) is the mean strength of these connections , defined to be the time integral of the postsynaptic potential change due to a spike afferent on a neuron from one of type . Action potentials are produced at the axonal hillock when the soma potential exceeds a threshold [10] , [22] , [25] . When averaged over a local population of neurons , a good approximation to the firing rate is ( 3 ) where , is the maximum firing rate , and and are the population mean and standard deviation of the threshold [10] , [22] , and is the mean soma potential averaged over a local population of neurons . We discuss the origins of this relationship below . Prior work has shown that the mean fields of axonal signals , , , and , propagate approximately as if governed by damped wave equations [22] , [27] , [28] , one form of which is ( 4 ) where ( 5 ) ( 6 ) where , with the axonal velocity , and the characteristic axonal range [22] , [27] . Equations ( 4 ) – ( 6 ) incorporate spatiotemporal coupling between neurons . This is more easily seen through the corresponding Green-function ( i . e . , propagator ) formulation [22] , [26]: ( 7 ) ( 8 ) for , with the Green function satisfying for to ensure causality , and where translation invariance of the system has been assumed . We choose the form ( 8 ) , which follows from ( 5 ) and ( 6 ) , to preserve the timing and shape of narrow pulses in one dimension , since we want to compare spike-based coupling with field-theoretic coupling in the present work . The spatial coupling vs . is found by integrating ( 8 ) over , [22] which gives ( 9 ) Integrating ( 9 ) over yields a normalization of unity , which reflects the fact that each pulse that enters an axon ultimately reaches its end . Equations ( 4 ) – ( 6 ) thus represent signals that propagate along axons at a uniform velocity , but where the number of axons reaching a distance decays exponentially as a function of , with characteristic range . This is a reasonable first approximation to the coupling of cortical neural populations by axons in one dimension . If we were to replace ( 6 ) by , we would recover the form introduced in by Robinson et al . [22] , which yields broader temporal pulses in response to a delta input . The latter form is actually more realistic in general , especially in two dimensions , since axons are neither identical in velocity nor exactly straight , thereby making delta-function propagation of a mean pulse field very much an idealization . Here we retain ( 6 ) – ( 8 ) to obtain ( 9 ) , which is commonly assumed in spike-based analyses . Moreover , this form provides a more stringent test of complementarity with rate-based analyses because it involves no temporal smoothing of the propagated signal , which would tend to make the two cases more similar . Delayed integrodifferential equations such as these have been well studied , both in general and in the context of neuronal modeling [29]–[31] . The presence of delayed feedback leads to Hopf bifurcations and other dynamic phenomena such as traveling waves [32] . We expect to see such features in the models discussed here . We now briefly review the Hindmarsh-Rose fast-spiking neuron model [2] , [33] , [34] , and how it can be put in a form compatible and comparable with NFT . Conductance-based equations for the rate of change in membrane potential in a single fast-spiking neuron , appropriate to the mammalian neocortex , can be written [2] , [14]–[16] , [25] , [33] , [34] ( 10 ) is the capacitance per unit area , is an externally imposed input current per unit area ( e . g . , due to synaptic input from other neurons ) , is a leakage current per unit area , and and are the and currents per unit area , respectively , and is a transient potassium current that enables these neurons to fire at very low spike rates when is small . Note that use of the script font indicates a voltage measured relative to the extracellular fluid ( i . e . , a membrane potential ) rather than a measurement taken relative to the resting state — there is a constant offset between and equal to the resting potential ; i . e . , . Each of the currents is assumed to obey Ohm's law , with ( 11 ) where is the conductivity per unit area and is the equilibrium potential of the ion . Numerous authors have investigated Eqs ( 10 ) and ( 11 ) for fast-spiking neurons , the main population in the mammalian neocortex , and have found simplified expressions for their dynamics , which can be closely approximated by just two equations that gave an adequate description of spiking dynamics [2] , [33]–[36] . There is one equation for the membrane voltage and one for a dimensionless recovery variable that describes the coupled opening of channels and corresponding closure of channels [2] , [34] . The equations are ( 12 ) ( 13 ) where ( 14 ) ( 15 ) with being the reversal potential , the reversal potential , , , , , , , , , and for fast-spiking neurons . The dynamics of ( 12 ) and ( 13 ) have been discussed in detail elsewhere ( e . g . , [2] ) , so we summarize very briefly here . At low they have three steady-state solutions: at these are a stable node with and , an unstable saddle point at somewhat higher , and an unstable spiral point at still higher . The first of these represents the resting ( non-firing ) state . As increases , the two lower fixed points approach one another , then generate a saddle-node bifurcation when they coalesce at the critical current with and . This gives rise to a limit cycle that encircles the resulting spiral point . Each orbit of the limit cycle corresponds to the generation of one spike; hence the picture of spike generation being due to a nonlinear oscillator . The frequency of the limit cycle ( i . e . , the firing rate ) satisfies [5] , [15] , [37] ( 16 ) for and for , which corresponds to a continuous increase from zero firing rate as increases beyond . Simulations show in Eq . ( 16 ) [2] , [15] , [34] . We next show that we can couple individual model spiking neurons together in a way that can be compared directly with NFT of the same system . In NFT , the mean membrane potential of a population of cells is driven by the incoming axonal pulse rate . However , in the spike theory , membrane potential is driven explicitly by current entering the cell body from the dendritic tree . Standard cable equations imply that this current is proportional to minus the spatial derivative of the voltage at the soma boundary . Hence , the functional form of the driving current to a cell induced by a delta function spike at a synapse has the same temporal dependence as , apart from a dimensional constant of proportionality [15] , [25] . Thus , it obeys ( 17 ) where is the time course of the part of the afferent signal that is above the channel opening threshold and is defined in Eq . ( 2 ) . This can be approximated as ( 18 ) in NFT notation , where the quantities ( ) have units of conductance per unit area and the connection strengths have been introduced . These incorporate the membrane conductance per unit area and will be used in the model to control the relative strengths of the direct and loop feedback , and external drive . Henceforth , we discuss the model in terms of connection strengths rather than the . This formulation allows communication between neurons via the intermediate fields ( ) , which can propagate spike profiles , not just average rates , provided we now replace ( 4 ) by ( 19 ) where is a constant which we determine shortly . That ( 19 ) reproduces ( 4 ) can be seen by averaging ( 19 ) over timescales much longer than a spike width . Explicitly , averaging Eq . ( 19 ) over the inter-spike interval ( which varies as a function of space and time ) gives ( 20 ) where the angle brackets denote the average over . Since changes over time-scales much longer than a spike , we can write , leaving ( 21 ) If ( 22 ) then it is clear that ( 23 ) where is the spike rate . Here we have assumed that the integral over is only significant within the vicinity of the spike , i . e . can be caluclated from Eq . ( 22 ) by considering a stereotypical spike profile . We assume ( 22 ) henceforth . In dealing with rates in populations of neurons the idealized square root form ( 16 ) of the response curve discussed earlier must be convolved with a distribution ( e . g . , a Gaussian ) of some width that encapsulates fluctuations in the properties of the neurons and their input: e . g . , variations in number and strength of synaptic connections , and in the various channel conductances , especially from neuron to neuron . Such convolutions smear ( 16 ) over a width [15] . A further source of broadening is fluctuation in arrival rate of spikes and associated changes in membrane voltage [38] . A good approximation that also captures saturation effects is the sigmoidal function ( 24 ) which is equivalent to the rate-voltage relationship ( 3 ) via ( 25 ) where [15] is a conductance per unit area and . We now in a position to write explicitly a set of coupled differential equations for our 1D chain of identical neurons , in a form that is consistent with NFT in the relevant limit . For each neuron at a point in space , we use the Wilson neuron model to describe its membrane potential and recovery variable . We emphasize here that the spikes are carried through a field rather than through pairwise interactions , which corresponds to the neuron-in-cell approach recently introduced by Robinson and Kim [14] . The input to the neuron comes from both synaptic input from other neurons ( through a current term , which is explicitly modeled below ) , and the input from the external drive term , labeled . The set of coupled differential equations is now obtained from ( 12 ) and ( 13 ) for the neural dynamics , ( 17 ) and ( 18 ) for the synaptic dynamics , and ( 19 ) for the propagation of fields along axons . To model a level of random external inputs , a white noise current density term is added to the neural dynamics on a grid , where and with a constant . The resulting equations are ( 26 ) ( 27 ) ( 28 ) ( 29 ) ( 30 ) where is a constant external drive . The variables , , , , and describe the state of the system , with distinguishing the locations of the neurons . Values of constants used in this paper are mostly taken from previous work [2] , [23] , [39] or , in the cases of , , and , numerical analysis of the Wilson model neuron [2] , and are listed in Table 1 . The level of noise , through the parameter , is chosen so that fluctuations are small and linear approximations are valid when used . Here we have separated the direct and loop feedbacks ( 29 ) and ( 30 ) , respectively , to enable the use of , and in general . Before we discuss the numerical implementation of the equations we emphasize that we have not proved that there are well-behaved solutions to these . However , wave equations are well understood physically and numerically; e . g . [22] , [27] , [28] , [40] . Moreover , numerical simulations as discussed below produce results that do not diverge with time . In numerical implementation of the model , the Eqs ( 23 ) – ( 27 ) are discretized on a 1D spatial grid . When spatially discretizing , several issues must be considered: ( i ) we must ask whether the numbers of neurons and system size are sufficiently large to ensure results adequately represent real brain dynamics and are not numerical artifacts . The suitability of the number of neurons can be estimated by asking the question of how many input spikes are needed to generate an output spike . In the human brain this is large , with each neuron receiving input of order 10 spikes per second at each of thousands of synapses . Overall , if the effective soma integration time leading to a spike is s , several hundred presynaptic input spikes contribute to each postsynaptic spike [41] . In our simulations with , each neuron is locally coupled to neighboring neurons in approximately 8 cm of tissue . By placing neurons approximately 0 . 25 cm apart , this gives us locally coupled neurons . Typically in the simulations neurons fire at a rate of spikes per second , giving spikes arriving at each neuron per second , implying that each spike is generated as a result of a neuron receiving input spikes during the relevant integration time . This is much lower than in the cortex; however , computational demands , which scale linearly with the number of neurons , necessitate the use of relatively few neurons . However , we have also carried out some larger runs with considerably more sampled spikes , to begin to explore the effects of relaxing this limitation . No qualitative difference is observed , suggesting that our levels of temporal and spatial discretization are sufficient . ( ii ) We also anticipate that a system that is too small would introduce artifacts: e . g . , with periodic boundary conditions if the system is too small , long-wavelength modes of activity are not captured . Moreover , a model that is of order or smaller in size would be affected by wrap-around of connections through the periodic boundary conditions . However , biologically , it should be remembered that the cortex is not of infinite size; the ratio of to cortical radius is approximately 0 . 61 . A system size of 20 cm is used for most runs; this is adequate in terms of removing numerical artifacts and computer resources and does not represent an implausible size biologically . Some simulations have been carried out with a larger system size and results are not significantly different . Initially , the variables , and are assigned the value zero for all spatial points . The membrane potential and recovery variable are assigned the values they would have at equilibrium when no external current is applied , namely and 0 . 279 respectively . The equations are integrated forward in time with a second-order stochastic predictor-corrector method [40] . In order to generate initial activity a high driving current is applied for the first second of simulation and then removed . The Courant condition requires that the time step must be smaller than the grid spacing divided by the velocity of a pulse to ensure numerical stability [42] . The typical step size of s is comfortably within this limit . We also treat the system of Fig . 1 using the complementary neural field approach of coupling neurons using rate of firing , rather than individual spikes . These rates are propagated using the same Green functions ( and same wave equations ) as for spikes in the spike-based approach , but individual spikes are not tracked . In the NFT approach , each grid point is taken to represent the average dynamics of a local population of neurons . To do this we replace the equations for the membrane potential ( 26 ) and recovery variable ( 27 ) with a single equation that relates the firing rate to the input current , via the square-root function ( 16 ) , whose parameters were calibrated to reproduce the dynamics of fast-spiking neurons in previous work [14]–[16] . This rate is used to provide input to the wave equation , rather than using the potential term explicitly . A small amount of white noise is added to the current , where , , where is a constant . The noise provides a small perturbation to the system to allow it to quickly explore phase space and ensure that no two simulations are identical . Therefore we obtain the following nonlinear set of four coupled equations for the variables , , and . ( 31 ) ( 32 ) ( 33 ) ( 34 ) Numerically , this set of neural field equations can be integrated forward in time using the same approach as for the spike-based case . In this case , the time step can be made larger than for the spike based model , although subject to the Courant condition for numerical stability , because the spike profiles are not modeled explicitly . This is a major advantage of field-based approaches over spike-based approaches . Equation ( 31 ) is appropriate rather than Eq . ( 24 ) since for simplicity we will consider homogeneous parameters . Equation ( 31 ) allows us to compare explicitly the firing rates predicted by the neural-field approach to those of the spike-based approach . If inhomogeneous parameters were used , Eq . ( 24 ) , with values of , and specific to the parameter distribution used , would be appropriate . Equations ( 31 ) – ( 34 ) can be used to compute the firing rate at various points in space and time; i . e . , the mean firing rate of all neurons in the vicinity of each grid point vs . time . Hence , for this to give a good representation of average dynamics , each grid point should correspond to multiple neurons . In the present case , this means the separation of grid points in the neural field model must be much larger than the length scale between neurons , which is satisfied in the present work . Therefore , in carrying out detailed comparisons between the spiking model and the field model , the results of the spiking model need to be coarse-grained ( i . e . , averaged over the appropriate length scale ) . We again emphasize that in this work we have carried out simulations at various length scales and neuron densities , and results are qualitatively unaltered by changing the scale ( i . e . our discretization is fine enough for the purposes of this work ) . It is found that the system ( 31 ) – ( 34 ) has at least one spatially uniform equilibrium state , which is obtained by setting all the temporal and spatial derivatives to zero . In general there may be one or three solutions ( plus a special case of two solutions ) ; however if as in this work , there is only a single solution . Equations ( 33 ) and ( 34 ) can then be solved to obtain equilibrium values , , and so via Eq . ( 32 ) : ( 35 ) Equation ( 31 ) gives ( 36 ) for , and otherwise . Squaring Eq . ( 36 ) and substituting Eq . ( 35 ) for gives a quadratic equation for that is easily solved for the positive firing rate solution . By writing the deviations from their equilibrium values of , , , , and the noise input in terms of their Fourier components in both space and time , we can establish the power spectrum of fluctuations in both temporal frequency and spatial frequency . To calculate these quantities , we linearize Eqs ( 31 ) – ( 34 ) in small deviations from equilibrium , and write the Fourier form ( 37 ) with similar expressions for , , and . We also note that the noise has an equilibrium value of zero and omit the from henceforth . This gives us the linearized equations ( 38 ) ( 39 ) ( 40 ) ( 41 ) These equations can be solved for ( or any other of the variables ) in terms of the noise input to give us , where the transfer function is given by ( 42 ) In general , Eq . ( 42 ) is difficult to analyze further analytically , especially because of the term . However , a useful limiting case can be seen when the system has only loop feedback whose time delay is much longer than the timing of the synaptic current pulses and wave events; i . e . , , and . In this case we can make the approximations and to give us the transfer function at : ( 43 ) The response ( 43 ) will have a resonance when the phase of the complex exponential is a multiple of , which gives resonances at angular frequencies of and its harmonics [16] . When both direct and loop feedback are present , these resonances modulate the combined spectrum to produce peaks , as found originally by Robinson et al . [43] , [44] . It might appear that use of a neural field model , where only spike rates are calculated , might remove all information about individual spike times; however , this is not the case [14] . Neural field theory yields instantaneous spike rates as functions of position and time , so the integral of the local rate over some time period is the expected number of spikes that occur at that location in this time period; i . e . , ( 44 ) Moreover , when the integral increments by one , we know that there must be exactly one spike during this interval , so ( 45 ) defines the expected time at which the next spike occurs , given that one occurred previously at . To construct a membrane potential time series from the spike timings , where is an integer , one can write the potential as a function of the noninteger part of the integral ( 44 ) . Therefore , ( 46 ) where is the integer part of and is a function that describes the spike profile . If required , this profile can be quickly computed from a look-up table [14] . Since this approach requires only the tracking of rather than spike profiles , it is computationally less intense than a spike-based approach , since larger time steps can be used . Implementing this approach requires initial phases to be specified for the neurons , or for enough time to pass that the system loses memory of its initial conditions . Before presenting the results , we emphasize that in all cases fields are used to describe propagation of signals between neurons , either carrying spikes from neuron to neuron through Eqs . ( 29 ) and ( 30 ) or conveying rates fields between spatial locations through Eqs . ( 33 ) and ( 34 ) .
To start , we demonstrate typical behavior of the spike-based state variables , , and . For this illustration we use a small positive loop feedback; i . e . , , . The external drive current is chosen to be equal to the critical current . This external current would put an individual neuron at the point of spiking , so that the positive feedback between neurons ensures that they obtain a modest spike rate; this allows us to explore the interaction between spike-based and collective oscillations . Biologically , it is reasonable that a neural system can organize to be near a critical point [45] , [46] . A very short time delay is used , s so that delays between the direct and delayed feedbacks are negligible compared to the timescales of the dominant neural activity in this case ( i . e . , the interspike interval ) . Fig . 2 shows the membrane potential of each neuron over a typical 1 second period . The spike events are clearly shown , indicating a spike rate of about . There is clearly evidence of spatial structure in the firing pattern , which we elucidate through the spectrum below . The current is plotted in Fig . 3 . It is much more smoothly varying than the spiking voltage . However , firing events can still be discerned via their associated rapid increases in versus time , meaning that a firing rate , as opposed to modulation in rate , will be the most obvious feature on any spectrum . The variables and are shown in Fig . 4 . In order to show the spatio-temporal structure of these fields , only a small part of the spatio-temporal domain is shown here . The scales are different for Figs . 3 and 4 . These variables denote the propagation of signals between neurons . The plots show signals emanating from each firing event as ‘’-shaped features . The apex corresponds to the firing event , while the two arms of the ‘’ show the propagation of the signal forward in time at constant speed in both spatial directions . The gradient of the arms of the ‘’ for the direct feedback term is , corresponding to the signal speed given by ; likewise , the gradient of the features for the loop feedback is , which equals . It is also clear from the length of the arms that the direct feedback events in Fig . 4 ( a ) have a longer spatial range than the loop-mediated ones in Fig . 4 ( b ) , in accord with cm and cm here . We now examine the power spectra for and . To complete comparisons , we have carried out simulations for the spike-based model , Eqs . ( 26 ) – ( 30 ) , the NFT , Eqs . ( 31 ) – ( 34 ) , and evaluated the theoretical field prediction through the transfer function of Eq . ( 42 ) . To consider the effect of the remnants of spike features on , we also have constructed the power spectrum of a series of stereotypical spike features in , which can be added to the NFT predictions of . To illustrate these spectra , we have carried out simulations for the case of , , and = 0 . 06 s . In Figs . 5 and 6 we show results for analyses of the current density term and membrane potential , respectively . In Fig . 5 , the four rows , in order , represent analyses of the spike-based model , the simulations of the NFT equations , the theoretical analysis of the spectrum of the NFT through Eq . ( 42 ) , and a theoretical analysis of NFT as for the third row but augmented with spike features . The three columns represent the breathing mode power , the spatial correlation function [from the inverse Fourier transform of ] and the full spatio-temporal power spectrum . Panel A shows for the spike-based model . The power is large at zero frequency and falls with increasing frequency; however , it is dominated in this case by features related to the spike rate; namely peaks at about 22 Hz and its harmonics . Panel B shows the correlation function , showing that there is some significant spatial order in the system; with decaying to in about 2 cm . Panel C shows ; here we see that there are large features at 22 Hz and 44 Hz associated with spiking behavior superposed on a more smoothly varying background with a maximum at ( 0 , 0 ) . The spatial frequency extent of these features , about , is equivalent to the correlation length seen in panel B . Panels D–F show the equivalent for the NFT simulation . The obvious difference is the lack of spike-features , since the NFT simulation does not contain spiking events . Otherwise , the shape ( but not the magnitude ) of the behavior is very similar to that of panels A–C . Panel G shows a theoretical calculation of from Eq . ( 42 ) ; it is evident that it is very similar to that of the simulation of panel D . Panel H shows the spatial correlation function; it has a similar correlation length to those of B and E; however , it does not have the same minimum at approximately 0 . 07 m that is the case for panels B and E . This negative correlation in panels B and E may be attributable to the toroidal boundary conditions in space . Panel I shows , which agrees with the simulation of panel F and the background of the panel C for the spike-based model . Panel J depicts calculated from Eq . ( 42 ) , with the addition of spike features arising from the spectrum of spikes . This compares well qualitatively to Panel A . The major discrepancy is the magnitude of the power . This is due to the interplay between the spike-based mode and the rate-based mode . The rate-based oscillation influences the synchrony of the spike-based mode , thus magnifying the power when resonances occur , such as for this set of parameters . The size of the major resonance can therefore vary tremendously as a function of . Panel K shows the NFT spatial correlation function , and panel L the NFT prediction of to which has been added the power spectrum due to a series of spike remnants in . Panel L compares moderately well with panel C; the major discrepancy is the greater extent of the resonant features in , corresponding to less synchronization of neurons than is seen in the spike-based simulations in panel C . Overall , for Fig . 5 , when features attributable to spikes are taken into account , we note that the NFT theory and simulation generally predict well the underlying shape of the power spectra ( though not its magnitude ) . In Fig . 6 we show a similar analysis for the membrane potential . The first row represents the simulation of the spike-based model; the second the reconstruction of a spike sequence from the simulations of the NFT model . Note that there is no NFT linearized prediction in this case since the NFT theory does not consider explicitly . The three columns represent the breathing mode power , the spatial correlation function ( from the inverse Fourier transform of ) and the full spatio-temporal power spectrum . Panel A shows for the spike-based model . It is dominated by features related to the spike rate; namely peaks at about 22 Hz and its harmonics . Fluctuations due to non-spike ( e . g . subthreshold ) processes are much lower in magnitude . Panel B shows the spatial correlation function , showing , as in Fig . 5 , decaying to in about 2 cm . Panel C shows ; here we see that there are large features at 22 Hz and 44 Hz associated with spiking behavior; the spatial frequency extent of these ( around ) approximately equals the inverse of the spatial correlation length . Panels D–F show the equivalent for the NFT simulation , in which spikes have been generated through the process described earlier . One notes that Panel D shows a similar ( but not exactly identical ) spectrum to Panel A; for example , the spike rates are slightly different and the spikes are less broad , consistent with less variation in inter-spike interval . Panel E shows that there is no discernable correlation between neighboring neurons in this method . Panels F shows the spectrum of the reconstructed spike sequence; the swaths at the spike rates represent very distinct firing frequencies that are uncorrelated in space . There is no equivalent NFT theoretical prediction since the NFT does not contain spiking events explicitly . Overall , for Fig . 6 , we note that the NFT theory and simulation predict temporal structure of spiking well , but are not as accurate spatially . This is attributed to the problem of defining initial conditions from the reconstruction of spikes . One could in principle , knowing the result of the spike-based approach , define initial phases to produce a similar correlation . We have not done this . Spike rates also show more fluctuation in the spike-based model than the NFT model . To summarize , we observe with Figs . 5 and 6 that the analysis of shown in Fig . 5 is more appropriate for analyzing the correspondence between the spike-based model and NFT where there is specific interest in the behavior of the NFT model ( e . g . where collective modes dominate behavior ) . However , the latter analysis of is likely to be appropriate when spike-rates are the dominant issue to consider . We now consider how the behavior of the models change as key parameters are varied . To do this , we carry out simulations of both sets of equations ( 26 ) – ( 30 ) and ( 31 ) – ( 34 ) for the spike-based and NFT models , respectively; and use the methods of [14] to reconstruct spike-sequences from the NFT prediction . The power spectra or as appropriate for both situations are then compared . Plots of the power spectrum against temporal frequency are then stacked to represent graphically the changes in resonances and power fluctuations in response to a variation in a parameter . Fig . 7 demonstrates the effect of a change in the external driving current . Higher naturally leads to a higher firing rate . Part A shows the breathing mode power for the spike based model as a function of drive current . There is an abrupt change in the spectrum at ; the maximum of shifts from 5 . 5 Hz to 11 Hz . From this point , the major frequency feature increases in frequency as drive current increases , corresponding to a mean firing rate of the neurons that is in agreement with neural field theory . Part B shows the predictions of the neural field model in terms of the power spectrum of the reconstructed voltage trace . The two graphs show that the resonances occur at similar frequencies , with a trend of increasing frequency with increasing . However , these frequencies are not exactly the same . For example , at , the spike-based approach in Panel A gives a simulated rate of 11 Hz , whereas in the NFT model shown in Panel B a rate of 9 Hz is observed . At , the spike-based and field-based frequencies are 17 Hz and 14 Hz respectively . In this case the spike-based model is dominated by the regular spiking behavior predicted by NFT . Note that the lumpy structure is a result of the resolution limit of the plot; it is not a chain of discrete peaks . In Fig . 8 we show the effect of varying the time delay for the case of a high external drive current and fairly high loop feedback . Since is well above there is a high firing rate and the feedback is not required to maintain activity . Direct feedback is set to zero . The temporal frequency spectrum at is plotted against time delay . Also shown on the plot are the predictions of resonances from the NFT , through Eq . ( 42 ) . These are shown by the solid lines . The half-integer multiples of these are shown by the dashed lines . The plot clearly indicates a firing rate of around 22 Hz . A harmonic at around 44 Hz is also present on the plot . However , the firing rate is not completely independent of , and varies between approximately 21 Hz and 22 Hz . There are two clear regions , at about and when the power is very large , indicating a sharp resonance in activity of the breathing mode , a property of the network . A feature of this plot is the dependence on of the magnitude of the resonance at the spike-rate of 21–22 Hz . A large response occurs when the firing frequency is that of the fundamental predicted by Eq . ( 43 ) . A low response occurs when the firing rate is exactly double that of the prediction of Eq . ( 43 ) . We emphasize that in this case synaptic coupling between a neuron and its neighbors is weak compared with the driving current , implying that each neuron has a well established limit cycles for its firing , dominated by . The reason for the discrepancy between the predicted resonances of Eq . ( 42 ) and the resonances seen in simulation appears to be the loss of spatial synchrony of the neurons at the time delays predicted by Eq . ( 42 ) to be resonances ( i . e . the solid lines of Fig . 8 ) . Instead of near synchronous firing , more intricate spatial patterns , e . g . traveling waves [47] , [48] , are formed causing a reduction in . This phenomenon is not seen when spike rates are significantly reduced by lowering , as discussed below . Next , the effect of a wide range of time delays is demonstrated for a low firing case . In order to elucidate the interaction between the loop resonances captured by the mean field approach and the effects of spike firings , a small loop connection strength has been chosen , with no direct feedback , and a drive current equal to the critical current . This ensures that any positive feedback will result in significant activity . This time , the appropriate analysis is with the power in the current fluctuations , . Part A of Fig . 9 shows the breathing mode power as a function of time delay , for a simulation of the spike based equations . Part B shows the same plot , but as predicted by the rate-based theory through Eq . ( 42 ) . One can see three clear regimes in Fig . 9A . For , the major feature on the power spectrum is a resonance at about 10 Hz , which is double the neuron spike rate . However , there is a hint of power at 5 Hz at . An examination of the neural firing patterns shows that there are strong correlations between neighboring neurons ( e . g . , a neuron firing at 5 Hz out of phase with its neighbor ) leading to the 10 Hz feature being more prominent than the 5 Hz one . For , the major feature is the resonance induced by the delay loop . The neurons adopt a firing rate that is equivalent to the resonance frequency predicted by the NFT ( i . e . , the resonances of Fig . 9B ) . Harmonics of this frequency are also clearly visible . The key result in this graph is that the firing rates seen on Panel A for the spike-based model when are in the positions predicted by NFT , while for the NFT resonance is not strong enough to capture this mode and the firing rate reverts to that of the spike-based resonance . A close-up of part of Fig . 9A is shown in Fig . 10 , where the solid lines show the loop frequency prediction of Eq . ( 43 ) ; the two are very closely related . At there is an abrupt change in the power spectrum and for the lowest frequency peak falls in magnitude as increases until it vanishes at . Here , the neurons are no longer able to fire slowly enough to follow the loop resonance frequency which is low for large , and instead the firing rate switches to ( nearly ) double the loop resonance frequency . However , this transition is subtle and Fig . 10 shows a very slight downward shift in the frequency compared to double the loop frequency . The above behavior is similar to that found for a population-based neuron model with loop feedback [16] . In that model the authors found that their system could jump between two regimes of behavior as the time delay was varied . In one regime , the system fired with a rate equal to the reciprocal of the time delay , or an integer multiple of this frequency ( i . e . , it decreased as the time delay increased ) ; in the other regime , it produced a firing rate independent of the time delay . The system alternated between these regimes as the delay time increased . In our model we also see this break between a firing rate roughly independent of time delay ( for ) , and one where the rate approximately follows the reciprocal of the delay time ( for ) . The other major parameters that can be changed are the connection strengths . We illustrate this case by studying the effect of altering the balance between the direct and loop connection strengths and , respectively . The mean field solution for firing rate depends upon the sum of and . However , fluctuations in firing rate are expected to be different . By setting , we ensure that the equilibrium firing rate is the same in all cases and we can study how the resonances and stability change as the balance between the terms changes . The other key parameters are selected as and . Fig . 11 shows the breathing mode power as a function of the direct ( non-loop ) connection strength . In Panel A power is plotted in the form of contours; Panel B shows a three-dimensional representation of the same information . Note that in the plot , , implying that the loop connection strength is negative . The most distinctive feature in this plot is a bifurcation at , as a result of a strong loop negative feedback through . At , the system oscillates at about 5 Hz between a rapidly firing state and a non-firing state . For , the system fires at about 12 Hz , as predicted by NFT . Close to bifurcation the system experiences large fluctuations in firing rate , as expected by NFT [49] . There is some evidence of an increase in power at about 5 Hz just before the bifurcation , for . This part of the spectrum is indicated explicitly on both parts A and B of Fig . 11 . The fluctuations for are shown explicitly in Fig . 12 which shows against time and space for one second of time . The plot shows propagating fronts of activity . The velocity of propagation has a range of approximately 0 . 4– and a typical value of around , and this variation results in the firing rate of each neuron showing considerable fluctuation with time . The point of bifurcation is also presented in more detail through Fig . 13 . Panel A shows ; we see a decrease in power with increasing frequency , with a hint of a peak at around 4 Hz . Panel B shows the spatial correlation function ; there is long-range order here with dropping to in about 3 cm . Panel C shows the for the spike based simulation . There is a background of activity that peaks at ( 0 , 0 ) ; on top of this there is a diagonal line of peaks with gradient of approximately ; corresponding to a traveling wave with velocity of about , consistent with the typical velocity of a wavefront in Fig . 12 . This is significantly larger than the loop propagation speed illustrating that the rate of propagation of activity along axons is not the sole determining factor for the wavefront velocity . Indeed , Bressloff has demonstrated that propagation of waves in a one-dimensional network of integrate-and-fire neurons is dispersive and dependent upon synaptic strength and delays in addition to axonal properties [50] . Panels D , E and F show the equivalent calculations from the NFT through Eq . ( 42 ) . In panel D we see a power spectrum with a strong peak at 3 Hz ( similar to the peak of panel A ) and then clear resonances at higher frequencies . Panel E shows the correlation function ; there is large long range order predicted , consistent with being on the edge of an instability . Panel F shows the predicted . One can see a peak at about 3 Hz and zero spatial frequency , similar to Panel C for the spike-based simulation . However , the major feature of this plot are the deep dips lying on a line of gradient . It is interesting to remark that this gradient is about half that seen for the resonances in Panel C . In these plots it is clear that the spike-based simulations and linearized NFT predictions are considerably different . This is not surprising given that this system lies very close to an instability where linear predictions are expected to break down . It is also possible that the critical point in NFT might be at a slightly different parameter value than for the spiking model .
We have explored the relationships between spiking-based and rate-based neural models by using both approaches to model the same test system — a one dimensional array of neurons coupled both directly and via a delayed feedback loop . The dynamics predicted by both approaches has been compared predominantly through the power spectra of the membrane potential and current density . We have focused on the relationships between resonances associated with the firing of single cells and populations of cells , particularly the overlap and transitions between these two regimes . We have shown how the dynamics , especially prominent resonant effects , depend on the key parameters of the model ( specifically delay-loop time , loop connection strengths and drive current density ) . The spike-based approach of Eqs ( 26 ) – ( 30 ) supports two modes of oscillation . First , there is the highly nonlinear spiking mode in which each neuron spikes according to its input . This mode , for a single Wilson neuron , has been well-studied [2] , [49] , [51] . Also , a collective mode can exist , in which the firing rate undergoes small or large oscillations . The spectrum of these oscillations can be determined through neural field methods [as in Eq . ( 42 ) ] . Both modes can be obtained through an analysis of an equivalent neural field model: the spiking modes from a reconstruction of the voltage , and demonstrated most directly through ; the collective modes through an analysis of the current fluctuations and demonstrated most directly using . At this point we again stress the distinction between the firing rate and fluctuations in the firing rate . Neural field theory predicts both , namely the firing rate itself , and how the firing rate fluctuates with time and space . In an extreme case , this could take the form of bursting — a neuron fires rapidly for a short period of time , and then is silent for a period of time . Thus there are two different time scales here , the inverse of the firing rate , and the period for the bursting–silent oscillation . Generally , however , the collective modes are not of this extremely nonlinear bursting form , but can be modeled by the linear analysis of Eqs . ( 35 ) – ( 42 ) . Results of spike-based and NFT simulations and predictions can be compared through plots of the power spectrum in current density , and membrane potential , . In the latter case , since the membrane potential does not feature in the NFT equations explicitly , a sequence must be reconstructed from knowledge of other variables , either the mean firing rate through time integration or the current density through a neuron-in-cell method [14] . Analysis of , and particularly the breathing mode power proves useful when there is significant power in the collective modes of oscillation; however , when spike-based behavior dominates , an analysis of gives more direct insight . A disadvantage of analyzing the membrane potential is that the reconstruction of spike sequences from the NFT solution does not produce the spatial correlations that are predicted through or seen in the spike-based model . This is because in the reconstruction of a voltage from a firing rate , [e . g . , Eq . ( 46 ) ] spatial effects manifest themselves differently; the reconstructed spike series depends upon the initial conditions which are not known a priori as a function of space . The collective and spike-base modes are not entirely independent of each other , particularly when the two time scales are the same ( or integer or half-integer multiples ) of each other . Indeed , Wu et al . [16] found for a rate-based model of a Wilson neuron that receives feedback from itself , the behavior can be dominated by either type of resonance . At certain points a small change in parameters is sufficient to cause the behavior to switch between one resonance and the other . In this spatial model , we see similar behavior . However , the spatial dimension adds a complexity to the behavior that is not present in simpler models . This manifests itself for example in the intricate traveling-wave firing patterns that are demonstrated in Figs . 2 and 12 for example , that have been described by Osan et al . [47] . In some cases , the collective and spike-based modes are both present in a system , without significant signs of interaction . For example , Fig . 5 shows that the NFT predicts the underlying power spectrum , on which features due to spiking events sit . For example , in this case there is a peak at zero temporal and spatial frequency , as is frequent for neural field models away from a resonance condition ( e . g . , [39] ) . However , there are situations where one of the two modes can dominate In Fig . 7 the spiking mode dominates; neurons fire at a rate that is close to that predicted by the NFT but show little modulation in this rate — i . e . there is little collective oscillation present . In Fig . 9A , and in more detail in Fig . 10 , for , the collective oscillation dominates to the extent that it captures the spiking oscillation . This can be seen from the correspondence between the solid lines in Fig . 10 and the regions of high power . In this case the firing rate is no longer that predicted by the equilibrium value of in Eq . ( 31 ) , specifically 5 . 5 Hz , but is equal to the position of the predicted resonance in the rate . Specifically , the larger , the lower the resonant frequency . We emphasize that this is not a trivial result . Resonances in spike rate as predicted by NFT ( i . e . the peaks in the spectrum ) are not in general the same as the mean firing rate itself . Analysis of Fig . 9 requires the use of both the spike-based and rate-based paradigms in order to fully elucidate the results . Analysis of the spiking pattern ( not shown ) shows that the system is also highly synchronized in space — in other words the collective oscillation has entrained the spiking oscillation . This is consistent with the prior result that a particular firing-rate based model agreed with a corresponding integrate-and-fire spiking model when interactions were sufficiently slow [17] . In Fig . 9B , the predicted power for the NFT case shows clearly the NFT resonances moving to lower frequency as increases . It is also interesting to note the transition between different dominant modes of behavior in this system . For , the strength of the collective oscillation is not sufficient to entrain the firing rate , which reverts to its equilibrium predicted value of 5 . 5 Hz . The major feature on Fig . 9A is at twice this , 11 Hz; analysis of the firing patterns ( not shown ) show that neurons are approximately paired; each fires in approximate anti-phase with its neighbor . Such behavior is common in neural simulations and its prevalence depends upon the strength of coupling between neurons , randomness in the couplings , noise and time delays . An anti-phase mode would be most likely when coupling strength and randomness are low [17] , [52] , noise is low [48] , but for a small range of time delays [52] , [53] . In our simulations we have not used random connections and have kept connection strengths low in order to ensure firing rates are of similar magnitude to resonances in the neural field simulations . Both of these favor the existence of an anti-phase state . Fig . 9A is similar to the results seen by Wu et al . [16] in which a simpler spatially uniform model was shown to exhibit similar transitions between spike-based modes and collective modes of behavior as the loop delay was changed . In Ref . [16] , spiking rates for one parameter set entrained alternately to either one of the two modes as was increased . Several switches between the modes were observed as delay time was increased from 0 to 0 . 7 s . In the current study , delay times were limited to what is physical within a thalamocortical system , and only a single switch between a spike-based mode and a collective mode is observed . For a different parameter set , Wu et al . [16] also demonstrated a doubling of the primary frequency of oscillation with as a result of a small change in time delay , similar to the doubling observed in this study in Fig . 9A . There can also be more complicated interplay between the two forms of oscillation . For example Fig . 8 demonstrates that the power in fluctuations in at a particular frequency can be strongly influenced by the relationship between this frequency ( in this case 22 Hz ) and the frequency of a collective resonance predicted by NFT shown in the figure by the solid and dashed lines . This variation in power requires particular comment . Naively , one might expect that where an NFT resonance , with positive gain , corresponds to the mean spike rate , there would be an enhancement of power . However , the opposite is the case here; the power is much reduced when the resonance predicted by Eq . ( 43 ) corresponds to the spike rate . There are two points to discuss by way of explanation . First , resonances predicted by Eq . ( 43 ) are weak when , so interaction between the two might be expected to be small for this parameter range . Second , the spatial nature of the model is important . Consider the propagation of activity in space . When a neuron fires , there is a time frame over which an effect is generated on the axon , through Eq . ( 30 ) . There is then a delay time for the signal to traverse the thalamocortical loop , followed by a time for an impact to be felt on the receiving neuron through Eq . ( 28 ) . A signal therefore takes a time to return to the same neuron in the cortex; longer times allow for spatial propagation via the delay loop . At a spike rate of each neuron receives strong input that arose from itself one spike-interval in the past , and consequently spatial communication between neurons is relatively weak . The weak communication encourages the formation of a variety of spatially patterned states [17] , [47] , [52] which are not synchronized in space and therefore lead to a reduction in . Such a pattern of alternating synchronous and asynchronous behavior has been found in previous studies [17] . It is possible that the dynamics of the spike based system is unduly influenced by the homogeneity of the parameters used [52] . For realistic systems , one would expect a wide range of values for the axonal lengths , synaptic decay times , etc . A system consisting of identical neurons may be particularly sensitive to modes of oscillation ( e . g . , synchronous in-phase or antiphase firing of all neurons ) that are less likely to be seen in practice . Using homogeneous values in a model has the advantage of increased analytic tractability; however , implications of such homogeneity require further study . The methods discussed are easily generalizable to two dimensions with appropriate choice of wave equation ( 6 ) . Results would be less easy to present , since two spatial dimensions and one temporal dimension would be present . Other neuron models ( e . g . the bursting model used by Robinson et al . [15] ) could be used by changing the forms of Eqs . ( 12 ) – ( 15 ) , and finding the equivalent rate equation ( 16 ) . An inhibitory population could be added with another set of variables . To conclude , we remark that we have demonstrated considerable overlap between the spike-based and the neural-field approaches . Where neural-field resonances are strong , spiking rates can be entrained to these resonances . A system that allows both modes to feature can show interactions between the two . Both spike-based and rate-based paradigms must be used to fully analyze the system . A spatial dimension adds complexity to the situation discussed previously by Wu et al . [16] . The theories can be considered as complementary methods of approaching the neural modeling problem , each offering a different physical emphasis . | We develop and demonstrate a model that allows us to examine how the predictions of spiking and rate-based models of neurons and their interactions are related . First , the behavior of a chain of neurons is explored by simulating each spiking neuron and spike-mediated interactions between neurons individually . Second , the same chain is studied using approximations based on the firing rate of the neurons . The predictions for these two approaches are closely compared and it is found that the simpler , rate-based approach captures the major system behaviors of the spike-based approach , namely spiking rates and modulations in those rates . Strong interactions between these modes take place when the frequency of one mode is an integer or half-integer multiple of the frequency of the other mode . | [
"Abstract",
"Introduction",
"Models",
"Results",
"Discussion"
] | [
"physics",
"statistical",
"mechanics",
"biophysic",
"al",
"simulations",
"computational",
"neuroscience",
"neural",
"networks",
"biophysics",
"simulations",
"biophysics",
"theory",
"biology",
"computational",
"biology",
"biophysics",
"neuroscience"
] | 2012 | Complementarity of Spike- and Rate-Based Dynamics of Neural Systems |
The fitness landscape captures the relationship between genotype and evolutionary fitness and is a pervasive metaphor used to describe the possible evolutionary trajectories of adaptation . However , little is known about the actual shape of fitness landscapes , including whether valleys of low fitness create local fitness optima , acting as barriers to adaptive change . Here we provide evidence of a rugged molecular fitness landscape arising during an evolution experiment in an asexual population of Saccharomyces cerevisiae . We identify the mutations that arose during the evolution using whole-genome sequencing and use competitive fitness assays to describe the mutations individually responsible for adaptation . In addition , we find that a fitness valley between two adaptive mutations in the genes MTH1 and HXT6/HXT7 is caused by reciprocal sign epistasis , where the fitness cost of the double mutant prohibits the two mutations from being selected in the same genetic background . The constraint enforced by reciprocal sign epistasis causes the mutations to remain mutually exclusive during the experiment , even though adaptive mutations in these two genes occur several times in independent lineages during the experiment . Our results show that epistasis plays a key role during adaptation and that inter-genic interactions can act as barriers between adaptive solutions . These results also provide a new interpretation on the classic Dobzhansky-Muller model of reproductive isolation and display some surprising parallels with mutations in genes often associated with tumors .
Introduced by Wright , the fitness landscape describes the possible mutational trajectories by which lineages evolve in a stepwise manner from genotypes that lie in regions of low fitness to ones of higher fitness [1] , [2] . When viewed as a whole , this metaphorical landscape represents a species' possible paths of adaptive evolution towards the optimal genotype in a particular environment . An outstanding question is whether the selective surface of fitness landscapes is smooth , containing a single global fitness optimum , or rugged , where selective constraints on differing mutational trajectories create multiple local fitness optima [3]–[7] . If the landscape is smooth , any path leading to the optimal genotype that continuously increases the population's fitness will be selectively favored , and the population will reach the global optimum on the landscape . However , if the landscape is rugged , adaptation will be constrained by the mutations available to increase the population's fitness [3] . This ruggedness can hamper the efficacy of natural selection compared to a smooth landscape by slowing the rate of adaptation due to pervasive genetic constraint [8] . Genetic constraint on fitness landscapes is due to fitness epistasis , where a mutation's adaptive value depends on the genetic background in which it arises [6] . Epistasis is a key component in such processes as reproductive isolation and speciation [9] , [10] , the evolution of sex and recombination [11] , [12] , as well as human diseases [13] . Theory shows that a form of epistasis called “sign epistasis” is necessary to constrain mutational trajectories on fitness landscapes [14] . Sign epistasis occurs when mutations are beneficial within the context of some genetic backgrounds , but detrimental within others . However , it is an extreme form of sign epistasis , recently dubbed “reciprocal sign epistasis” , that is necessary to create the local peaks and valleys on fitness landscapes [7] , [14] , [15] . Reciprocal sign epistasis occurs when the mutational path between two genotypes is selectively inaccessible due to intermediate , low-fitness genotypes . Such valleys are less likely to be crossed by natural selection alone , depending on the mutation rate and population size [16] . Therefore , genotypes that reside at local fitness optima are likely dead-ends for natural selection: even if a higher fitness peak exists elsewhere on the landscape , the neighboring fitness valley impedes adaptation to the global fitness optimum . Experimental studies aimed at testing mutational constraint on fitness landscapes due to epistasis have focused on testing engineered , biased amino acid substitutions , such as mutating residues at enzymes' active site ( s ) [17] , [18] , engineering likely evolutionary intermediates between ancestral and adapted versions of a single protein [19]–[21] , or quantifying interactions between mutations in different genes that display an observable phenotype [22] , [23] . Such work suggests that genetic constraint due to sign epistasis is prevalent and that adaptation can take surprisingly few mutational paths to the optimal genotype on the landscape . Some genotype-phenotype mapping studies using molecular data have inferred a multi-peaked landscape using proxies for fitness , but the extent of the role played by local optima during adaptation was either unknown [23] or limited [17] , [24] . Here we describe a rugged fitness landscape that arose during an experimental evolution . We identify the molecular nature of the mutations resulting from the evolution , and describe which mutations are individually adaptive . Adaptive mutations in two genes appear several times in different adaptive lineages , and we determine that they are selectively mutually exclusive due to reciprocal sign epistasis . The genetic constraint between these two mutations causes a rugged fitness landscape , as both mutations occur multiple times during the evolution and are highly adaptive individually , while highly maladaptive in concert . This work shows that inter-genic interactions can act as barriers between adaptive solutions and adds to the mounting experimental evidence that the constraint caused by epistasis is of central importance in evolutionary biology .
We have further characterized a previously described population of asexually-propagated haploid S . cerevisiae that were experimentally evolved under glucose limitation in continuous culture for 448 generations [25] . In that study , the chemostat was seeded with equal quantities of three otherwise isogenic haploid S288c strains that each expresses a different fluorescent protein constitutively ( GFP , YFP or DsRed ) , and the proportions of the three colored lineages were tracked over time using flow cytometry . Expansions and contractions of the colored subpopulations were monitored , and a total of five adaptive clones ( M1–M5 ) , were isolated from the various colored subpopulations at generations 56 , 91 , 196 , 266 and 385 , respectively ( see Materials and Methods for additional details on experimental design ) . That study identified a total of 12 independent mutations in these clones , using tiling microarrays [25] . However , through analysis of progeny of these clones , we discovered that some of them harbored additional , adaptive mutations of unknown identity , so we performed whole-genome sequencing on all five clones and their ancestor in order to identify these mutations . Sequence coverage of the nuclear genome ranged from 21× to 45× ( Table S1 ) . As expected , we discovered additional mutations: a total of five additional single nucleotide polymorphisms ( SNPs ) in M2 , M3 and M5 ( Figure 1 ) . All SNPs , indels and copy number variants found previously using tiling microarrays [25] were detected with this genome sequencing approach , though our sequence data were not able to discover the LTR insertion in GPB2 in M5 that we previously characterized , which is likely a limitation of the single-end sequencing . We also estimated the copy number of the HXT6/7 amplifications in M4 and M5 using real-time quantitative PCR ( qPCR ) , targeting the nearly identical HXT6 and HXT7 coding regions as well as the HXT7 promoter . These data were normalized to an ancestral strain without the amplification , which has one copy each of the HXT6 and HXT7 ORFs flanking the HXT7 promoter ( see Figure S1 for the region's structure ) . Given the model of mitotic recombination proposed by [26] to explain the HXT6/7 amplification , the number of HXT6/7 ORFs should be one greater than the number of HXT7 promoter regions , regardless of the number of amplifications that have occurred . The qPCR results indicate that there are ten HXT6/7 ORFs and nine HXT7 promoters in M4 , while M5 has 8–10 HXT6/7 ORFs and 7–9 HXT7 promoters ( Figure S2 ) . These data suggest there were a minimum of four mitotic recombination events for M4 and three for M5 to produce each array of HXT6/7 genes . Comparing these results to the sequencing coverage of the HXT6/7 region show that the coverage-based analysis underestimated the copy number for both clones ( Figure S1 ) , which may be due to the mapping algorithm used . Each mutation could be the result of positive natural selection , or alternatively , could be a neutral or slightly deleterious mutation that hitchhiked along with one or more adaptive mutations . Thus , we segregated the mutations and determined each one's fitness effect in genetic isolation using competition experiments [27] , where the only genetic difference between the mutant and wild-type competitor strain was the single mutation . Surprisingly , these data indicate that only 1–2 mutations per clone confer a significant fitness increase when considered singly , regardless of the total number of mutations in an adaptive clone ( Figure 2 ) . This increase in the number of seemingly non-adaptive mutations in later clones ( M4 and M5 ) could be due to a lack of sensitivity in our single mutation fitness assay , an accumulation of neutral or deleterious hitchhiking mutations ( Muller's ratchet [28] ) , or to non-additive fitness effects between mutations ( positive , synergistic epistasis [6] ) . Given the 1 . 2×107 bp size of the S . cerevisiae genome , a mutation rate of 5 . 12×10−10 bp−1 generation−1 [29] , and the number of generations passed for these evolved clones , the expected number of neutral mutations is 1 . 6 for M4 and 2 . 4 for M5 , while the probability of accumulating 4 or more neutral mutations in M4 is 0 . 084 and 5 or more in M5 is 0 . 092 ( Poisson distribution ) . While these probabilities are low , they do not allow us to reject the null hypothesis that these mutations are neutral . Thus , our results are consistent with there being 1–2 adaptive mutations per clone , with the remaining mutations being neutral , though we cannot unequivocally rule out there being additional adaptive mutations in M4 and M5 . Notably , we have also observed a nonsense mutation in MUK1 in an independent evolution experiment performed under glucose limitation ( Wenger and Sherlock , unpublished ) , suggesting that it may be adaptive either singly or in concert with another mutation . To test for evidence of additional adaptive mutations , we calculated the sum fitness effect of all the singly adaptive mutations from a particular clone to determine if that sum recapitulates the overall fitness of the clone . If additional adaptive mutations exist , the sum effect will be less than the fitness of the clone , assuming a no epistasis model . We found that the additive fitness effects of the individual adaptive mutations recapitulated the fitness of adaptive clones M1–M3 and M5 ( Figure 3 ) . In M4 , the additive fitness effect of the two adaptive mutations is significantly larger than the fitness of the adaptive clone . This suggests that negative , antagonistic epistasis between two mutations in M4 causes a reduction in its overall fitness . This may be because the fitness effect of each individual mutation is so large that when combined using an additive model , the additive fitness effect is larger than the upper bound of fitness in the given environment . The genes containing singly adaptive mutations are enriched [30] for the GO terms “hexose transport” and “negative regulation of Ras protein signal transduction” ( p = 1 . 36e-4 and p = 5 . 8e-3 , respectively; FDR<0 . 01% ) , suggesting that adaptation was due to increased glucose transport and signaling through the Ras/cAMP pathway , as previously hypothesized [25] . We compared the mutations found in our experiment with S . cerevisiae polymorphism data [31] , and found that MTH1 and RIM15 both have naturally-occurring premature stop codons alleles in environmental isolates . This suggests that the mutations leading to premature stop codons we see in these genes during experimental adaptation to limiting glucose are also ecologically-relevant mutations that may provide a fitness advantage in nature . Furthermore , we have also observed variation in the copy number of the HXT6/HXT7 locus in a survey of ∼70 yeast strains ( B . Dunn and G . Sherlock , unpublished ) , again suggesting ecological relevance of these mutations . The within-population reproducibility of adaptation between different lineages was striking - three independent mth1 adaptive nonsense mutations arose in clones M1–M3 ( hereafter referred to as mth1-1 , mth1-2 and mth1-3 ) , while the amplification of the tandemly arrayed glucose transporter genes HXT6 and HXT7 arose independently in clones M4 and M5 , as well as within the green subpopulation by the end of the evolution experiment ( see Figure 3 in [25] ) . These repeated independent changes suggest that the presumptive loss of MTH1 function or increased HXT6/7 copy number - both of which result in increased HXT expression [26] , [32] - are effective mechanisms by which yeast can adapt to limiting glucose . Strikingly , despite the common occurrence of these two mutations independently , none of the five clones we characterized had them both . To further investigate this trend , we genotyped 22 randomly picked clones from the yellow subpopulation at generation 266 for the mth1-3 allele and HXT6/7 amplification . While the subpopulation was heterogeneous , none of the 22 clones had both mutations ( Table S2 ) . Additionally , when examining our estimates of allele frequencies throughout the evolution experiment , we found that the mth1-3 allele began decreasing in frequency concurrent with the increase in frequency of the HXT6/7 amplification in the yellow subpopulation , further suggesting that the two mutations did not co-exist within the same clonal lineage ( see generations 200–300 , Figure 4 ) . We also genotyped 24 random clones isolated from the generation 448 terminal green subpopulation , and again the two mutations were never seen to co-exist , with all 24 clones carrying the HXT6/7 amplification but not the mth1-1 allele ( Table S2 ) . Since the MTH1 coding sequence has a large capacity for nonsense mutations ( 169/434 codons differ by only one nucleotide from stop codons ) , and since we were only assessing the specific mth1-1 allele within the terminal green population , we sequenced the full-length MTH1 coding sequence in 4 of the 24 terminal green subpopulation clones , and found that all four had the wild-type coding sequence ( data not shown ) . Taken together , these results suggest that the mth1 mutation and HXT6/7 amplification did not exist together in the same clone to reach detectable frequencies during our experiment , suggesting they may be selectively mutually exclusive . Reciprocal sign epistasis for fitness can constrain evolutionary trajectories and even create multiple peaks on a fitness landscape [7] , [14] , [15] , a situation that would create selectively mutually exclusive mutations . Thus , we hypothesized that reciprocal sign epistasis underlies mutual exclusivity between the observed mth1 mutations and HXT6/7 amplification . To test this hypothesis , we constructed double mutant strains ( as outlined in Figure S3 ) containing either an mth1-2 or mth1-3 nonsense mutation and an HXT6/7 amplification allele , and competed this double mutant against either a wild-type strain , an mth1 mutant , or an HXT6/7 amplification mutant . As controls , we also competed single mutants and wild-type spores from the same dissection against a wild-type strain . As expected , the single mutants were again more fit than the parental wild-type strain . However , the double mutant was significantly less fit than the wild-type strain as well as both the mth1 and HXT6/7 amplification single mutant strains ( Figure 5A and Figure S4 ) . This shows that within the genetic contexts we tested , a clone with both a nonsense mutation in mth1 and an amplification in HXT6/7 is highly maladaptive , and is thus unlikely to reach an appreciable frequency during glucose-limited evolution . To determine if sign epistasis during our evolution experiment was a common phenomenon , we constructed pairwise combinations of non-co-existing adaptive mutations from M1–M5 and competed them against the wild-type parental strain . While we saw no further evidence of reciprocal sign epistasis between these pairs of mutations , the RIM15 and GPB2 mutations do show standard sign epistasis , where the double mutant is less fit than one of the single mutants ( GPB2 ) but more fit than the other ( RIM15 ) ( Figure 5B ) . In addition , negative epistasis is also prevalent , defined as the double mutant having a fitness effect that is less than the additive effects of the single mutants , but still greater than the fitness effects of each single mutant . The pairs IRA1/ ( HXT6/7 ) yellow , IRA1/RIM15 , MTH1/GPB2 all have significant negative epistasis between the single mutations . Since negative epistasis between adaptive mutations results in a double mutant with a fitness greater than both single mutants , it is selectively favorable and does not act to constrain the fitness landscape [7] . As with the inferred negative epistasis between mutations found in M4 ( Figure 3 ) , it is possible that the calculated negative epistasis between these mutations is due to the large magnitude of their individual fitness effects exceeding the maximum fitness , and the seemingly pervasive negative epistasis may be non-specific , possibly occurring between any two mutations of individually large enough fitness effect .
We have characterized the fitness of individual mutations as well as genetic interactions between mutations that arose during experimental evolution of yeast under glucose limitation . Our exhaustive analysis of single mutation fitness determined that only one to two mutations per adaptive clone individually result in a fitness gain , regardless of the total number of mutations in an adaptive clone ( Figure 2 ) . The remaining mutations' effects are consistent with the null hypothesis of neutrality . Our data also show that , given an additive model , the identified singly adaptive mutations are sufficient to explain the fitness of adaptive clones M1–M3 and M5 , and in one case ( M4 ) , there is evidence for negative epistasis between mutations ( Figure 3 ) . The effect of negative epistasis may act like the idea of diminishing returns – adaptive mutations of large individual effect result in a smaller fitness effect when they occur together . The idea of diminishing returns of adaptation over time in a constant environment is further supported by data from long-term E . coli evolution experiments , where the rate of fitness improvement was initially fast but quickly decreased over time and then remained low [33] , [34] . Perhaps this decrease in the adaptive value of accumulated mutations is due to a maximum intrinsic fitness for a particular environment . This makes intuitive sense in the light of the mutations in our study having high relative fitness coefficients ( 1 . 1 to 1 . 45 ) , which when combined under an additive model would result in extremely large fitness coefficients . This non-specific , negative epistasis-like phenomenon is further exemplified by the interactions between non-co-existing adaptive mutations ( Figure 5B ) . Thus , it is possible that the pervasive negative epistasis observed is genetically promiscuous and not a result of the specific interactions between the two mutations , and any two mutations with large enough fitness increases will display such interactions . To our knowledge , this phenomenon has not been addressed in the literature , and there is as yet no distinction between true negative epistasis between specific sets of alleles and a non-specific , negative epistasis-like phenomenon due to a fitness upper bound , even though both scenarios would be classified as negative epistasis by its mathematical definition . We have also determined that reciprocal sign epistasis between mutations in MTH1 and HXT6/7 constrains adaptation , which causes these two mutations to be mutually exclusive during the evolution experiment , a result that is consistent with data from an independent glucose-limited yeast evolution experiment [35] . This constraint can be represented in an empirical fitness landscape , which maps out the mutational evolutionary trajectories available for natural selection , and our data make possible the construction of this landscape using direct fitness effects of combinations of mutations . Reciprocal sign epistasis between the adaptive mutations in MTH1 and HXT6/7 can be represented as a simple two-locus fitness landscape ( Figure S5 ) . Here , the mth1 mutation is at a local optimum , as the only mutational steps available lead to a decrease in fitness , and it is not the fittest genotype on the landscape ( Figure 5A and Figure S4 ) . The HXT6/7 amplification mutation is the fittest genotype and is therefore the global optimum for this landscape . A consequence of this landscape is that lineages with an adaptive mutation in MTH1 are stuck on a local adaptive peak and may not be able to reach the higher fitness peak where the HXT6/7 amplification mutation lies . This is the likely reason why the mth1 and HXT6/7 mutations remain mutually exclusive for the duration the experiment . The issue of how populations move from one peak to another in nature ( the “peak shift” problem ) has been disputed since Wright conceived of the fitness landscape metaphor [1] , [2] , [36]–[38] . Furthermore , the existence of multi-peaked fitness landscapes themselves has recently been questioned at the theoretical level due to their immense multidimensionality causing neutral ridges connecting genotypes of high fitness on the landscape ( the holey landscape model ) [4] , [39] . These ridges thereby eliminate the classical peak shift problem , and there is some experimental support for such ridges ( e . g . [21] , [40] ) . Since the adaptive mutations in mth1 and HXT6/7 remain mutually exclusive in our experiment , even after sampling several clones in different lineages , this supports the argument of an adaptive valley on the fitness landscape rather than a ridge connecting the two peaks . Alternatively , a ridge connecting the two peaks might be long and circuitous , and our experiment was not performed for sufficient evolutionary time for neutral evolution on a fitness ridge to occur . This is likely the case in [41] , where an unresolved potentiating mutation is thought to have occurred 20 , 000 generations into the evolution experiment , allowing an innovative phenotype to evolve . In either case , the constraint is such that it would be difficult to adapt from one peak to the other . In reality , all possible genotypes are present on a fitness landscape . In our case , while it is impossible to experimentally test all combinations of all possible mutations in conjunction with the mth1/ ( HXT6/7 ) double mutant to prove that this portion of the fitness landscape does indeed have two peaks [3] , the large population size of the culture ( 2×109 ) means that a large spectrum of mutations should be sampled by natural selection often ( ∼107 new SNP mutations per generation , based on the previously measured mutation rate [29] ) . This suggests that natural selection may be rejecting the double mutant in myriad genomic contexts . Determining whether further evolution of an mth1 single mutant strain ever results in an HXT6/7 amplification mutation will shed additional light on the feasibility of adaptation from one peak to the other . If the HXT6/7 amplification can indeed appear on the mth1 background in such an evolution experiment , this would suggest that a compensatory mutation or mutations provide a fitness ridge between the peaks . While observing how an evolution experiment proceeds cannot indisputably prove the shape of the fitness landscape , it can inform the most relevant and repeated paths of adaptation . It has been understood for quite some time that epistasis is a fundamental component of adaptation [6] , but only have recent technological developments facilitated the discovery and testing of individual nucleotide changes and how they interact with one another to create function and fitness [6]–[8] , [18]–[22] . We speculate that for the reciprocal sign epistasis between mth1 and HXT6/7 , the fitness defect may be caused by an overabundance of hexose transporter proteins in the cell , due to the fact that both individual mutations act to increase hexose transporter transcription [25] , [26] , [32] . It is possible that the devotion of too many resources to the production of hexose transporters may take resources away from other essential functions . For example , the secretion machinery by which hexose transporters are localized to the plasma membrane may be overwhelmed by their overabundance . Alternatively , the large number of hexose transporters may take up space on the membrane's surface , preventing other transporters from being correctly localized or negatively impacting the fluidity of the membrane . Another scenario is that the overabundant hexose transporters may aggregate and form plaques in the cell due to their 12 hydrophobic trans-membrane domains . Understanding the mechanism of the reciprocal sign epistasis between mth1 and HXT6/7 will be important for understanding the underpinnings of molecular fitness landscapes and is worthy of further investigation . The rugged mth1/ ( HXT6/7 ) fitness landscape has far-reaching evolutionary implications . In the Dobzhansky-Muller model of postzygotic reproductive isolation , two species are separated from each other by a pair of genomic loci that interact negatively to create a hybrid organism that is of lower fitness compared to its parents [42] , [43] . This model bears striking similarity to the epistatic interaction we see between mth1 and HXT6/7 amplification ( also see [23] ) . In our case , if two lineages became fixed for the mth1 or HXT6/7 mutations , the low fitness of hybrid double mutant offspring may lead the two lineages on a path to reproductive isolation , though in contrast to a typical D-M pair , clones containing the individual mutations are more fit than wild-type clones . This is in contrast to a recent finding of a D-M pair between two strains of S . cerevisiae that were experimentally adapted to different environmental conditions [9] , [44] . In this case , the two mutations were adaptive in the conditions in which they evolved , but maladaptive when made to co-occur in one of the conditions , which is the traditional way in which D-M pairs are thought to arise . Mutually exclusive mutations are also known to exist in cancers [45] , [46] . Of specific relevance is the recent observation that mutually exclusive mutations in the Ras pathway are individually adaptive under limiting glucose in colorectal cells , an environmental condition believed to be relevant to cancers in vivo [46] . Intriguingly , these mutations lead to an increased expression of the glucose transporter GLUT1 resulting in increased glucose uptake by cancer cells [46] . These results from human cancers closely parallel the mutations we see in the glucose sensing and Ras/cAMP pathways , which lead to an increase expression of the hexose transporters . Thus , human cancers and yeast may respond to the same selective pressures by mutating the same pathways , and these parallels beg the thought that reciprocal sign epistasis might be the mechanism by which these cancer mutations are mutually exclusive .
The details of the experimental setup yielding the adaptive clones used in the present study have been described previously [25] . Briefly , three strains of haploid S288c that are isogenic , except that each expresses a different fluorescent protein constitutively ( GFP , YFP or DsRed ) , were seeded in equal quantities in a 20 ml chemostat device . The population was evolved for 448 generations at steady state under glucose limitation ( 0 . 08% ) at a dilution rate of 0 . 2 h−1 . During this evolution , the proportions of the three colored lineages were tracked using flow cytometry . At five points throughout the evolution experiment , the population was sorted into its component colored subpopulations using fluorescence activated cell sorting ( FACS ) , and 7 clones from the visibly adaptive subpopulation were isolated , competed to determine the clones' fitnesses , and the most fit clone was selected for subsequent investigations . The five resulting clones are labeled M1–M5 ( Table S3 and Figure 1 ) . Strains used and constructed in this study are shown in Table S3 . All batch and competitive chemostat cultures were grown as described previously [25] . Adaptive clones M1–M5 ( GSY1171 , GSY1180 , GSY1194 , GSY1200 , GSY1208 ) and an ancestral strain ( GSY1135 ) were single-end sequenced using the Illumina Genome Analyzer ( GAI or GAII ) . Single-end sequencing libraries were constructed for each clone using the Illumina Genomic DNA sample prep kit from 5 µg of genomic DNA and each library was sequenced on 2 flow cell lanes . Sequence analysis was performed as follows , using default parameters unless otherwise noted . Reads were mapped to the S288c reference genome ( downloaded from SGD on Dec 3 , 2008 ) using bwa-short in BWA v0 . 5 . 7 [47] and variants were found with SAMtools v0 . 1 . 7-6 [48] and filtered with SAMtools varFilter ( -d 5; -D 100 , 000; -S 20; -i 50 ) . For each genomic position , if the ancestral strain and adaptive clone shared the consensus genotype , the variant was eliminated . The resulting variants were filtered unless they passed the following heuristic filters . SNP: proportion of “N” bases covering position <0 . 1; majority non-reference base must be >80% of all non-reference bases at the position; when comparing ancestral to evolved , the proportion of non-reference bases must be <0 . 1 in one strain and >0 . 5 in the other . Indels: coverage >10; indel calls must make up >50% of coverage; proportion of the sum of two most frequent indel calls >0 . 8; proportion of reference matches <0 . 5; when comparing ancestral to evolved , the difference in proportion for shared alleles must be >0 . 3 . All novel SNPs were confirmed by Sanger sequencing ( primers in Table S4 ) . Three SNPs were excluded from further consideration because they were present in both the ancestral strain and each evolved strain of a particular color: chr02:353579 ( C to A ) in GSY1136 and M1; chr09:275382 ( PKP1 , G to T ) in GSY1135 , M2 and M4; and chr16:401704 ( MOT1 , G to C ) in GSY1137 , M3 and M5 . In addition , the COX18 mutation found in M2 and M4 as reported in [25] was also excluded , because it was determined the mutation was already fixed in the red subpopulation at the earliest sampling . Thus the mutation was not the result of evolution in the chemostat but was most likely a random mutation in the colony used to initiate the red population in the chemostat; we have also determined that this mutation is not adaptive ( data not shown ) . Real-time quantitative PCR ( qPCR ) of genomic DNA ( gDNA ) was used to determine the copy number of the HXT6/7 coding region , as well as the HXT7 promoter , using a previously described protocol [49] and primers listed in Table S4 . Results were normalized to the copy number of UBP1 , a non-varying locus also on chromosome 4 , to control for slight differences in input gDNA amount . Experiments were performed in triplicate . 95% confidence intervals were calculated as mean ± 1 . 96 * SEM . HXT6/7 copy number was also visualized using sequencing coverage . Adaptive clone coverage was normalized to the average coverage of the ancestral strain and this was divided by the ancestral strain coverage . A running median was calculated over this proportion to smooth the data . M1–M5 were backcrossed to a wild-type S288c ancestral strain containing the same fluorescent protein ( GSY1221–1223 ) . Diploids were sporulated and dissected , and resulting spores were genotyped for mutations using allele-specific colony PCR ( primers in Table S4 ) . The spores were also tested for mating type by cross-stamping two tester strains ( GSY2476 and GSY2670 ) onto YPD master plates of the dissections , grown overnight , followed by replica plating and selection for mated diploids on SC-ura+G418 plates . Backcrossing was repeated as often as necessary to get a single mutation segregating per cross ( Figure S6 ) . Double mutants for epistasis experiments were constructed by mating haploid single mutant strains and the resulting diploids were sporulated , dissected and genotyped . Pairwise competitive chemostats were performed as described [25] , but were sampled every 6 h over 20–25 generations . Single mutation competition experiments were performed in at least biological triplicate . As a control , we also performed competition experiments in at least biological triplicate of wild-type sister spores derived from the same backcross . Selection coefficients were calculated as described [25] , and normalized by subtracting the wild-type mean selection coefficient from the mutant mean selection coefficient for each mutation . Fitness was calculated relative to the competing wild-type strain , such that . The mth1/ ( HXT6/7 ) epistasis competitive chemostats were performed in at least biological duplicate , and the remaining epistasis competitive chemostats were performed once . For each of these experiments , a wild-type strain from the same dissection was used as a control . The linear phase of growth was determined , and the selection coefficient was taken as the slope of the linear regression line as described [35] and normalized as above . 95% confidence intervals were inferred for the slope of the regression using the confint ( ) function in R . Analysis of covariance ( ANCOVA ) was used to determine if the selection coefficient of the mutant strain was significantly different than the wild-type control . In this analysis , a model in which the mutant and wild-type experiments were allowed independent slopes was compared to a model in which they had a common slope . Epistasis was quantified as , where s is the normalized selection coefficient and x and y are mutant alleles of two different genes [6] . Error of epsilon was calculated by the method of error propagation: [22] , and a 95% confidence interval of epsilon was calculated as . If the value of epsilon fell outside this confidence interval , we considered there to be significant epistasis between alleles x and y . For each adaptive clone , the selection coefficients of all individually adaptive mutations derived from that clone were added and a pseudo relative fitness was calculated from the selection coefficient as above . This was compared to the relative fitness data for each clone from [25] . To make this comparison quantitatively , we used the epistasis framework above to calculate a value and confidence interval for epsilon . If indeed there are additional adaptive mutations , this would manifest itself as a positive value for epsilon , as the additive effects of the individual mutations would be less than the fitness of the clone . Quantitative Sanger sequencing [35] and the software PeakPicker [50] with default settings were used to determine the frequency of mth1-3 in triplicate throughout the evolution . The Illumina sequence data are available at the NCBI Sequence Read Archive under accession SRA020606 . 1 . | How organisms adapt to their environment is of central importance in biology , but the molecular underpinnings of adaptation are difficult to discover . Fitness landscapes illustrate possible steps adaptive evolution can take to increase the evolutionary fitness of individuals within a population , and the shape of the fitness landscape determines the accessibility of the fittest point on the landscape . On a rugged landscape , negative interactions between mutations cause fitness valleys separating fitness peaks , which can constrain adaptation and act as an adaptive barrier . Here , we comprehensively characterized the fitness of mutations that arose in clones during a yeast experimental evolution and found that mutations in two loci , MTH1 and HXT6/HXT7 , arose multiple times independently and are individually adaptive . However , when forced to co-occur , the double mutant has a lower fitness than either single mutant and even the wild-type strain . This negative interaction forces these two mutations to remain mutually exclusive during the experimental evolution and results in a rugged fitness landscape , where genetic constraint prevents lineages carrying the MTH1 mutation from reaching the higher fitness peak of HXT6/HXT7 . These results show that genetic interactions are central in shaping a very active portion of this fitness landscape . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genome",
"evolution",
"evolutionary",
"selection",
"epistasis",
"genome",
"sequencing",
"biology",
"adaptation",
"natural",
"selection",
"heredity",
"genetics",
"genomics",
"evolutionary",
"biology",
"genomic",
"evolution",
"evolutionary",
"processes",
"genetics",
"and",
"genomics"
] | 2011 | Reciprocal Sign Epistasis between Frequently Experimentally Evolved Adaptive Mutations Causes a Rugged Fitness Landscape |
Auxin underlies many processes in plant development and physiology , and this makes it of prime importance to understand its movements through plant tissues . In stems and coleoptiles , classic experiments showed that the peak region of a pulse of radio-labelled auxin moves at a roughly constant velocity down a stem or coleoptile segment . As the pulse moves it becomes broader , at a roughly constant rate . It is shown here that this ‘spreading rate’ is larger than can be accounted for by a single channel model , but can be explained by coupling of channels with differing polar transport rates . An extreme case is where strongly polar channels are coupled to completely apolar channels , in which case auxin in the apolar part is ‘dragged along’ by the polar part in a somewhat diffuse distribution . The behaviour of this model is explored , together with others that can account for the experimentally observed spreading rates . It is also shown that saturation of carriers involved in lateral transport can explain the characteristic shape of pulses that result from uptake of large amounts of auxin .
Auxin is the key integrating signal in plants [1 , 2] , and the dynamics of its movement within the plant is crucial to understanding development [3–9] and a multitude of physiological responses [10–12] . A traditional way to observe these dynamics in stems or coleoptiles is to apply a pulse of radioactively-labelled auxin at one end of a segment of the tissue , allow it to be transported for some period of time , and then cut up the segment into small pieces and measure the amount of label in each piece [13–15] . The profiles one obtains this way allow one to make some inferences about the underlying mechanism [14 , 16] . One parameter that is often measured is the velocity of the pulse . It provides information about the underlying permeabilities and diffusion constants of the transport mechanism [16 , 17] . I show here that one can also measure the rate at which the pulse spreads out , and that the values one obtains imply further constraints on the underlying mechanism . These constraints lead one to reject single channel models and to consider a variety of ways in which multiple auxin channels may be coupled together . ( Here the term ‘channel’ will be used for a file of cells specialised for auxin transport , though later we also consider intracellular compartments as potential channels . Proteins that move auxin in or out of cells will be called transporters , more specifically importers or exporters . ) Models of auxin transport with many channels have certainly been considered before; in particular , the flow and counterflow pattern of auxin in the root has been modelled and applied to gravitropism and growth control [18–20] . However , the aim here is somewhat different , namely , to use the shape of an auxin pulse to diagnose properties of the stem transport system .
The classic experiments on auxin transport in coleoptiles [13] show a pulse moving basipetally and broadening as it goes . One would like to measure both the velocity , v , of this pulse and what will be called the spreading rate , denoted by ρ , which is defined to be the rate at which the variance of the pulse increases . It is not always straightforward to measure the velocity , because there are typically several components to the auxin transport profile [14 , 21] . In particular , there are fixed or slower-moving components that trail behind the main pulse and may obscure its shape . These features cause even greater difficulties when measuring the variance , as small components far from the peak can cause a large increase in the variance . One strategy would be to try to fit the data to a composite model that attempts to explain all the features in the profile . An alternative strategy , taken here , is to try to isolate the main peak , separating it from trailing or fast moving components . There are reasons to expect the shape of a pulse due to a single channel to be gaussian in form ( S1 Text ) . The data were therefore fitted to a gaussian using least-squares , in the hope that this would pick out the main peak . This works quite well for the data from Goldsmith [13]; see S1 Fig . Given these fits , one can plot the mean and variance against time to obtain estimates for the velocity and spreading rate; see Table 1 and Fig 1A . Goldsmith’s data points represent averages of several coleoptile segments , four for the zero time point and two for all other time points . This introduces the danger that the spreading rate may be over-estimated , since averaging two pulses moving at slightly different velocities will give rise to an artifactual increase of variance with time as the pulses separate . The overestimate will probably not be very large , being of the order of the variance of the velocity distribution . However , one can entirely avoid this objection by using the counts from individual segments . Here we used the single-segment data ( S1 Data ) underlying the averaged curves in Fig 4 in Morris et al . [15] . These were kindly provided by Dr Morris . The raw data are of course noisier , and the problem is to determine the shape of the main peak in a reliable manner . Least squares performs poorly in this noisier setting ( S2A Fig ) , as it attempts to fit the entire profile . However , one can modify the fitting algorithm so that it more effectively selects the peak region and ignores flanking regions . We call this procedure maximal fit ( S1 Text and S2B Fig ) . One can also select the peak region by eye , trying to err towards excess width at short times and narrower regions at long times , so as to give a conservative estimate of the spreading rate . There is good agreement between this procedure and the maximal fit algorithm at long time intervals ( S3 Fig ) ; at shorter times the match is less good ( S4 Fig ) because of edge effects . Nonetheless , the estimates of spreading rate obtained by the two methods agree well , Table 1 , and the estimates of velocity and spreading rate obtained from Goldsmith’s data by least squares and maximal fit also agree reasonably well , within their error ranges . The conclusion of this section is that it is possible to measure both the velocity and the spreading rate of the main peak of an auxin pulse , though the spreading rate is more vulnerable to noise and questions of interpretation . Next we ask what the velocity and spreading rate can tell us about the underlying transport mechanism . Consider first a simple model consisting of a row of cells , where auxin is assumed to reach a uniform distribution rapidly inside cells and the flux ϕ between cell n−1 and cell n takes the form ϕ = p a n - 1 + q ( a n - 1 - a n ) , ( 1 ) where ai denotes the auxin concentration in cell i . Thus there is a polar component to auxin transport between cells given by the permeability p and a symmetric , diffusion-like component given by the permeability q . In the next section it will be shown how p and q may relate to known auxin carriers . Note that a non-zero q can be produced by combinations of polar carriers , and does not necessarily imply that there is an underlying physical diffusive coupling . The simple model is appropriate where the cells are small enough for intracellular diffusion to be rapid ( e . g . 10 microns or less in length ) . For instance , an implicit model of this kind underlies treatments of canalization or up-the-gradient mechanisms near to the apical meristem . It is probably not such a good model for the auxin pulse experiments we are considering , since the relevant cells , e . g . the xylem parenchyma , are long enough ( e . g . 100 microns ) for intracellular diffusion times to be significant . Nonetheless , it serves as a useful warm-up exercise . One can show ( S2 Text ) that the velocity , v , and the spreading rate , ρ , are given by v = p , ( 2 ) ρ = L ( p+2q ) , ( 3 ) where L is the cell length . Given an estimate of cell length L , knowing the velocity and the spreading rate fully characterises the model . In fact , the pulse at time t is well approximated by a gaussian of mean vt and variance ρt; see S2 Text . Fitting the data with a gaussian , as was done in the last section , therefore has some theoretical justification . The data in Table 1 allow one to calculate p and q . The velocity gives the permeability p , and taking L = 100 microns and using Eq ( 3 ) gives q . Instead of giving q itself in the Table , we give the ratio q/p , which is small when the movement of auxin is essentially polar and large when the diffusion-like coupling between cells dominates . As can be seen , the inferred value of q/p is large in all the data sets . A version of this model has been used by Renton et al . [22] ( their model I ) . The units in their version are not individual cells but 2 mm segments of stem ( the length of the pieces into which the stem is subdivided for counting labelled auxin ) . They assume that these segments are coupled by a completely polar flux; i . e . q = 0 . They are able to fit the peak region in their data by taking p = 9 mm/hr . Putting L = 2 mm , p = 9 mm/hr and q = 0 in Eq ( 3 ) gives ρ = 18 mm2/hr . This is larger than our estimate in Table 1 , which could be due to the fact that they use averaged data and different fitting criteria . There is a caveat here . In the above model , and in other models where a piece of tissue is used as the computational unit , e . g . the ‘metamers’ in [23] , this computational strategy may offer a gain in speed and simplicity . But care is needed in interpreting the results at a cellular level . In the above model , we can ask what parameter values would be needed to achieve the same velocity and spreading rate with cells as the computational units instead of 2 mm pieces . Taking L = 100 microns , with a velocity of 9 mm/hr and spreading rate of 18 mm2/hr ( as calculated above ) , Eq ( 3 ) gives 18 = Lp ( 1 + 2q/p ) = 0 . 1 × 9 × ( 1 + 2q/p ) , or q/p = 9 . 5 . Thus the completely polar segment-based model translates into a model at the cell level where the diffusive component is much larger than the polar component . We now turn to the key question: is this simple model plausible ? We have already pointed out that one needs to take account of intracellular diffusion in a realistic model . Another potentially unrealistic feature of the model is the dominance of diffusive coupling over polar transport between cells , implied by the values of q/p in Table 1 . Yet the strongly polar localization of the auxin transporter PIN1 [24] in what are presumed to be the principal auxin-transporting tissues of the stem suggests a high overall polarity of transport . We consider next whether this view is justified . Several families of auxin importers and exporters assist the movement of auxin between plant cells . One family of exporters consists of the PIN-formed family ( PINs ) [25 , 26] . Some PINs , e . g . PIN5 [27] , PIN6 [28] and PIN8 [29] are involved in movement between compartments in cells , and will be considered later . The remaining PINs are associated with the plasma membrane , and often show polar localisation within cells , e . g . a basal location of PIN1 in the stem [24] , or a more complex pattern of polarity at the shoot apex [30] , or a basal localization of PIN2 in epidermal root cells [31] , or a gravitropic movement of PIN3 to the bottom side of Arabidopsis hypocotyl endodermal cells [12] . Another family of exporters consists of the ABCB transporters ( also known as P-glycoproteins , or PGPs ) [32 , 33] . These are generally uniformly distributed over the plasma membrane , but can sometimes be polar . Thus ABCB1 is known to be uniformly distributed at the shoot and root apices but polarly localized in the mature root cortex and endodermis [32] , where it is found in the basal membrane , as are PINs [34] . Some ABCBs , e . g . ABCB21 [35] and ABCB4 [36] , may act both as exporters and importers , according to the balance of internal and external auxin concentrations . However , the predominant class of importers is the AUX/LAX family [26 , 37 , 38] . In general , these too seem to be uniformly distributed in the plasma membrane , though in one special case , the protophloem of Arabidopsis roots , AUX1 is asymetrically distributed [38] . Auxin can also be imported passively , without a specific protein channel , because of the lipid solubility of the protonated acid and the fact that the apoplast is acidic ( pH typically 4 . 5 to 5 ) . As the pK of the carboxyl group of auxin is 4 . 7 , about half the auxin in the apoplast will be protonated , and therefore able to enter a cell fairly easily . This is one of the twin pillars of the chemiosmotic theory [39–42] , the other being that auxin inside a cell , where the pH is higher ( cytoplasmic pH 7 . 2 [43] ) , will pass through the membrane far less readily , being about 99 . 7% in the charged anion form . Thus auxin is in effect trapped inside cells , and exporters are required for intercellular auxin movement , whereas importers would seem less necessary . However , recent measurements of the permeability of the cell membrane to protonated auxin are about two orders of magnitude lower than those available when the theory was formulated . For instance Rutschow et al . [44] find values for PIAAH close to those of Delbarre et al . [45] , at around 4 × 10−5 cm/sec , whereas older measurements gave 10−3 cm/sec [41] or 3 . 3 × 10−3 cm/sec [46] . These high estimates may have come about because the contribution of AUX importers was not appreciated . It seems likely , therefore , that auxin influx is dominated by AUX importers [44] . Now consider two adjacent cells in a channel , and let a1 be the cytoplasmic concentrations of auxin at the basal end of the uppermost cell , and a2 the concentration at the apical end of the cell below it ( see Fig 2 ) . Let b1 and b2 be the concentrations at the apical and basal boundaries of the apoplast between the two cells , and let L0 be the vertical width of this apoplastic region . Thus we allow there to be an auxin gradient within the apoplast; let the diffusion constant of auxin in the apoplast be D0 . Let α1 be the total permeability for exporters , i . e . the sum of permeabilities for both PINs and ABCBs , in the basal cell membrane of the upper cell , and α2 the total permeability for the apical membrane of the lower cell . Let us assume that AUX/LAX importers are uniformly distributed , so the combined permeabilities for import via AUX/LAX and passive movement of protonated auxin are equal at the basal and apical membranes , and have a value β . The flux ϕ per unit area between the cells is given by ϕ = α1a1−βb1 , ( 4 ) = D0L0 ( b1 - b2 ) , ( 5 ) =βb2 − α2a2 . ( 6 ) Eliminating b1 and b2 gives ϕ = ( α 1 a 1 - α 2 a 2 ) / ( 2 + r ) , ( 7 ) where r = βL0/D0 . This can be put in the form of Eq ( 1 ) , i . e . ϕ = pa1 + q ( a1 − a2 ) , by taking p = ( α 1 - α 2 ) / ( 2 + r ) , ( 8 ) q = α2/ ( 2 + r ) . ( 9 ) From this we get q / p = α 2 α 1 - α 2 . ( 10 ) Since PIN1 , which is probably the principal exporter in the xylem parenchyma in stems , is strongly localised at the basal end of cells [24] , the contribution of PIN1 permeability to α1 will be much larger than that to α2 . On the assumption that ABCBs are non-polarly distributed , the ABCB permeability contributions to α1 and α2 will be equal . However , they must surely be very small compared to the PIN1 contribution to α1 , since otherwise the isotropic distribution of ABCBs implies that auxin would be pumped out at a high rate in all directions . Thus α1 must greatly exceed α2 , and hence q/p must be small , highlighting the implausibility of the values of q/p in Table 1 . The simple model assumes complete mixing inside cells , which is unlikely to be a good approximation for the cells that transport auxin in stems and coleoptiles . Let us therefore assume that auxin diffuses through the cytoplasm with a diffusion constant D that is 60–90% [47] of its value in aqueous solution; the latter being 6 . 7 × 10−6 cm2/sec [48] . Suppose the cells transporting auxin have length L , which we take to be 100 microns . Then the analogues of Eqs ( 2 ) and ( 3 ) , derived in S3 Text , are 1v = 1p + L2D ( 1 + 2qp ) , ( 11 ) ρ ≤ L v ( 1 + 2q p ) , ( 12 ) with the inequality in Eq ( 12 ) approximating an equality as D becomes large . These two formulae give constraints on q/p and D . For instance , from the data of Morris et al . [15] we have v = 9 mm/hr and ρ = 10 mm2/hr , from which inequality Eq ( 12 ) implies q/p > 5 . 06 , and Eq ( 11 ) gives D ≥ 1 . 4 × 10−5 cm2/sec . Thus we still have the problem of a large value of q/p , and the unrealistic requirement for instant mixing inside cells has been replaced by the requirement for a diffusion constant that is twice the measured value for diffusion in water [48] . Possibly some intracellular transport mechanism could faciliate diffusion . Cytoplasmic streaming seems not to play an important part [49] , but there could be some other kind of active process , such as the movement of carriers along fibres postulated in model II in Renton et al . [22] . Against this , Kramer et al . [17] have shown that there is a broad agreement between auxin transport speeds and the bound 2D/L based on passive diffusion [16] . Moreover , even if there is enhanced diffusion , the problem of the large value of q/p still remains . Fig 3 compares a plausible single-channel model , with q/p small and a realistic cytoplasmic diffusion constant ( q/p = . 05 , D = 5 × 10−6 cm2/sec ) , with Goldsmith’s Fig 1A and 1D , for zero time and 30 minutes , respectively . The zero time curves match approximately , but the 30 minute single-channel distribution is much too narrow compared to the data , as expected from the arguments above . Is there some obvious modification of the single channel model that would predict larger spreading rates ? In general , any source of noise makes response curves broader , and the longer an auxin pulse runs , the more noise it would be expected to accumulate . However , incorporating various plausible sources of noise , such as random variation in cell length ( so that neighbouring files are no longer in register ) , and randomised permeability p , has little effect on the spreading rate ( see S4 Text ) . One might also imagine that the loading of auxin from a source could have an effect on the spreading rate . Certainly , a sustained period of loading will lead to a broader pulse . However , the rate of broadening is unchanged , whatever the temporal pattern of loading . This is because the variance of a pulse is the sum of the variance due to an instantaneous loading event and the variance of the loading distribution ( see S5 Text ) . The latter variance term , being a constant , disappears when we take the time-derivative to get the spreading rate . Another possible explanation for a large spreading rate would be saturation of polar transporters . Saturation will cause the high concentration part of a pulse to be held back while the lower concentration part can move forwards , thus stretching out the pulse . Consider the situation where only the PINs , usually considered to be the principal polar transporters , show saturation , all other import and export mechanisms being treated as linear . Thus the flux due to PINs obeys Michaelis-Menten kinetics of the form flux = Vmax a/ ( Km + a ) . We simplify the situation by neglecting ABCB efflux . Then instead of Eq ( 7 ) we obtain ϕ = κ ( 1 ) a 1 K m + a 1 - κ ( 2 ) a 2 K m + a 2 , ( 13 ) where κ ( i ) =Vmax ( i ) / ( 2+r ) , and r = βL0/D0 . This expression for flux will come in useful later . For now , to illustrate the basic qualitative behaviour , we simplify even further by assuming that PINs are exclusively present in the basal ends of cells , so κ ( 2 ) = 0 . This model can easily generate the required large spreading rates , by the stretching process described above . However , it leads to a characteristically-shaped pulse , with an abrupt edge at the apical end and an extended basal tail; see Fig 4 . This is quite unlike the symmetrical gaussian which , as we have seen , gives a good fit to the peak region . This point is discussed in Renton et al . [22] ( see their Fig 7c and 7d ) . In their model II , saturation of the auxin carriers can produce broad enough peaks to match the data , but the shape of the peak is then incorrect . This suggests that saturation does not explain the spreading rate of pulses in the experiments discussed so far . However , Brewer et al . [50] showed , by using sources with differing auxin concentrations , that it was possible to change the total uptake by a factor of ten in Arabidopsis and almost a hundred in pea . Under these conditions , effects of saturation are seen . However , as will be shown later , these effects point to saturation of lateral transport , which brings us to the topic of auxin movement between multiple channels . It appears not to be possible to generate a plausible match to the data with a single channel , and the natural next step is to consider a number of channels . If one has a collection of independent channels with a scatter of velocities , one might expect them to combine to give a pulse that grows broader as its components drift apart . However , it is unlikely that the channels are truly independent . Various experiments , such as those of Sachs [8 , 9] on the induction of vascular strands by external application of auxin , suggest that auxin can move fairly freely , at least through some tissues . One can model coupling of neighbouring cells by expressing the lateral flux per unit area formally in the same way as the axial flux , using Eq ( 1 ) , or Eq ( 13 ) if there is saturation . In this section we assume that there is no saturation and also that the coupling is symmetric , so one can write ϕ = s ( a n - 1 - a n ) , ( 14 ) where s is the symmetric lateral permeability and an−1 and an denote auxin concentrations in cells that lie side-by-side . Consider first a minimal model that illustrates lateral coupling . It consists of two adjacent channels , one polar and the other apolar . The situation is depicted in Fig 5A . When s is large enough the two channels are synchronised , producing a single sharp pulse travelling at their average velocity; see the curve for log s = −3 in Fig 5B . As the lateral permeability decreases , the single peak eventually separates into two distinct peaks ( curve for logs = −7 ) . However , there is an intermediate region where the peak broadens while still keeping an essentially gaussian form ( curve for log s = −5 ) . To explore this further , we observe how the velocity of the pulse in a channel varies with the lateral permeability ( see Fig 5C ) . Starting with large s , the two channels have the same velocity . As s decreases , a critical point is reached where the velocities begin to separate , and with sufficiently small s the channels behave essentially separately , the apolar channel pulse having zero velocity . Fig 5D shows the spreading rate , which rises as the lateral permeability decreases , and then eventually declines . The dotted region on the curve corresponds to a range of permeabilities where the variance does not depend linearly on time , so the spreading rate is not properly defined . This is due to a “streak” of auxin left behind the fast moving peak , which causes the variance to increase slightly faster than linearly . However , before this dotted segment is reached , the spreading rate is well-defined and can match those seen experimentally . We use Goldsmith’s data as a test set for various models . The vertical line in Fig 5C and 5D marks the value s = 7 . 1 × 10−6 cm/sec , where the 2-channel model gives a reasonable fit ( by eye ) to her Fig 1A and 1D at zero time and 30 minutes , respectively . Fig 6A and 6B show the corresponding auxin distributions . One can scale this two-channel model up to larger numbers of channels , where there is a group of polar channels adjacent to a group of apolar channels ( see the thumbnail sketches in Fig 6 ) . This might be the situation , for example , when there is a specialised transporting tissue adjacent to a tissue where auxin moves diffusively . In these scaled-up models , the lateral coupling is assumed to be symmetric and of equal strength , s , everywhere . Thus one can regard ws , where w is the width of a cell , as a diffusion coefficient for lateral movement . More precisely , diffusion within cells with rate D will combine with the intercellular coupling [51] so the effective diffusion constant , Deff , is given by: 1 D e f f = 1 D + 1 w s . ( 15 ) Now one intuitive explanation for the broadening of a pulse by coupling of channels is that it is due to the lateral diffusion of auxin . The diffusion distance , or the average distance that a molecule diffuses in time t , is 2 D t , ( e . g . expression ( 3 . 40 ) in [52] ) . If one is comparing models with varying numbers of channels , and hence varying total width L , the time needed to diffuse the distance L is given by L = 2 D t , or t = L 2 4 D . So if the amount of broadening depends on this time t , we expect models to give similar shapes of pulses when the diffusion constant , Deff , is proportional to L2 , or the square of the number of the channels . One can test this by finding the best matches to our test set for various numbers of channels; see Fig 6 . The resulting values of Deff do indeed scale in a quadratic manner , and the same is true for s when ws ≪ D , since the term 1/ ( ws ) then dominates 1/D in Eq ( 15 ) . However , the scaling only holds approximately ( Table 2 ) . It seems this intuition only partially captures the complicated dynamics . Note that there is a limit to lateral movement set by diffusion within the cytoplasm . From an engineering point of view , to achieve rapid lateral movement without wasted resources , the permeability s for movement between cells should be matched to diffusion within them . This means that the equivalent diffusion constant , ws , ( where w is the width of a cell ) should be equal to the diffusion constant D for auxin in cytoplasm , or equivalently that the two terms in Eq ( 15 ) should be equal . Taking w = 20 microns and D = 5 × 10−6 cm2/sec , this gives s = 2 . 5 × 10−3 cm/sec , which is not far from the best-fitting value for the 21-channel model ( Table 2 ) . With this value , Deff = 2 . 5 × 10−6 cm2/sec , which means that auxin can move rapidly enough through tissues to register changes at distances of the order of a millimetre within an hour ( since t = L2/Deff with L = 1 mm is 1 . 1 hrs ) , which seems biologically reasonable . The biological parallel for this engineering argument would be that evolutionary pressure to achieve rapid signalling would lead to increasing values of the permeability s . However , at the point where ws = D , increasing s gives ever smaller gains in Deff , and selection for larger values of s becomes weak . There are many other ways of varying the minimal model that still allow a good fit to Goldsmith’s data . For instance , one can consider two channels that are both polar , but with strengths that are in some ratio α : 1−α ( Fig 7A ) . One can plot the spreading rate against coupling strength ( as in Fig 5D ) , and Fig 7B shows a number of such graphs for different values of the ratio α : 1−α . As can be seen , the maximum spreading rate decreases as the polarities of the two channels become more similar . This implies that one can only find a match to Goldsmith’s data when the polarities are sufficiently different . Fig 7C shows that there is a match for 0 . 8:0 . 2 , but not for 0 . 7:0 . 3 , where the peak splits as the coupling is weakened in an attempt to get a sufficiently large spreading rate . The spreading of a pulse due to lateral movement through the cells of a tissue could in principle also be produced by local movement within a cell: instead of looking at coupling between cells , we can look at coupling between subcellular compartments . One obvious example is the vacuole , which takes up most of the volume of a plant cell , in mature tissues at least . We can represent the cytosol/vacuole pair by the minimal model , i . e . Fig 5A , with n = 2 . The cytosol is the polar channel and the vacuole the apolar channel; they are coupled laterally via the tonoplast ( Fig 8 ) . One difference is that there is no coupling between successive cells in the vacuolar channel , i . e . p = q = 0 for apical and basal faces , whereas the apolar channel in the minimal model has the same permeability on all cell faces ( See Models , Fig 6 ) . However , setting the apical and basal permeabilities to zero has only a very small effect on the dynamics , especially for the case n = 2 where s is small ( Table 2 ) . Another example is the endoplasmic reticulum ( ER ) . Auxin is thought to be moved from the cytosol to the ER by PIN5 [27] , and PIN6/PIN8 may move auxin in the reverse direction [28] . Again we can represent the cytosol/ER pair by the minimal model , with PINs 5/6/8 providing the lateral coupling . The question is whether the minimal model , applied to these compartments , can account for the observed spreading rate , or for some fraction of it ( lateral movement between and within cells could both be operating at the same time ) . Consider first the vacuole . Two new features have to be added to the minimal model: asymmetry in the size of the compartments , and also a possible asymmetry in the inward and outward permeabilities , so in place of Eq ( 14 ) , or ϕ = s ( a1−a2 ) for the flux ϕ from channel 1 ( cytosol ) to channel 2 ( vacuole ) , we have ϕ = scyt a1−svac a2 , where scyt is the permeability from cytosol to vacuole , and svac is the permeability in the opposite direction . In the minimal model it was assumed that each channel ( cell file ) was 20 microns wide . Let us assume that the cytosol and vacuole have widths dcyt and dvac microns , respectively . We can rescale the permeabilities to take account of these widths , so that the dynamics of the best-fitting minimal model will be reproduced if we take scyt = s ( dcyt/20 ) , svac = s ( dvac/20 ) , where s is given by Table 2 , n = 2 , as 7 . 1 × 10−6 cm/sec . Let us pause to consider the geometry of the cell and its vacuole . A cell is represented computationally by a brick shape whose long axis measures 100 microns and the two sides 20 microns; thus it has a square cross-section . The cytoplasm is assumed to form a thin layer , 1 micron in depth , just inside the boundary of the cell . Thus the vacuole to cytoplasm volume ratio is 182 × 98:202 × 100−182 × 98 , or about 4:1 . This is low compared to the average of about 10 : 1 [53] , but probably more typical for the elongated xylem parenchyma cells that transport auxin efficiently in the stem . In comparing dcyt and dvac , we need to take into account that there are two layers of cytoplasm on opposite sides of the cell , so we can pair each layer of cytoplasm with the adjacent half of the vacuole , whose width is half that of the whole vacuole , namely 9 microns . Thus we take dcyt = 1 and dvac = 9 . With these values , and with the best-matching s we get scyt = s ( dcyt/20 ) = 3 . 6 × 10−7 cm/sec and svac = s ( 9dvac/20 ) = 3 . 2 × 10−6 cm/sec . Thus we obtain a good match to Goldsmith’s data if scyt and svac have the above values . We can compare them with the values expected from passive diffusion . With a pH for the cytoplasm of 7 . 2 [43] ) , and using pIAAH = 4 × 10−5 cm/sec [45] , we get scyt = 1 . 2 × 10−7 cm/sec . Similarly , taking the pH of the vacuole to be 5 . 5 [42 , 43 , 54] , we find svac = 5 . 6 × 10−6 cm/sec . As can be seen , these are of the same order as the best-fitting values . However , scyt is about 3 times smaller and svac almost twice as large as these best-fitting values , so there is a bias towards excluding auxin from the vacuole . These calculations show that , given passive diffusion , only about a quarter of the auxin in a cell resides in the vacuole , despite its large relative volume . The effect of this exclusion is that coupling with the vacuole produces only a small amount of broadening of the pulse: about 2 . 6 mm2/hour with passive diffusion . In addition to passive diffusion there may also be auxin transporters in the tonoplast; indeed , there is evidence for an exporter from the vacuole , the WAT1 transporter [55] ( though see also [56] ) . It seems possible that a combination of low pH and exporters ensures that auxin is largely excluded from the vacuole , in which case its contribution to the spreading rate will of course be negligible . We take a final look at the geometry of the vacuole to explain why it is plausible , when looking at lateral movement between cells , to treat each cell as having a 20 micron width , despite the vacuole filling much of the space . The reason is that one can view lateral flux as taking place in sheets of cytoplasm 1 micron thick but 20 microns wide within each cell , and with corresponding sheets in adjacent cells ( marked in red in Fig 9 ) . The cytoplasmic sheets at right angles serve as an extended interface between adjacent cells . This is of course a gross oversimplication , but at least provides a tractable computational model . Turning next to the ER , if auxin is both imported and exported by specific channels [28] , then the permeabilities are likely to be much larger than for passive diffusion , so we will be in the left-hand part of the graph in Fig 5D where the spreading rate is low . Thus if we regard the ER simply as a space through which auxin diffuses , it is unlikely to explain the spreading rate . However , there is evidence that auxin undergoes metabolic processing once it enters the ER [27] , and this could plausibly account for the frequently observed accumulation of label in the wake of a pulse ( see for instance Fig 8 in [14] ) . We now return to the question of saturation . The pea data of Fig 4A–4E in Brewer et al . [50] , kindly supplied by Dr Brett Ferguson ( S1 Data ) , are redrawn in Fig 10A . They show how the shape of a pulse changes when increasing amounts of auxin are applied and larger amounts of auxin taken up by plant tissues ( determined by the total counts in the stem segments ) . There is a gradual build-up of auxin near to the apical end of the stem which gradually swamps the peak as the uptake increases . This does not resemble the effect of saturation of the polar transporters in a single channel model described earlier , where the trailing edge of the pulse remains sharp as the uptake increases ( Fig 4 ) . A better match to the experimental data can be obtained by assuming that there is saturation of the lateral movement between channels . As a first try at modelling this , S6 Text , we add lateral saturation to the n = 6 version of the model shown in Fig 5 ( with the minor modification that we make the two polar channels of different strengths , which gives a better eventual fit ) . We assume that auxin initially moves from the externally applied source into the apolar channels , from which it can move laterally into the polar channels . The idea is that , as the uptake of auxin increases , lateral transport will be limited by saturation , and a diffusion gradient of residual auxin in the apolar channels will be created , accounting for the apical build-up seen in the data . There is no saturation of the axial ( polar ) transporters , only of those that move auxin laterally . S6A Fig shows how this model behaves . The 26 . 5 ng peak is retarded relative to the peaks with smaller loadings . There is also a marked apical build-up with the 95 . 9 ng pulse . However , there is little or no apical build-up with the lower uptakes , as there is in the pea data . Weakening the lateral coupling partially brings this about but distorts the curves in other ways . However , selectively weakening of the lateral coupling between the fifth and sixth channels , as shown in Fig 10B , does achieve a fairly good fit to the data . Thus a model of modest complexity can capture the qualitative features of the family of data curves . If we take this model and keep all parameters unchanged except that we switch the saturation from lateral to axial transporters , the simulated pulses are as shown in Fig 10C . This is clearly a very poor fit to the data , with large shifts in the positions of peaks , like those seen in Fig 4 , and no build-up at the apical end .
Finding an explanation for the large spreading rate of auxin pulses turns out to be an intriguing puzzle . One can rule out a single-channel model , and also variants of it that introduce noise , e . g . by randomising cell length or the degree of polarity in successive cells in the channel . These models fail by a large margin to achieve a large enough spreading rate . I propose here that the explanation lies in the coupling together of several , perhaps many , channels of different degrees of polarity . When the coupling strength is large , the channels are essentially bound together into one “super-channel” that has too sharp a peak to match the data . When the coupling is sufficiently weak , pulses in the different channels move at different velocities and separate . There is , however , an intermediate range of lateral coupling strength where the pulses in all the channels move coherently but the resulting pulse spreads out . It is this intermediate coupling regime that matches the data best . A good fit can only be obtained if the coupled channels have sufficiently disparate polarities . This might mean , for example , that some of them correspond to specialised polar transport tissue and others to apolar , or very weakly polar , tissues . Alternatively , the specialised transport tissues themselves might consist of channels with varying degrees of polarity , due to their stage of maturation for example . There is also a wide range of possible physical scales that can produce the observed phenomena , ranging from a minimal model of two channels ( with a total width of 40 microns ) to extended models with 20–30 parallel channels ( 0 . 4 to 0 . 6 mm ) . This brings home the fact that there are many ways of modelling the same data , which is hardly surprising , given that one is only fitting two parameters: the velocity and spreading rate of a pulse . There are more constraints once one brings saturation into the picture and is required to explain not just the peak but also the build-up at the apical end when large amounts of auxin are taken up , as with the data from Brewer et al . ; Fig 10A ) . The six-channel model of Fig 6 can be adapted to give a good match to the data , by introducing saturation of lateral permeabilities and weak coupling between some of the channels , which allows them to act as a slowly emptying reservoir of auxin that is waiting to enter the rapid transport stream ( Fig 10B ) . However , even with this more complex scenario , the model is not unique . An alternative model , S6 Text and S6B Fig , can achieve as good a match . In this model there is an asymmetry in the lateral permeabilities , which ensures that auxin can only enter , and not exit from , the polar channels . There is some evidence for this kind of polar lateral transport ( e . g . [57] , Fig 1 ) , and it could in principle have a strong influence on the lateral distribution of auxin . Thus the difference between the two models bears on the problem of balancing basipetal flow with lateral propagation . Although the channels have been envisaged as files of cells , it is also possible that they could be compartments within cells . However , it has been argued that two conspicuous compartments , the vacuole and the endoplasmic reticulum , are unlikely to play this role , the former because the coupling with the cytosol is too weak , the latter for the opposite reason . Thus the cell file interpretation seems more plausible , and one attractive conjecture is that the 21-channel expanded version of the minimal model , with a corresponding physical width of 0 . 4 mm , comes closest to reality for the coleoptile , as it fits not only Goldsmith’s data but also accords with the ‘engineering’ criterion of matching lateral diffusion between channels to diffusion within cells . An intriguing final observation is that the best-fitting parameters for Goldsmith’s data imply a coupling strength that is close to the point where the velocities in the channels begin to separate; the vertical line in Fig 5C lies close the split . A similar conclusion applies to the data of Morris et al . [15] . This suggests that the polar transport system is only just able to carry along with it the auxin flow in the surrounding apolar tissues . This could be an indication of energetic economy , or perhaps a way of creating responsiveness to changes .
In all the models , cells are assumed to be 100 microns long and 20 microns wide . To simulate intracellular diffusion , cells are represented as rectangles divided up lengthwise into N compartments , where N was generally taken to be 5 . If lateral coupling constants are large enough so that lateral intracellular diffusion needs to be taken into account , this is done by a further lengthwise subdivision of each compartment into two . A permeability having a computational value of p′ is converted into a permeability p measured in cm/sec by p = p ′ L N Δ , where L is the cell length in cm and Δ is the time step , generally taken to be 1/20 sec . The computational parameter D′ for intracellular diffusion is converted into a diffusion constant D in cm2/sec by D = D ′ L 2 N ( N - 1 ) Δ . This makes Eq ( 11 ) equivalent to 1 v ′ = 1 p ′ + N - 1 2 D ′ ( 1 + 2 q ′ p ′ ) , as observed in [16] ( see the passage following Eq ( 40 ) ) . The constants and simulation details for the models used for generating the figures are as follows: Fig 3 . For the single channel model in this figure , p = 4 × 10−4 cm/sec , q = 2 . 08 × 10−5 cm/sec , D = 5 × 10−6cm2/sec . Fig 4A Here we take Vmax ( i ) =10−6 cm2/sec , Km = 10−3 , q = 0 , and D = 5 × 10−6 cm2/sec . The total amount of auxin uptake , treated as an instantaneous pulse , takes the values 0 . 3 , 1 and 2 in arbitrary units . The simulation runs for the equivalent of 90 minutes . B The same data as in A normalised . Fig 5 For the polar channel , p = 1 . 4 × 10−3cm/sec , q = 0; for the apolar channel p = q = 0 . For all channels , D = 5 × 10−6 cm2/sec . It could be argued that q should be comparable to s , since an apolar cell should have uniformly distributed permeabilities . However , we wish to study the effect of changing the lateral coupling alone , so q is fixed at zero . And in fact , the picture does not change qualitatively if we put q = s rather than q = 0 . Uptake of auxin was simulated by adding at each computational step a fixed amount to the concentration in all the channels at the apical end for the loading period . At the end of the loading period , the distribution was normalised to a total of 1 . We used Goldsmith’s protocol for loading , with 15 minutes of uptake at a constant rate at the apical end of the segment . This was followed by 45 minutes of transport , corresponding to Goldsmith’s 30 minute data , since she includes a 15 minute wait after loading before “zero time” . In the dotted region of the graph of spreading rate in D , a trail of auxin behind the frontal peak makes the variance increase faster than linearly , so the spreading rate is not defined . Instead , what is plotted is ( Var ( 0 ) -Var ( t ) ) /t , where Var ( t ) denotes the variance of the distribution at time t , and t is taken to be 30 mins . If one had chosen t to be 15 mins , for example , the solid-line parts of the curve would have been identical , and the dotted line part slightly lower . Fig 6 For all polar channels , q = 0 . For apolar channels we take p = 0 and put q = s , where s is the value of the lateral permeability needed to match Goldsmith’s data given the number of channels in the model . Thus permeabilities are the same on all cell faces in apolar cells . For the polar channels we take the following values of p: n = 2 , p = 1 . 4 × 10−3 cm/sec , n = 6 , p = 6 × 10−3 cm/sec , n = 21 p = 6 . 4 × 10−3cm/sec , and n = 30 , p = 8 . 8 × 10−3cm/sec . In all cells D = 5 × 10−6 cm2/sec . Loading follows the protocol given above for Fig 5 . Graphs in B were scaled so that the peak heights coincide . Fig 7 To make the graph on the left , the two-channel model is used with p = 1 . 44 × 10−3 , so pleft = αp , and pright = ( 1−α ) p , where α:1−α takes the values 1 : 0 , 0 . 9 : 0 . 1 , etc . , indicated on the graph . For both channels , q = 0 , D = 5 × 10−6 cm2/sec . To make the right-hand graph , we again take q = 0 , D = 5 × 10−6 cm2/sec in both channels , and now seek values of p and s that optimise the fit ( or come as close as possible ) , for the indicated ratios α:1−α . For 1:0 we take p = 1 . 44 × 10−3 cm/sec , s = 1 . 4 × 10−5 cm/sec . For 0 . 8:0 . 2 we take p = 9 . 6 × 10−4 cm/sec , s = 9 . 6 × 10−6 cm/sec . For 0 . 7:0 . 3 we take p = 8 . 8 × 10−4 cm/sec , s = 4 . 0 × 10−7 cm/sec . Fig 10B Saturation kinetics is used only in the lateral walls ( marked by an ‘S’ in the sketch ) . The lateral coupling between the channels is symmetric , with κ1 = κ2 = Km s in Eq ( 13 ) , where Km = 10−4 and s = 2 . 4 × 10−5 cm/sec for the green arrows and 4 × 10−7 cm/sec for the red arrows . The other ( unsaturated ) permeabilities are p = 3 . 2 × 10−3 cm/sec in channel 1 , p = 4 × 10−4 cm/sec in channel 2 , and for channels 3 to 6 p = 0 and q = 8 × 10−3 cm/sec . For all channels , D = 5 × 10−6cm2/sec . Loading is assumed to be instantaneous in the initial computational cycle and to be into channels 2 to 6 only . The computation runs for the equivalent of 4 hours . The parameters in C are as in B , but now saturation kinetics are used for the axial permeability , so in Eq ( 13 ) for channels 1 and 2 we have κ1 = Km p , κ2 = 0 . There is no saturation in the lateral coupling . | Auxin is one of the most important signalling molecules in plants . It is a key player in the development of veins and plant organs , and in responses that the plant makes to light and gravity . Yet we have a rather limited understanding of how auxin moves around plant tissues . I show here that a classic experiment , first carried out almost 50 years ago , has more information hidden in it than one might suppose . In this experiment , one studies how a pulse of radioactive auxin moves down a stem . The first lesson is that the peak of the pulse moves with a well-defined velocity , usually around 1 cm per hour . It has also been observed that the pulse spreads out as it moves , which is perhaps unsurprising , since any source of noise would be expected to have this effect . I show , though , that it is not easy to account for the degree of spreading in terms of noise . Instead , I propose that auxin travels down stems via several , probably many , coupled channels with differing transport rates . This gives one some insight into how plants manage their auxin signalling and puts some constraints on the underlying parameters . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Models"
] | [] | 2015 | The Shape of an Auxin Pulse, and What It Tells Us about the Transport Mechanism |
The mosquito Aedes aegypti was recently transinfected with a life-shortening strain of the endosymbiont Wolbachia pipientis ( wMelPop ) as the first step in developing a biocontrol strategy for dengue virus transmission . In addition to life-shortening , the wMelPop-infected mosquitoes also exhibit increased daytime activity and metabolic rates . Here we sought to quantify the blood-feeding behaviour of Wolbachia-infected females as an indicator of any virulence or energetic drain associated with Wolbachia infection . In a series of blood-feeding trials in response to humans , we have shown that Wolbachia-infected mosquitoes do not differ in their response time to humans , but that as they age they obtain fewer and smaller blood meals than Wolbachia-uninfected controls . Lastly , we observed a behavioural characteristic in the Wolbachia infected mosquitoes best described as a “bendy” proboscis that may explain the decreased biting success . Taken together the evidence suggests that wMelPop infection may be causing tissue damage in a manner that intensifies with mosquito age and that leads to reduced blood-feeding success . These behavioural changes require further investigation with respect to a possible physiological mechanism and their role in vectorial capacity of the insect . The selective decrease of feeding success in older mosquitoes may act synergistically with other Wolbachia-associated traits including life-shortening and viral protection in biocontrol strategies .
A number of the major mosquito vectors of disease , including the various anopheline species that transmit malaria and Aedes aegypti that transmits dengue and other arboviruses , do not naturally carry the common insect endosymbiont Wolbachia pipientis . Recently , A . aegypti was artificially transinfected in the laboratory with a virulent , life-shortening strain of Wolbachia , wMelPop , that is native to Drosophila melanogaster [1] . In D . melanogaster the infection is thought to shorten the insect's lifespan by over-replicating and causing rupture of host cells [2] . A . aegypti were transinfected with this Wolbachia strain with the aim of developing a novel biocontrol strategy based on reducing mosquito lifespan in the field . Such a shift in mosquito population age structure could theoretically reduce or even eliminate dengue virus transmission given that only old mosquitoes transmit dengue [3] , [4] . The wMelPop infection demonstrated a range of effects on its non-natural host including the predicted and sought after phenotypes of cytoplasmic incompatibility and life-shortening [1] . The former phenotype is common to most insect-Wolbachia associations and is predicted to serve as a driving mechanism for spread of Wolbachia infections in the field [5] . The latter is a trait that appears to be uniquely associated with the wMelPop strain [2] . The wMelPop-infected mosquitoes also exhibit increased daytime locomotor activity and metabolic rates [6] . These two Wolbachia-associated effects suggest physiological differences between infected and uninfected insects that could affect complex behaviour like mate seeking and foraging . In blood-feeding insects , performance in these complex behaviours can dramatically affect the fitness of individual mosquitoes and the frequency of disease transmission [7] , [8] . Wolbachia-associated changes in insect activity have also been previously reported for both a Drosophila parasitoid [9] and two Drosophila species [10] . Infection-induced increases in hunger and foraging rates have been demonstrated for other insect systems including microsporidia infected honeybees [11] and parasitoid infected aphids [12] . Increased hunger is one possible explanation for the activity of Wolbachia-infected A . aegypti , as a sugar water source was present in the chambers where mosquito activity was measured [6] . Wolbachia could either be causing increased hunger indirectly , by forcing the host to expend energy by mounting an immune response [13] or directly , by accessing host resources for its own nutrition [11] . To date there is no documented evidence of Wolbachia inducing an immune response in insect hosts [14] , although the effects of wMelPop infection have been less well explored . Sequencing of the Wolbachia genome , however , has revealed clear points where Wolbachia must rely on hosts due to missing or incomplete biochemical pathways , particularly with respect to amino acid biosynthesis [15] , [16] . The presence of any such Wolbachia-associated drain on host resources may historically have been missed in fitness assays performed in nutrient rich , ideal laboratory conditions [17] . Changes in insect activity due to Wolbachia could , however , simply be a by-product of the infection process . Drosophila with systemic bacterial infections , for example , commonly show altered circadian rhythms and sleep patterns [18] . In the Wolbachia system , the use of the term “virulence” has traditionally been reserved for describing the life-shortening phenotype of wMelPop [19] . Our definition of virulence may need to expand to include these other behavioural and physiological consequences of infection associated with a broad range of strains [9] , [10] . As Wolbachia infections are highly dispersed throughout the bodies of insects [2] , [20] , [21] they have the potential to alter the function of a wide range of tissues . Here we sought to examine the effects of wMelPop infection on A . aegypti blood-feeding behaviour . In both small population and individual trials , we assessed mosquito response time to human hosts , frequency of successful blood feeding , and blood-meal weights . These characteristics were compared for wMelPop infected and uninfected control mosquitoes at a range of ages . The aim was to test for evidence of Wolbachia-associated energetic drain , as revealed by increased foraging behaviour , or wMelPop-associated virulence effects on mosquito feeding . An examination of both old and young mosquitoes was critical to capture any potential increase in virulence with age as predicted by the D . melanogaster-wMelPop association [2] .
This study was conducted according to the principles expressed in the Declaration of Helsinki . The study was approved by the Medical Research Ethics Committee at the University of Queensland ( Project #2007001379 ) . All patients provided written informed consent for the collection of samples and subsequent analysis . For all experiments two laboratory lines of A . aegypti were used , the PGYP1 line , previously generated by transinfection with wMelPop and PGYP1 . tet , its Wolbachia cured control line [1] . Mosquitoes were reared at 26±2°C , RH 75% with 12 h∶12 h light/dark cycle . Larvae were fed 0 . 1 mg/larvae of TetraMin Tropical Tablets once per day . Females were separated from males at the pupal stage and placed into 300 mm3 cages for emergence at a density of 400 individuals per cage . The females were fed 10% sucrose solution ad libitum until the day before feeding trials . Mosquito lines were screened to confirm presence ( PGYP1 ) or absence ( PGYP1 . tet ) of infection every two generations using a PCR based assay . Five days after eclosion , DNA was extracted from 10 females using DNeasy spin columns ( QIAGEN , Australia ) , following the manufacture's protocol . PCR was then carried out using primers for the IS5 transposable element present in Wolbachia [22] . Reaction conditions were as follows: 0 . 01–0 . 09 µg of each DNA sample , 2 µl of 10× Buffer , 0 . 5 µl 1 mM dNTPs , 0 . 5 µl of 20 µM IS5 primers , 0 . 15 µl Taq DNA polymerase and water up to 20 µl . Samples were denatured for three minutes at 94°C then cycled 34 times for 30 seconds at 94°C , 30 seconds at 55°C and one minute at 72°C . This cycle was followed by a final 10-minute extension at 72°C . Presence of the expected size product was then confirmed by agarose gel electrophoresis . Experiments were conducted with five , 26 and 35-day-old adult mosquitoes . Behaviours were measured in either small populations ( proportion of population fed and number of attempted bites ) or for single mosquitoes ( response time to human , blood-meal weight and length of attempted bites ) depending on feasibility . The afternoon prior to each trial the required number of mosquitoes were removed from their rearing cages and stored in mesh-covered holding buckets at a density of five mosquitoes per bucket . At the same time an additional population of five mosquitoes were set aside to replace any mosquitoes that died during the starvation period . Mosquitoes were starved of sucrose but given access to water for ∼16 hours until trials began the next morning . Prior to each trial , mosquitoes were transferred from holding buckets into a 645 cm3 cage and allowed to acclimatise for 5 minutes . All human volunteers cleaned both of their forearms with 70% isopropyl alcohol wipes , rinsed their forearms with distilled water and dried them with paper towel , and placed latex gloves on both their hands before feeding . All population trials were carried out in two cages placed next to one another . One cage contained five PGYP1 mosquitoes and the adjacent cage contained five PGYP1 . tet mosquitoes . The position ( left or right ) of the two lines was alternated throughout the experiment . Volunteers inserted an arm into each of the two cages and rested their hands on buckets placed within each cage . Both the volunteer and an external assistant monitored the number of attempted bites each mosquito made on the volunteer's forearm . An attempted bite was recorded when a mosquito landed and actively attempted to probe the volunteer's skin at a location . A single mosquito could attempt to probe multiple times at a single location , but if a mosquito moved to a new position and attempted to probe again , then this new location was recorded as another attempted bite . Mosquitoes in both cages were monitored for 15 minutes before the volunteer shook their arms and withdrew both arms from the cages . Mosquito abdomens were examined for presence of a blood meal and the proportion of the population that imbibed a blood meal was recorded . This experiment was replicated with six volunteers ( 3 female , 3 male ) ×4 replicate trials for each of the three adult mosquito age classes . A single mosquito from each line was separately aspirated into a pre-weighed 1 . 5 ml microcentrifuge tube and weighed on an analytical balance . Each mosquito was then released simultaneously into the adjacent 645 mm3 cages . Alternation of cage position , mosquito settlement time and trial length were as per population trials . The volunteer inserted their arms into the cage and the times at which mosquito's made their first attempted bite ( host-seeking time ) and the length of each attempted bite were recorded by the volunteer into a voice recorder . After the trial , mosquitoes were transferred back into the tubes they were originally weighed in and the tubes were re-weighed . The weight of the blood-meal imbibed by each mosquito was then calculated . The volunteer ( male ) hosted four groups of 10 mosquitoes from each of the three age classes . All analysis was conducted using STATISTICA v8 ( StatSoft , Inc ) . The variables , host-seeking time , length of attempted bites and blood-meal weight were normally distributed . The number of attempted bites was square root transformed to achieve normality . The role of infection and age on these variables was examined using general linear mixed models . The role of human volunteer was not examined as there were only 6 replicate individuals and they were internally controlled . When infection status was significant , t-tests were then used to further identify specific differences between infected and uninfected lines within each of the three age classes . The proportion of mosquitoes obtaining a blood meal did not respond to transformation and so non-parametric Mann Whitney U-tests tests were employed instead of linear models to examine differences between infected and uninfected mosquitoes for all three ages .
If the Wolbachia-infected mosquitoes were hungrier than uninfected counterparts they might be expected to exhibit a more rapid response to the human forearm . Over the short distance in the laboratory cage environment , infected mosquitoes did not differ ( F = 0 . 10 , df = 1 , P = 0 . 77 ) in the time it took them to land on the human volunteer and initiate an “attempted bite” ( Fig . 1 ) . Age of the mosquitoes was also not a significant determinant of length of time to first “attempted bite” ( F = 0 . 99 , df = 2 , P = 0 . 43 ) . These data suggest that wMelPop does not alter mosquito capacity to sense and respond to human hosts in the laboratory-based assay . The number of “attempted bites” made by infected mosquitoes was also examined as a possible indicator of hunger . As per our methods , an attempted bite encompassed both probing and attempted probing in a particular region on the arm . Given the cage sizes and numbers of mosquitoes involved we could not visually differentiate between a probing event that broke the skin and one that did not . Infection status ( F = 13 . 37 , df = 1 , P = 0 . 014 ) , age of mosquitoes ( F = 5 . 72 , df = 2 , P = 0 . 021 ) , and the interaction between age and infection status ( F = 5 . 76 , df = 2 , P = 0 . 021 ) were significant determinants in the number of attempted bites made . In particular , Wolbachia-infected mosquitoes at 26 ( t = −3 . 70 , df = 238 , P<0 . 001 ) and 35 days of age ( t = −5 . 35 , df = 235 , P<0 . 001 ) attempted to bite more than their uninfected counterparts ( Fig . 2 ) . This was not the case for five-day-old mosquitoes ( t = −1 . 12 , df = 236 , P = 0 . 26 ) . The significant interaction between infection status and age as reported above is seen in the increase in biting attempts by infected mosquitoes in the older age classes ( Fig . 2 ) . For example , if we directly compare infected 26-day-old versus 35-day-old mosquitoes we see an increase ( t = −2 . 70 , df = 235 , P = 0 . 0073 ) in the mean number of attempted bites while this is not the case for uninfected mosquitoes ( t = 1 . 72 , df = 238 , P = 0 . 085 ) . Lastly , we also measured the length of time each mosquito spent on an attempted bite ( data not shown ) , which was not influenced by infection status ( F = 0 . 75 , df = 1 , P = 0 . 45 ) or age ( F = 1 . 68 , df = 2 , P = 0 . 26 ) of the mosquitoes . These data suggest that as Wolbachia-infected mosquitoes age they are exhibiting a greater number of attempted bites than uninfected mosquitoes , but are not spending more time on any one attempt . Blood-meal weight ( Fig . 3 ) was examined as a measure of feeding success in the infected mosquitoes . Linear models revealed that blood-meal weight could be partially explained by the infection status ( F = 87 . 07 , df = 1 , P<0 . 001 ) and age of mosquito ( F = 16 . 87 , df = 2 , P<0 . 001 ) . There was also a significant interaction between age and infection status ( F = 5 . 59 , df = 2 , P = 0 . 004 ) . The blood-meal weight of wMelPop-infected mosquitoes was smaller than uninfected mosquitoes for all ages examined , ( 5 d , t = −2 . 80 , df = 67 , P = 0 . 007; 26 d , t = −7 . 15 , df = 67 , P<0 . 001; 35 d , t = −6 . 09 , df = 66 , P<0 . 001 ) with the differential increasing with age ( Fig . 3 ) . If infected mosquitoes were on average smaller than their uninfected counterparts , then smaller blood-meal weights would also be expected . A comparison of average weights of the infected and uninfected mosquitoes' pre-blood meal indicated there were no size differences between the lines , PGYP1 and PGYP1 . tet ( df = 204 , t = 1 . 57 , P = 0 . 11 ) . The median proportion of mosquitoes that imbibed a blood meal ( Fig . 4 ) was also reduced for infected 26 ( Z = 4 . 10 , P<0 . 001 ) and 35-day-old ( Z = 5 . 39 , P<0 . 001 ) , but not 5-day-old ( Z = 0 . 83 , P = 0 . 74 ) mosquitoes relative to uninfected . These data indicate that as Wolbachia infected mosquitoes age , an increasing proportion of the population fails to successfully obtain a blood meal and that when they do feed the meals are smaller . Normally during biting a mosquito may probe unsuccessfully , but will ultimately insert its stylet into a host ( Fig . 5A ) . In this study , infected mosquitoes were observed in which the proboscis repeatedly bent as the mosquito pushed its head towards the skin while probing ( Fig . 5B , Video S1 ) . This phenotype was present only in wMelPop-infected mosquitoes , typically of 35 days of age .
We have identified changes in A . aegypti blood-feeding behaviour in response to infection with the virulent strain of Wolbachia , wMelPop . As infected female mosquitoes age they experience increasing difficulty in successfully obtaining a blood meal and the size of the blood meal becomes smaller . With increasing failure to feed , we also see an increase in the number of “attempted bites” made by the infected mosquitoes . Five-day-old females do not show decreased success in feeding , but do show decreased blood-meal size relative to uninfected mosquitoes . Infected mosquitoes , despite all of these differences , do not exhibit a change in responsiveness to human hosts . Lastly , we observed a behavioural trait we term “bendy proboscis” in wMelPop-infected mosquitoes that could potentially explain failure of individuals to successfully obtain blood meals . There are two outstanding questions generated by this study . The first is that while we quantified the numbers of “attempted bites” we were not able to differentiate the proportion of individuals whose stylet successfully pierced the skin . An understanding of this trait is critical as it directly relates to the insect's capacity to vector pathogens . A subsequent study by our laboratory is currently exploring how wMelPop infection affects the frequency of successful skin piercing as the mosquito ages . Second , given the discovery of the “bendy proboscis” phenotype late in the study in older individuals , its occurrence relative to age and infection status was not quantified . We are currently developing a quantitative relationship between the bendy trait , infection status , age , and lack of feeding success to test our hypothesis regarding its role in failed feeding . Our original aim was to use response time to humans , blood-meal size and feeding success to test for underlying differences in hunger between infected and uninfected mosquitoes . Feeding increases in response to parasites have been documented previously . In many of these cases , where the parasite is transmitted via the blood meal , increased feeding tends to be explained as a parasite manipulation of the host for improving its own transmission [23] , [24] , [25] . As Wolbachia is maternally inherited its transmission is not confounded with blood feeding . Models like that of microsporidia in honey bees [11] and parasitoids of aphids [12] are more appropriate for comparison and do show evidence of direct effects of parasites on host hunger . The heightened daytime activity of wMelPop infected A . aegypti , revealed by videography [6] , could have been explained by more frequent trips of the mosquito to the sugar water source present in the recording arena . Further studies in our laboratory also indicate that infected individuals on average take smaller sugar-water meals than uninfected mosquitoes ( unpublished data ) . Without an equal ability to obtain similar sized meals or equal probability of success with respect to feeding , basal “hunger” of infected and uninfected mosquitoes could not easily be compared . The reduced blood-feeding success and evidence of the “bendy proboscis” defect in wMelPop-infected mosquitoes , point toward a model of infection induced virulence . Disruption of host activities or damage at the level of the tissue or cell could lead to consequences for host physiological function and hence complex behaviours like feeding . In particular , host neuropeptides that regulate feeding behaviour are an appealing target for mechanistic exploration [26] . The pattern seen here of worsening effect with increasing age parallels the predicted virulence in the wMelPop:D . melanogaster association , where as bacterial densities rise , host cells lyse in association with host mortality . Indeed there are few published studies examining wMelPop properties such as growth kinetics and induced damage at the level of individual tissues . Even in D . melanogaster , 20 years after its initial report [2] , the virulence properties of wMelPop and the nature of the host response may have evolved [19] . With wMelPop now stably infecting D . melanogaster [2] , D . simulans [19] , and A . aegypti [1] and present in the somatic tissue of Anopheles gambiae [27] , these associations are ripe for comparative study at the tissue level . The suite of effects identified here in A . aegypti render the nervous tissue , musculature , proboscis and digestive tract of particular interest . This work highlights the growing diversity of Wolbachia-induced host somatic phenotypes , and offers the first report of effects for a blood-feeding insect . The nature of the behavioural changes suggests they are the result of infection-induced virulence . A determination of the physiological basis of the virulence effects requires further examination at the cellular or tissue level . The changes in feeding behaviour documented here may further improve wMelPop's ability to limit transmission of human pathogens in association with the traits of life-shortening [1] and the more recently documented dengue viral protection ( Moreira , unpublished . ) in A . aegypti . By limiting the blood-feeding success of old mosquitoes we could further reduce the period where the mosquito could function as a disease vector . As the blood-feeding effect is present only in older mosquitoes the reproductive capacity of young mosquitoes should theoretically be preserved . The balance of the onset of feeding difficulties versus the primary window of mosquito reproductive output will in part determine the virulence of the wMelPop on the host and possibly affect its long-term stability in an evolving population . | The primary mosquito vector of dengue virus , Aedes aegypti , has recently been artificially infected with a symbiotic bacterium called Wolbachia pipientis . This bacterium occurs naturally inside the cells of ∼66% of insect species . The Wolbachia used to infect A . aegypti shortens the insect's lifespan . Because only old mosquitoes are capable of transmitting dengue virus , the Wolbachia infection could theoretically reduce dengue virus transmission if infected mosquitoes were released into the wild . Here we have examined the effects of this Wolbachia infection on the mosquito's ability to obtain blood meals from human hosts . Blood is required for females to produce eggs , and so successful completion of this behaviour is necessary if Wolbachia-infected mosquitoes are to be competitive in the wild . Blood feeding on humans is also the time when viruses like dengue are transmitted , so changes in this behaviour can have consequences for the transmission rate of viruses . We show that Wolbachia-infected mosquitoes are less able to obtain blood meals , but only in old age . The reduced feeding success may be explained by a defect in the insect's proboscis . The finding is exciting as it may allow young mosquitoes to breed as normal but help reduce the lifespan and success of old mosquitoes , which are the primary transmitters of virus . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"ecology/behavioral",
"ecology"
] | 2009 | Wolbachia Infection Reduces Blood-Feeding Success in the Dengue Fever Mosquito, Aedes aegypti |
RNA interference ( RNAi ) -related pathways target viruses and transposable element ( TE ) transcripts in plants , fungi , and ecdysozoans ( nematodes and arthropods ) , giving protection against infection and transmission . In each case , this produces abundant TE and virus-derived 20-30nt small RNAs , which provide a characteristic signature of RNAi-mediated defence . The broad phylogenetic distribution of the Argonaute and Dicer-family genes that mediate these pathways suggests that defensive RNAi is ancient , and probably shared by most animal ( metazoan ) phyla . Indeed , while vertebrates had been thought an exception , it has recently been argued that mammals also possess an antiviral RNAi pathway , although its immunological relevance is currently uncertain and the viral small RNAs ( viRNAs ) are not easily detectable . Here we use a metagenomic approach to test for the presence of viRNAs in five species from divergent animal phyla ( Porifera , Cnidaria , Echinodermata , Mollusca , and Annelida ) , and in a brown alga—which represents an independent origin of multicellularity from plants , fungi , and animals . We use metagenomic RNA sequencing to identify around 80 virus-like contigs in these lineages , and small RNA sequencing to identify viRNAs derived from those viruses . We identified 21U small RNAs derived from an RNA virus in the brown alga , reminiscent of plant and fungal viRNAs , despite the deep divergence between these lineages . However , contrary to our expectations , we were unable to identify canonical ( i . e . Drosophila- or nematode-like ) viRNAs in any of the animals , despite the widespread presence of abundant micro-RNAs , and somatic transposon-derived piwi-interacting RNAs . We did identify a distinctive group of small RNAs derived from RNA viruses in the mollusc . However , unlike ecdysozoan viRNAs , these had a piRNA-like length distribution but lacked key signatures of piRNA biogenesis . We also identified primary piRNAs derived from putatively endogenous copies of DNA viruses in the cnidarian and the echinoderm , and an endogenous RNA virus in the mollusc . The absence of canonical virus-derived small RNAs from our samples may suggest that the majority of animal phyla lack an antiviral RNAi response . Alternatively , these phyla could possess an antiviral RNAi response resembling that reported for vertebrates , with cryptic viRNAs not detectable through simple metagenomic sequencing of wild-type individuals . In either case , our findings show that the antiviral RNAi responses of arthropods and nematodes , which are highly divergent from each other and from that of plants and fungi , are also highly diverged from the most likely ancestral metazoan state .
RNA interference-related ( RNAi ) pathways provide an important line of defence against parasitic nucleic acids in plants , fungi , and most animals [1–5] . In plants and fungi , which lack a distinct germline , Dicer and Argonaute-dependent RNAi responses suppress the expression and replication of viruses and transposable elements ( TEs ) through a combination of target cleavage and/or heterochromatin induction [6 , 7] . This gives rise to a characteristic signature of short interfering RNAs ( siRNAs ) derived from both TEs and viruses [8–12] . In contrast , the best-studied animal ( metazoan ) lineages display two distinct signatures of defensive RNAi . First , reminiscent of plants and fungi , arthropods and nematodes exhibit a highly active Dicer-dependent antiviral pathway that is characterised by copious virus-derived siRNAs ( viRNAs ) peaking sharply in length between 20nt ( e . g . Lepidoptera ) and 22nt ( e . g . Hymenoptera ) . These are cleaved from double-stranded viral RNA by Dicer , and loaded into an Argonaute-containing complex that targets virus genomes and transcripts via sequence complementarity [13 , 14] . Second , and in contrast to plants and fungi , animals also possess a Piwi-dependent ( piRNA ) pathway that provides a defence against TEs in germline ( Drosophila and vertebrates ) and/or somatic cells ( e . g . [15–18] ) . This pathway is usually characterised by a broad peak of 26-30nt small RNAs bound by Piwi-family Argonaute proteins , and comprises both 5'U primary piRNAs cleaved from long ‘piRNA cluster’ transcripts by homologs of Drosophila Zucchini[19] , and secondary piRNAs generated by ‘Ping-Pong’ amplification . This pathway is thought to target TE transcripts for cleavage and genomic copies for heterochromatin induction in most animals [20] . The presence of abundant viRNAs in infected plants , fungi , nematodes , and arthropods suggests that Dicer-dependent antiviral RNAi is an ancient and conserved defence [1 , 2] . However , RNAi has been entirely lost in lineages such as Plasmodium [21] , some trypanosomes [22] , and some Saccharomyces [23] , and/or extensively modified in others . For example , antiviral RNAi was long thought to be absent from vertebrates [24 , 25] , at least in part because their viRNAs cannot easily be detected by high-throughput sequencing of the total small-RNA pool from wild-type individuals [25–30] . Recently , it has been suggested that vertebrates also possess a functional virus-targeting RNAi pathway in tissues lacking an interferon response [31–33] and/or in the absence of viral suppressor of RNAi [32 , 34 , 35] . However , there is still debate as to whether this occurs under natural conditions , and whether or not it represents an immunologically relevant defence ( compare [36 , 37] ) . Despite this clear interest in the phylogenetic distribution of antiviral RNAi , comprehensive experimental studies of antiviral RNAi in animals are not available . Instead , studies have focussed on arthropods such as insects ( reviewed in [38 , 39] ) , crustaceans ( [40] , and reviewed in [41] ) , chelicerates [42] , and on nematodes [43–45] and vertebrates [25 , 26 , 28 , 29 , 31–35 , 46] . In particular , there have been few attempts to identify viRNAs in ‘early-branching’ animal lineages such as Porifera or Cnidaria , in divergent Deuterostome lineages such as Echinodermata or Urochordata , or in Lophotrochozoa ( including the large phyla Annelida and Mollusca; See Fig 1 for the known distribution of RNAi-pathways across the Metazoa ) . Broadly consistent with a wide distribution of antiviral RNAi , Argonaute and Dicer genes are detectable in most animal genomes ( Fig 1; [47–50] ) . However , while Dicer and Argonaute genes would be necessary for an antiviral RNAi response , their presence is insufficient to demonstrate one , for two reasons . First , these genes also have non-defensive roles such as transcription regulation through miRNAs ( see [51 , 52] ) —and a single gene can fulfil multiple roles . For example , whereas in Drosophila there is a distinction between the Dcr2-Ago2 antiviral pathway and the Ago1-Dcr1 micro-RNA ( miRNA ) pathway [53] , in C . elegans a single Dicer is required for the biogenesis of both miRNAs and viRNAs [Fig 1; [43 , 54 , 55] ) . Second , RNAi pathways are labile over evolutionary timescales , with regular gene duplication , loss , and change of function [18 , 56–58] . For example , the Piwi-family Argonaute genes that mediate anti-TE defence in animals were ancestrally present in eukaryotes , but were lost independently in plants , fungi , brown algae , most nematodes , and dust mites [2 , 47 , 57–59] . In contrast , non-Piwi Argonautes were lost in many alveolates , excavates and Amoebozoa [59 , 60] while Piwi genes were retained in these lineages . At the same time , new RNAi mechanisms have arisen , such as the 22G RNAs of nematodes [57 , 61 , 62] , the recent gain of an antiviral role for Piwi in Aedes mosquitoes [63 , 64] , and the RNAi-mediated immune memory of some dipterans [65–67] . Taken together , the potential for multiple functions , and for gains and losses of function , make it challenging to confidently predict the phylogenetic distribution of antiviral RNAi from the distribution of the required genes alone ( see [49] ) . Thus , although antiviral RNAi is predicted to be shared by most extant eukaryotes ( see [68 , 69] ) , in the absence of experimental studies , its distribution across animal phyla remains largely unknown ( Fig 1 ) . This contrasts sharply with our knowledge of other RNAi-related pathways , such as the miRNA mediated control of gene expression , which is conserved across plants , brown algae , fungi , and almost all animals [70] , and the presence of TE-derived piRNAs in most animals: Porifera [15 , 71] , Cnidaria [15 , 72] , Ctenophora [73] , Vertebrata [74 , 75] , Arthropoda [76–79] , some Nematoda [57 , 80] , Platyhelminthes [81] , but not Placozoa [15] . In eukaryotes that lack direct experimental evidence for viRNAs , the presence of an inducible RNAi response to experimentally applied long double-stranded RNA might indicate a potential for antiviral RNAi ( Fig 1 ) . This has been reported for Excavata [82] , Heterkonta [83] Amoebozoa [84] , trypanosomes [85] , and among animals in Porifera [86] , Cnidaria [87] , Placozoa [88] , Arthropoda [89] , Nematoda [90] , and several lineages of Lophotrochozoa including planarian flatworms [91] , bivalve molluscs [92] , rotifers [93] and annelids [94] . Thus , although circumstantial evidence suggests a near-universal potential for antiviral RNAi in animals , we still lack experimental evidence of exogenous viral processing . This knowledge gap is probably attributable , in part at least , to the challenges associated with isolating and culturing non-model animals and their natural viral pathogens in the lab . Here we seek to examine the phylogenetic distribution of viRNAs , and thus elucidate the phylogenetic distribution of a canonical ( i . e . Drosophila- , nematode- or plant-like ) antiviral RNAi response , through metagenomic sequencing . We combine rRNA-depleted RNA sequencing with small-RNA sequencing to detect both viruses and viRNAs in pooled samples of six deeply divergent lineages . This metagenomic approach circumvents the need to isolate and/or culture non-model organisms in the laboratory , and can capitalise on the high diversity of viruses naturally infecting individuals in the wild . It also avoids any artefactual outcomes that might result from non-native host-virus combinations or non-natural infection routes . First , we include two early branching animal species: a sponge ( Halichondria panicea: Porifera , Demospongiae ) and a sea anemone ( Actinia equina: Cnidaria , Anthozoa ) that branch basally to the divergence between deuterostomes and protostomes ( Fig 1 ) . Second , a starfish ( Asterias rubens: Echinodermata , Asteroidea ) that branches basally to vertebrates within the Deuterostomia . Third , two divergent species of Lophotrochozoa—the clade which forms the sister group to Ecdysozoa within the protostomes: a dog whelk ( Nucella lapillus: Mollusca , Gastropoda ) and earthworms ( Annelida , Oligochaeta ) . Finally , to explore the deep history of antiviral RNAi within the eukaryotes , we included the brown alga Fucus serratus ( Phaeophyceae , Heterokonta ) , which represents an origin of multicellularity separate from those of plants , fungi , and animals . We detect 21nt 5’U small RNAs derived from both strands of an RNA virus in the brown alga , similar to virus-derived small RNAs seen in plants and fungi , despite the deep divergence among these multicellular eukaryotes . We also detect miRNAs and somatic TE-derived piRNAs in all the animal lineages , demonstrating that our small-RNA sequencing was successful , and suggesting that somatic piRNAs represent the ancestral state in Bilateria . However , although we find RNA viruses to be common and sometimes highly abundant , we do not find abundant viRNAs in most of the sampled animals . Specifically , we detect no viRNAs from RNA viruses infecting the earthworms , the sponge , or the sea anemone , suggesting that insect- or nematode-like antiviral RNAi is absent from these lineages , and thus potentially their common ancestor . In contrast , we do detect viRNAs from RNA viruses in a gastropod mollusc , the dog whelk . But , unlike the viRNAs of nematodes and arthropods , these peak broadly at 26-30nt , as would be expected of piRNAs—but they lack the 5'U or ‘ping-pong’ signature Finally , we identify primary piRNA-like 26-30nt 5'U small-RNAs derived from putatively endogenous copies of viruses in the sponge , the starfish , and the dog whelk , consistent with a piRNA-like response targeting endogenous virus copies . Together with the known differences among the antiviral RNAi response of plants , fungi , nematodes and arthropods , these findings imply that the true diversity of defensive RNAi strategies employed by eukaryotes may have been underestimated . They also suggest that antiviral RNAi may either be lacking from many animal phyla , or perhaps resembles the antiviral RNAi response reported for mammals .
Using the Illumina platform , we generated strand-specific 150 nt paired-end sequence reads from ribosome-depleted RNA extracted from metagenomic pools of each of six different multicellular eukaryotes: the breadcrumb sponge ( Halichondria panacea , Porifera ) ; the beadlet sea anemone ( Actinia equina , Cnidaria ) ; the common starfish ( Asterias rubens , Echinodermata ) ; the dog whelk ( Nucella lapillus , Mollusca ) ; mixed earthworm species ( Amynthas and Lumbricus spp . , Annelida ) , and a brown alga ( the ‘serrated wrack’ , Fucus serratus , Fucales , Phaeophyceae , Heterokonta ) . See S1 Table for collection data . Gut contents were excluded by dissection , and contaminating nematodes excluded by a PCR screen prior to pooling ( Materials and methods; S1 Table ) . Reads were assembled separately for each species using Trinity v2 . 2 . 0 [95 , 96] , resulting in between 104 , 000 contigs for the sponge and 235 , 000 contigs for the earthworms . Metagenomic analysis using Diamond v0 . 7 . 11 . 60 [97] and MEGAN6 [98] suggests the vast majority of these contigs derive for the intended host organism ( S1 Fig ) . For each of the six species pools , the raw meta-transcriptomic contigs generated by Trinity are provided in compressed ( gzipped ) fasta format as unannotated contigs at https://doi . org/10 . 6084/m9 . figshare . 6803885 . v1 . To identify viruses , we used Diamond to search with translated open reading frames ( ORFs ) from our contigs against all virus proteins from the NCBI nr database , all predicted proteins from Repbase [99] , and all proteins from the NCBI RefSeq_protein database ( see Materials and methods ) . After excluding some low-quality matches to large DNA viruses and matches to phage , this identified nearly 900 potentially virus-like contigs ( S1 Data ) . These matches were examined and manually curated to generate 85 high-confidence virus-like contigs between 0 . 5 and 12kbp ( mean 3 . 7Kbp ) that are the focus of this study . We have provided provisional names for these viruses following the model of Shi et al . , [100] and the sequences have been submitted to GenBank under accession numbers MF189971-MF190055 . The majority of these virus-like contigs were related to positive sense RNA viruses ( +ssRNA ) , including ca . 20 contigs from the Picornavirales , 10 Weivirus contigs , and around 5 contigs each from Hepeviruses , Nodaviruses , Sobemoviruses , and Tombusviruses . We also identified 18 putative dsRNA virus contigs ( Narnaviruses , Partitiviruses and a Picobirnavirus ) and 11 negative sense RNA virus ( -ssRNA ) contigs ( 5 bunya-like virus contigs , 3 chuvirus-like contigs , and two contigs each from Rhabdoviridae and Orthomyxoviridae ) . Our curated viruses included five DNA virus-like contigs , all of which were related to the single-stranded DNA Parvoviridae . Sequences very similar to our Caledonia Starfish parvo-like viruses 1 , 2 and 3 are detectable in the publicly-available transcriptomes of Asterias starfish species ( 101; S2 Fig ) . Although some of the virus-like contigs are likely to be near-complete genomes , including several +ssRNA viruses represented by single contigs of >9kbp , many are partial genomes representing only the RNA polymerase , which tends to be highly conserved [102] . We identified virus-like contigs from all of the sampled taxa , although numbers varied from only three in the earthworm pool to around 40 in the sponge . This may represent differences in host species biology , but more likely reflects the different range of tissues sampled [103] , and/or differences in sampling effort ( S1 Text ) . A detailed description of each putative virus is provided in S2 Table . After initially assigning viruses to potential taxonomic groups based on BLASTp similarity , we applied a maximum likelihood approach to protein sequences to infer their phylogenetic relationships . Many of the viruses derived from large poorly-studied clades recently identified by metagenomic sequencing [100 , 104] , and most are related to viruses from other invertebrates . For example , five of the sponge picornavirales were distributed across the ‘Aquatic picorna-like viruses’ clade of Shi et al . , [100] with closest known relatives that infect marine Lophotrochozoa and Crustacea . Associated with the breadcrumb sponge we identified sequences related to the recently described ‘Weivirus’ clade known from marine molluscs [100] , and from the beadlet anemone we identified sequences related to chuviruses of arthropods [100 , 104] . Some of the virus-like sequences were closely-related to well-studied viruses , for example Millport beadlet anemone dicistro-like virus 1 and Caledonia beadlet anemone dicistro-like virus 2 are both very closely related to Drosophila C virus [105 , 106] and Cricket Paralysis virus [107] . Others are notable because they lack very close relatives , or because they fall closest to lineages not previously known to infect invertebrates . These include the Caledonia dog whelk rhabdo-like virus 2 sequence , which is represented by a nucleoprotein that falls between the Rabies/Lyssaviruses and other rhabdoviruses , and Barns Ness dog whelk orthomyxo-like virus 1—for which the PB2 polymerase subunit falls between Infectious Salmon Anaemia virus and the Influenza/Thogoto virus clade ( Fig 2; the PA polymerase subunit shows similarity to the Thogoto viruses , but not other Orthomyxoviruses ) . Phylogenetic trees are presented with support values and GenBank sequence identifiers in S2 Fig , and the alignments used for phylogenetic inference and newick-format trees with support values are provided in S2 and S3 Data respectively . In addition to avoiding gut content and/or nematode contamination , we sought to provide four lines of corroborating evidence that these virus-like sequences represent infections of the targeted hosts . First , we estimated the representation of potential hosts in each pool by mapping RNA-seq forward reads to the contigs of Cytochrome Oxidase 1 ( COI , a highly expressed eukaryotic gene ) that could be identified in our assemblies . COI reads that could not be matched to the target host animals amounted to less than 0 . 2% of the target’s own COI reads in every case , arguing against substantial contamination with non-target taxa such as parasites or commensals . Contamination was higher in the brown alga , perhaps reflecting the challenge of recovering RNA from this taxon ( S1 Text ) . In this case we identified around 10 contaminating taxa , amounting to 5% of the COI reads , including taxa that we might expect to live as ectocommensals on seaweeds , such as a bryozoan with 3 . 6% and a tunicate with 1 . 2% . We also identified some cross-contamination and/or adapter-switching between libraries that shared an Illumina lane [108 , 109] , with a mean of < 0 . 2% of COI reads deriving from the other libraries in the lane . Nevertheless , an average of 99 . 78% of the mapped COI reads in each invertebrate library derived from the targeted species ( 93% in the brown alga ) , suggesting that any viruses of contaminating species would need to be at a very high copy-number to be detected and erroneously attributed to the target host ( read counts are provided in S3 Table ) . Second , we remapped reads to the 85 focal virus contigs to measure the number of virus-derived reads relative to host COI . We reasoned that sequence reads from genuine infections are likely to appear in a single host species and to have high representation , whereas viruses present only as surface or sea-water contaminants would be present at low copy-number and seen in association with the multiple hosts that were collected together . We only identified one virus present at an appreciable copy-number in more than one host pool , suggesting that our virus-like sequences do not in general represent biological or experimental contaminants , and that the majority of viruses infected only one of the sampled host species . The exception was a 1 . 3 kbp partiti-like virus contig ( Caledonia partiti-like virus 1 ) , which displayed substantial numbers of reads in both the anemone and the sponge—perhaps indicative of closely related viruses infecting these highly divergent taxa . Four viruses were present at a very high level ( >1% of COI in at least one library ) , including Caledonia beadlet anemone dicistro-like virus 2 , Millport beadlet anemone dicistro-like virus , Lothian earthworm picorna-like virus 1 , and in the brown alga , Barns Ness serrated wrack bunya/phlebo-like virus 1 . In total , 18 of the 85 virus contigs were present at >0 . 1% of host COI in at least one library , and all but 8 were present at >0 . 01% of COI ( S3 Table , S3 Fig ) . Third , we recorded which strand each RNA sequencing read derived from , as actively replicating DNA viruses and -ssRNA and dsRNA viruses generate substantial numbers of positive sense RNAs ( S4 Fig ) . As expected , all of the -ssRNA viruses in our sample ( Orthomyxoviridae , Rhabdoviridae , Bunyaviridae/Arenaviridae-like , chuvirus-like ) displayed substantial numbers of reads from both strands , consistent with active replication . We also detected negative-sense reads for many of the +ssRNA viruses , but not always at a substantially higher rate than seen for host mRNAs such as COI ( S3 Table , S4 Fig . Nevertheless , although +ssRNA viruses also produce complementary ( negative sense ) RNA during replication , the positive to negative strand ratio is usually very high ( e . g . 50:1 to 1000:1 in Drosophila C Virus ) , potentially making the negative strand hard to detect by metagenomic sequencing . These data provide strong evidence that all of the -ssRNA and dsRNA viruses we detected comprise active infections and are consistent with replication by the other viruses . Surprisingly , only one of the five DNA viruses ( Millport starfish parvo-like virus 1 ) showed the strong positive sense bias expected of mRNAs , whereas the others displayed a negative sense bias . This suggests that these parvovirus-like sequences derived from expressed Endogenous Viral Elements ( EVEs; [110] ) rather than active viral infections . Fourth , we selected 53 of the putative virus contigs for further verification by PCR ( Materials and methods; S2 Table ) . For most of these , we confirmed that the template was detectable by RT-PCR but not by ( RT-negative ) PCR , confirming that the viruses were not present in DNA form , i . e . were not EVEs ( Materials and methods; S2 Table ) . The exceptions were Caledonia dog whelk rhabdo-like virus 2 and ( as expected ) the DNA parvovirus-like contigs , which did appear in RT-negative PCR . We then estimated virus prevalence in the wild , using RT-PCR to survey all of our samples in pools of between 7 and 30 individuals . The majority of viruses had an estimated prevalence in the range 0 . 79–100% ( S4 Table ) , with some virus-like sequences present in all sub-pools of the species . These ‘ubiquitous’ sequences included Caledonia dog whelk rhabdo-like virus 2 , Caledonia starfish parvo-like virus 2 , Caledonia starfish parvo-like virus 3 , Caledonia beadlet anemone parvo-like virus 1 , and thirteen of the sponge viruses . This suggests that these sequences are common or that they are ‘fixed’ in the population , which could be consistent with integration into the host genome ( i . e . an EVE ) . However , given the sampling scheme , a sponge virus at >36% prevalence has a >95% chance of being indistinguishable from ubiquitous . In addition , with the exception of Caledonia dog whelk rhabdo-like virus 2 , none of the RNA viruses could be amplified from a DNA template . Taken together , the use of tissue dissection in RNA preparation , the distribution of viruses across sequencing pools , the host distribution of related viruses , the abundance and strand specificity of virus reads , the absence of DNA copies ( for all but one of the RNA viruses ) , and the variable prevalence in wild populations , support the majority of these sequences as bone fide active viral infections of the sampled species . Virus and TE-derived small RNAs have been well characterised in plants , fungi , and some animals , but other major eukaryotic lineages such as Heterokonta , Alveolata , Excavata and Amoebozoa have received less attention . In principle , a metagenomic approach could also be applied to these lineages , but the difficulty of collecting large numbers of individuals of a single lineage makes this challenging for single-celled organisms . Here we have taken advantage of multicellularity in the brown algae ( Phaeophyceae , Heterokonta ) to test for the presence of viRNAs using the serrated wrack , Fucus serratus . Based on a single pooled sample of tissue from 100 individuals , we identified large numbers of small RNAs with a tight distribution between 19 and 23nt , peaking sharply at 21nt ( S5 Fig ) . Almost all of the 21nt sRNAs were 5' U , as has been seen for sRNAs in diatoms ( Bacillariophyceae , Heterokonta; [111] ) and is seen for some small RNA classes in green plants [112 , 113] and fungi [114 , 115] . Although miRNAs have been described for two other brown algae , Ectocarpus siliculosus [116 , 117] and Saccharina japonica [118] , we were unable to identify homologues of known miRbase miRNAs among these reads . This may reflect a lack of sensitivity , as the miRNA complements of the studied brown algae are highly divergent [118] , and miRNAs of Fucus serratus may be sufficiently divergent again to be undetectable based on sequence similarity . In contrast , 1 . 8% of small RNAs corresponded to the subset of high-confidence TE contigs . These small RNAs were derived from both strands , but as expected given the absence of Piwi , displayed no evidence of ‘ping-pong’ amplification—with sRNAs from both strands showing a 5' U bias . Most interestingly , we also detected viRNAs corresponding to a -ssRNA bunya-like virus ( Barns Ness serrated wrack bunya/phelobo-like virus 1; Fig 3E , S6 Fig ) . Although numbers were relatively small , comprising 0 . 01% of all small RNA reads , these derived from both strands along the full length of the virus-like contig , peaked sharply at 21nt , and were almost exclusively 5'U . We did not detect a viRNA signature from a further two -ssRNA or from four dsRNA virus-like contigs , although their copy-number was very low compared to Barns Ness serrated wrack bunya/phelobo-like virus 1 ( S3 Fig , S3 Table ) . Based on our knowledge of antiviral RNA interference in arthropods and nematodes , we expected viral infections in our animal samples to be associated with large numbers of Dicer-generated viRNAs , with a narrow size distribution peaking between 20nt ( as seen in Lepidoptera; [119] ) and 22nt ( as seen in chelicerates , nematodes , and hymenopterans; [42 , 120 , 121] ) . Because animal piRNAs and viRNAs are generally modified by the addition of a 3' 2-O-methyl group , and some nematode small RNAs are generated by direct syntheses ( resulting in a 5’ triphosphate group ) our sequencing included small RNAs treated with 5' polyphosphatase ( to remove 5’ triphosphates ) and oxidised RNA ( to increase the representation of small RNAs bearing a 3' 2-O-methyl group ) . Furthermore , to ensure that we did not exclude viRNAs that had been edited ( e . g . by ADAR; [122] ) , or that contained untemplated bases ( e . g . 3' adenylation or uridylation; [123] ) , our mapping approach permitted at least two high base-quality mismatches within a 21nt sRNA . We also confirmed that remapping with local alignment , which permits any number of contiguous mismatches at either end of the read , did not substantially alter our results . We successfully recovered abundant miRNAs in all of the animal samples , with between 20% ( sponge ) and 80% ( starfish ) of 20-23nt RNAs from untreated libraries mapping to known miRbase miRNAs [124] . Consistent with the absence of a 3' 2-O-methyl group , these miRNA-like reads had much lower representation in the oxidised libraries , there comprising only 0 . 4% ( earthworms ) to 14% ( dog whelk ) of 20-23nt RNAs . We also identified characteristic peaks of small RNAs derived from ribosomal RNA at 12nt and 18nt in the sponge , at 12nt and 16nt in the sea anemone , and in oxidised libraries from all organisms . The only exception to this overall pattern was for the sea anemone , in which oxidation had no effect on the relative number of miRNAs , although did strongly affect the overall size distribution of rRNA-derived sRNAs . This suggests the presence of a 3' 2-O-methyl group in sea anemone miRNAs ( S5 Fig ) . A few small RNAs also mapped to contigs nominally identified as bacterial in origin ( S7 Fig ) but numbers were small in the Sea Anemone , Dog Whelk , and Starfish , while those the brown alga were predominantly degradation products from bacterial rRNA , and manual inspection suggests the vast majority in the sponge and earthworm derived from miss-classified TE sequences . Surprisingly , despite our identification of more than 40 RNA virus-like contigs associated with the sponge , 17 in the sea anemone , and three in the earthworms , we were unable to detect a signature of abundant viRNAs in any of these three organisms . On average , less than 0 . 002% of 17-35nt RNAs from these organisms mapped to the RNA virus contigs , and those that did map were enriched for shorter lengths ( 17-19nt ) , lacked a clearly defined size distribution , and were less common in the oxidised than non-oxidised libraries ( S5 Fig , S3 Table ) —features consistent with non-specific degradation products , rather than viRNAs . ( Note that the starfish sample lacked detectable RNA viruses , precluding the identification of RNA-virus viRNAs ) . The only metazoan sample to display a viRNA signature was the dog whelk ( Nucella lapillus ) , with 0 . 14% of oxidised small RNAs derived from four of the seven RNA virus-like contigs . These included both contigs of Barns Ness dog whelk orthomyxo-like virus 1 , Caledonia dog whelk rhabdo-like virus 1 , and Caledonia dog whelk rhabdo-like virus 2 . A Narnavirus-like contig and a very low copy-number Bunyavirus-like contig were not major sources of viRNAs . Given the absence of detectable viRNAs in the Sponge , Sea Anemone , and Earthworm , it is notable that the viRNA-producing viruses in the dog whelk were present at a much lower copy number than many viRNA-free viruses in those organisms ( e . g . Lothians earthworm picorna-like virus 1 , Barns Ness breadcrumb sponge hepe-like virus 1; S3 Fig ) . This suggests that , had viRNAs been present in those taxa , we were likely ( for many viruses ) to have been be able to detect them . Nevertheless , the virus-derived small RNAs seen in the dog whelk did not show the expected size , strand , or 2nt overhang signature of canonical Dicer-generated viRNAs ( Fig 3 , S6 Fig ) . Instead , viRNA lengths formed a broad distribution from 26 to 30nt ( peaking at 28nt ) , more consistent with piRNAs seen in the Drosophila and mammalian germlines . These small RNAs were derived almost entirely from the negative-sense ( i . e . genomic ) strand of Barns Ness dog whelk orthomyxo-like virus 1 ( Fig 3A and 3B ) and Caledonia dog whelk rhabdo-like virus 2 ( Fig 3D ) , but from both stands of Caledonia dog whelk rhabdo-like virus 2 ( Fig 3C and 3E ) . Although this size distribution is more consistent with the piRNA pathway , only those from Caledonia dog whelk rhabdo-like virus 2 ( a suspected EVE , see above ) displayed the strong 5'U bias expected of primary piRNAs ( Fig 3D ) , and none showed any evidence of ping-pong amplification . In all three cases , the putative dog whelk viRNAs were derived from the whole length of the viral genome—albeit with strong hotspots in Caledonia dog whelk rhabdo-like virus 2 . None of these findings were qualitatively altered by a requirement for perfect ( zero mismatch ) mappings , or by permitting local mapping . Relative to miRNAs , these RNA-virus derived viRNAs were much more strongly represented in the oxidised library than the untreated library , with the miRNA:viRNA ratio increasing 300-fold—consistent with the presence of a 3' 2-O-methyl group ( S5 and S6 Figs ) . DNA viruses are a source of Dicer-mediated viRNAs in arthropods and in plants , and antiviral RNAi pathways are important for antiviral immunity to DNA viruses in both groups ( reviewed in [125 , 126] ) . Although our RNA sequencing strategy was intended to detect RNA viruses , we also identified four novel parvo/densovirus-like contigs ( Parvoviridae; single-stranded DNA ) in the starfish , and one in the sea anemone . These sequences constituted a substantial source of small RNAs in both organisms , particularly the starfish—contributing 0 . 3% of small RNAs in the untreated libraries and 3 . 4% of small RNAs in the oxidised library . In four of the five cases these small RNAs were almost exclusively negative sense , were 26 to 30nt in length ( peaking at 28nt ) , and were very strongly biased toward U in the 5' position—resembling primary piRNAs ( Fig 4 ) . However , the high prevalence and/or negative strand RNAseq bias ( S4 Fig ) of these source contigs is consistent with expressed genomic integrations ( EVEs ) rather than active viral infections . In the case of Millport starfish parvo-like virus 1 , both positive and negative sense reads were detectable , the negative sense reads again displayed a strong 5' U bias , but the positive sense reads displayed a postion-10 ‘A’ ping-pong signature ( Fig 4B ) , as expected of piRNAs . Relative to miRNAs , these putative piRNAs were much more strongly represented in the oxidised library than the untreated library , consistent with the presence of a 3' 2-O-methyl group ( S5 and S6 Figs ) . Transposable elements and TE-derived transcripts represent a major source of piRNAs in the germlines of Drosophila [76] , C . elegans [127 , 128] , mice [129 , 130] , and zebrafish [75] , although the germline limitation seen in Drosophila is derived within the Arthropods [17] . Piwi-interacting RNAs are also detectable in Cnidaria and Porifera , although their tissue specificity is unclear [15] . Furthermore , TE transcripts in Drosophila and some other arthropods are also processed by Dicer to generate 21nt endo-siRNAs [17] . We therefore selected a total of 146 long , high-confidence , TE contigs from our assemblies to analyse TE-derived small RNAs ( these contigs were selected on the basis of length and similarity to repBase entries , and to best illustrate small RNA properties; contigs are provided in S4 Data ) . We identified large numbers of TE-derived putative piRNAs in the somatic tissues of all the sampled organisms ( Fig 5 ) . In total , between 0 . 17% ( starfish ) and 1 . 7% ( dog whelk ) of untreated small RNA reads mapped to the 146 high-confidence TE contigs ( S2 Data , S5 and S8 Figs ) . In every case except the anemone , the putative piRNAs were more highly represented in the oxidised library than in untreated or polyphosphatase-treated libraries ( 1 . 4–6% of oxidised reads ) , suggesting that they are 3' 2-O-methylated and result from cleavage rather than synthesis . Despite very large numbers of piRNAs for some TE contigs , we did not observe endo-siRNA -like small RNAs similar those observed in Drosophila and some other arthropods ( e . g . [17 , 131] ) . We observed putative piRNAs derived from one or both strands of the TEs ( Fig 5 ) . Where they derived predominantly from a single strand they were generally strongly 5'U-biased ( consistent with primary piRNAs ) . Where they derived from both strands , those from the second strand presented evidence of ‘ping pong’ amplification ( i . e . no 5' U bias , and a strong ‘A’ bias at position ten; Fig 5 , S8 Fig ) . However , the piRNA size distribution varied substantially among organisms and TEs . In the sponge , the length of the 5' U-biased piRNAs either peaked at 23-24nt , or presented a broader bimodal distribution peaking at 23-24nt and 27-29nt . Where piRNAs derived from both strands , the strand with a ping-pong signature showed a shorter length distribution ( 22-23nt ) . In a few cases the putative sponge piRNAs from both strands showed a strong 5'U bias with no evidence of ping-pong amplification . In the sea anemone we consistently identified a strong peak of 5'-U biased sRNAs peaking at 28-29nt on one strand , but a generally bimodal distribution from the second ‘ping-pong’ strand ( if piRNAs were present ) , peaking at around 23nt and 28nt . Again , both strands occasionally displayed a 5'-U bias and no evidence of ping-pong amplification . The patterns were again similar in the starfish and the earthworms , except that size distributions were unimodal , peaking at 29-30nt in the 5'-U biased strand and 25-26nt ( starfish ) and 26-27nt ( earthworms ) in the ‘ping-pong’ strand . As with viRNAs , the only exception to this general pattern was seen in the dog whelk . In addition to TE-like contigs that displayed a classical piRNA-like signature ( 28nt 5'U RNAs from one strand; 26-28nt ‘ping-pong’ RNAs from the opposite strand ) , a small number of TE-like contigs in the dog whelk had an sRNA signature that resembled that of Barns Ness dog whelk orthomyxo-like virus 1 and Caledonia dog whelk rhabdo-like virus 1 . In these TE-like contigs , the sRNAs were derived from one or both strands , peaked broadly at 26-30nt , and lacked any bias in base composition or evidence of ‘ping-pong’ ( Fig 5E and 5F ) . This indicates that some TEs are processed in the same way as the identified RNA viruses , ( e . g . Gypsy; S8D Fig ) . A minority of TE-like contigs displayed an intermediate pattern , with a weak 5'U-bias from one strand , and a broad peak that lacked a pong-pong signature from the other strand . Such an intermediate pattern could result either from a single TE targeted by two different mechanisms , or from cross-mapping of sRNAs derived from different copies of the same TE inserted in different locations/contexts . As before , our permissive mapping approach and re-mapping using local alignments reduces the possibility that a large category of sRNAs escaped detection , and a requirement for zero mismatches had no qualitative impact on our results . We sought to examine whether the phylogenetic distribution and expression of RNAi pathway genes in our samples was consistent with the small RNAs we observed . As expected , based on the presence of abundant miRNAs and/or an antiviral pathway and given what is known for their close relatives [15 , 132–138] , we identified two deeply divergent Dicer transcripts in the sea anemone , and a single Dicer transcript in each of the other animal species . The single Dicers seen in the starfish , dog whelk , and earthworms were more similar to Dicer-1 from the Drosophila miRNA pathway than to arthropod Dicer-2-like genes that mediate antiviral RNAi . Similarly consistent with an antiviral RNAi and/or a miRNA pathway , and with what is known for their close relatives [42 , 135 , 136 , 139–143] , we identified two deeply divergent ( non-Piwi ) Argonaute transcripts in the sponge and in the anemone ( S6 Table ) , and single Argonaute transcripts in the dog whelk and in the starfish . We identified three distinct Argonaute transcripts in the mixed-earthworm species pool , although these may represent the multiple earthworm species present . The dog whelk , starfish , and earthworm Argonautes were all more closely related to arthropod Ago-1 ( which binds miRNAs but rarely viRNAs ) and to vertebrate Argonautes , than to insect Ago2-like genes that mediate antiviral RNAi . It is likely that these genes mediate the miRNA pathway in these organisms , although it is possible that they may also mediate the production of novel viRNAs seen in the dog whelk . We also identified a single Dicer and Argonaute in the Fucus , which is consistent with what has been seen in other brown algae [116–118] , and with the presence of both miRNAs and viRNAs . Host-encoded RNA-dependent RNA polymerases ( RdRp ) play a key role in antiviral RNAi responses in plants [144] and nematodes [145 , 146] , underlining the substantial diversity in antiviral RNAi pathways across eukaryotes . However , their role in RNAi in other animals is unknown , and they have an extremely patchy distribution across the animal phylogeny with multiple independent losses . For example , they are absent from Vertebrata and Pancrustacea , but are present in Porifera , Cnidaria , Chelicerata , Nematoda , Bivalvia , Brachiopoda , some Platyhelminthes , and non-vertebrate Deuterostomia . We identified three host RdRps in the Sea Anemone , each closely related to sequences from Exaiptasia pallida . We also identified a single RdRp sequence in the sponge and three in the Earthworm , although these did not cluster with their closest sequenced relatives . We were unable to identify any RdRp sequences in the dog whelk or the starfish , or in the brown alga , but it remains possible that they are present and expressed at a level too low for us to detect . In animals , the piRNA pathway suppresses transposable element transcripts , and is mediated by homologs of the Drosophila nuclease ‘Zucchini’ and the Piwi-family Argonaute proteins Ago3 and Piwi/Aub . In mammals , fish , C . elegans and Drosophila , this pathway is primarily active in the germline and its associated somatic tissues [75 , 76 , 127–130] , whereas in sponges and cnidarians—which lack a segregated germline—and many other arthropods , Piwi homologs are ubiquitously expressed [17 , 71 , 147] . Consistent with our finding of TE-derived piRNAs displaying a canonical ‘ping-pong’ signature , we identified single Zucchini , Ago3 and Piwi homologs in four of the five animals surveyed ( S6 Table ) . The exception was the sea anemone , in which we could only identify a single Piwi ( more similar to Drosophila Piwi/Aub than to Ago-3 ) . Surprisingly , although we did not identify canonical piRNAs in the brown alga , we did identify a possible Piwi-like transcript . However , its relatively low expression and apparent similarity to Piwi genes from the Lophotrochozoa suggest it most likely derives from the contaminating bryozoan identified by COI reads ( above ) . Finally , consistent with the altered small RNA profile associated with oxidation , we were able to identify a single homolog of the RNA methyl transferase Hen-1 in each of the animal species , but not in the brown alga . These sequences have been submitted to GenBank under accession numbers MF288049-MF288076 .
Antiviral RNAi is an important defence mechanism in plants and many fungi , and in nematodes and arthropods , where it generates large numbers of easily detectable virus-derived small RNAs in wild-type individuals . Here we identified abundant viRNAs from RNA viruses in two of the six multicellular Eukaryotes we tested: from a bunya/phlebo-like virus in a brown alga ( Fucus serratus ) and from three different RNA virus-like contigs in the dog whelk ( Nucella lapillus ) . The viRNAs from the brown alga strongly resembled other classes of small RNA from brown algae [117 , 118] and viRNAs from fungi [114 , 115] and some viRNAs from plants [113] , consistent with an antiviral RNAi response in this species . The viRNAs from the dog whelk similarly displayed a distinct size distribution , derived from the full length of the viral sequence , and were over-represented after oxidation—implying the presence of a 3' 2-O-methyl group ( Fig 3 , S6 Fig ) . However , their broad length distribution around 28nt and the strong strand-bias were not consistent with Dicer processing , which is expected to generate sRNAs from both strands simultaneously and to result in a characteristic sequence length determined by the distance between the PAZ and RNaseIII domains [148] . We therefore suggest that these distinctive viRNA are consistent with an active , but divergent , antiviral RNAi pathway in this species . We have also considered three alternative explanations for these data . First , it is possible that the result is artefactual , and that all of the virus-like reads derive from another unknown source , such as environmental contamination . However , the large number of complementary ( mRNA ) sequences show the -ssRNA viruses to be active , the sequences were not identified in any of the other co-collected taxa , and the COI read counts in the dog whelk show contamination rates to be low . Contamination was higher for the brown alga , but the virus would need to be at extremely high copy number in the contaminating taxon to achieve the observed 3% of brown alga COI expression . Second , it is possible that the virus-like contigs represent expressed host loci , such as EVEs . However , sequences were not detectable by PCR in the absence of reverse transcription , and in the dog whelk the low and variable population prevalence means that any putative EVE must be segregating and at very different frequencies in different samples—more consistent with an infectious agent . Moreover , in a previous analysis of insect viruses , expressed EVEs were found to be rare relative to active viral infections: zero of 20 viruses identified by metagenomic sequencing in Drosophila [149] . Third , even if the virus-like sequences do represent real infections , it is possible that the small RNAs do not represent an active RNAi-like response . However , their distinctive size distributions and the presence of a 3' 2-O-methyl group in the dog whelk and near 100% 5'U in the brown alga , argue strongly that these viRNAs are the result of active biogenesis rather than degradation . In contrast , it seems probable that the shorter rhabdo-like virus fragment from the dog whelk ( Caledonia dog whelk rhabdo-like virus 2; Fig 3D ) is a host-encoded EVE . First , the only open reading frame is homologous to a nucleoprotein and we could not detect a polymerase—despite its close relationship with the nucleoprotein of Lyssaviruses ( Fig 2A ) . Second , RNA sequencing was dominated by negative-sense reads , suggesting a lack of mRNA expression , but consistent with host-driven expression of an integrated locus . Third , the small RNAs were exclusively negative-sense and 5'U , as sometimes seen for primary piRNAs derived from EVEs in other taxa . Fourth , the sequence was ubiquitous in our population samples , consistent with fixation and thus genome integration . Fifth , we were able to PCR amplify a band from a DNA template . If this sequence is an EVE , this could represent an alternative antiviral RNAi mechanism , akin to the piRNA-generating EVEs seen in Aedes mosquitoes [150] . Despite the presence of more than 70 high-confidence RNA virus-like contigs , we were unable to identify an abundant or distinct population of viRNAs from RNA viruses in the sponge , sea anemone , or earthworm samples ( the starfish sample lacked detectable RNA viruses ) . Whereas the -ssRNA viruses in the dog whelk produced 1–100 viRNA reads per RNAseq read ( S9 Fig ) , and Barns Ness serrated wrack bunya/phelbo-like virus 1 in the brown alga produced ca . 0 . 1 viRNA reads per RNAseq read ( S9 Fig ) , none of the other RNA viruses gave rise to ≥0 . 001 viRNA reads per RNAseq read . In contrast , in an equivalent analysis of Drosophila , all putative viruses produced viRNAs at approximately 10–1000 viRNAs per RNAseq read [149] . This represents a striking difference in the processing of RNA viruses between arthropods [17 , 38 , 39 , 42] and nematodes [44 , 45 , 120] , and the processing of viruses by sponges ( Porifera ) , anemones ( Cnidaria ) , and earthworms ( Annelida ) . Importantly , it suggests that these animal lineages either do not process RNA viruses into small RNAs in the way that plants , fungi , nematodes or insects do , or that they do so at a level that is undetectable through the bulk small RNA sequencing of wild-type organisms and viruses—as has been reported to be the case for mammals [31 , 33–35] . The broad distribution of dsRNA-inducible gene knockdown reported across the animals ( Fig 1 ) may support the latter ( cryptic small RNAs ) explanation . However , it is also possible that these knockdowns function through the Dicer that mediates the miRNA pathway , as it does in C . elegans . In either case our data imply that the antiviral RNAi mechanisms seen in arthropods and nematodes are highly derived and unlikely to represent the ancestral state in Metazoa . Nevertheless , it is necessarily hard to demonstrate that RNA viruses do not give rise to small RNAs in these lineages: an absence of evidence provides weak evidence of absence . For example , it is possible that small RNAs are abundant , but were not detected . This is highly unlikely as we were able to detect miRNAs , piRNAs , and small rRNAs , and we would also have detected viRNAs bearing a 5' triphosphate or 3' 2-O-methyl group , as well as viRNAs that had been edited or extended by untemplated bases at the 5' or 3' end . One alternative is that all of the RNA-virus like contigs that we identified from the sea anemone , sponge , and earthworm , were inactive and/or encapsidated at the time of collection , and thus not subject to Dicer processing . However , this is unlikely for three reasons . First , it can be ruled out for eight of the nine highest copy-number dsRNA viruses in the sponge , as these all showed a strong positive-strand RNAseq bias , consistent with gene expression . Second , it is not supported by the two -ssRNA virus contigs in the earthworms , which also displayed positive sense mRNA reads ( although the virus copy-number was extremely low , such that that we had little power to identify either positive sense RNAseq reads or viRNAs ) . Finally , although the small number of negative sense reads resulting from +ssRNA virus replication makes it hard to exclude the possibility that they were inactive , it would be surprising if all of the -ssRNA viruses and dsRNA viruses ( including those in the dog whelk and brown alga ) were active , but none of the +ssRNA viruses were . Perhaps a more plausible alternative is that the remaining viruses express viral suppressors of RNAi ( VSRs ) that completely eradicate the small RNA signature , such that it is undetectable through bulk sequencing of wild-type individuals . This appears to be the case for some mammalian viruses , where viruses genetically modified to remove their VSR do indeed form a much greater source of small RNAs [31 , 32 , 34 , 35] . However , it is not the case for the many insect and plant viruses that express well-characterised VSRs [151 , 152] , and while it could certainly be true for some of the 80 different viruses we detect , it would be surprising if it were true for all of them . It is also possible that abundant viRNAs are characteristic of a response against -ssRNA viruses in anemones , earthworms , and sponges , but are not characteristic of the response against +ssRNA or dsRNA viruses . This could also be consistent with our failure to detect viRNAs from putative dsRNA narnaviruses in the dog whelk and brown alga , and to a putative +ssRNA nodavirus in the brown alga . If so , then an apparent absence of antiviral RNAi in the sponge , sea anemone and earthworms may really reflect differences in the composition of the RNA virus community , with a preponderance of -ssRNA viruses in the dog whelk and their absence from the sponge or anemone . However , even if -ssRNA viruses , but not +ssRNA viruses or dsRNA viruses , give rise to viRNAs in most animal lineages , then this is still in striking contrast to the antiviral RNAi response in plants , fungi , nematodes and insects [9 , 38 , 153] , and again suggests that antiviral RNAi mechanisms are highly variable among eukaryotic lineages . Finally , it also remains possible that the majority of sponge , sea anemone , and annelid species do possess an active antiviral RNAi mechanism that generates abundant viRNAs from RNA viruses , but that the particular species we examined here have lost the ability . It is certainly the case that RNAi mechanisms are occasionally lost , as in one clade of the yeast genus Saccharomyces [23 , 154] . However , unless antiviral RNAi is lost extremely frequently in these three animal phyla—which is not the case in arthropods or plants—it is extremely unlikely that we would by chance select three lineages that have lost the mechanism while others retained it . We identified four parvo/denso-like virus contigs in the starfish , and one in the sea anemone . All of these sequences were detected as RNAseq reads and were associated with abundant 26-29nt piRNA-like small RNAs ( Fig 4 ) . However , RNAseq from three of the four starfish parvo/denso-like virus contigs , and the sea anemone contig , were dominated by negative sense reads . This is hard to reconcile with the normal functioning of ssDNA parvo/denso-like viruses , and may instead reflect host-driven transcription . For these four contigs , the small RNAs were also almost exclusively negative-sense and 5'U—as expected of primary piRNAs . In contrast , RNAseq and small RNAs reads from Millport starfish parvo-like virus 1 were almost exclusively positive ( mRNA ) sense , with the negative strand small RNAs showing a 5'U bias and positive strand sRNAs showing weak ‘ping-pong’ signature ( S6 Fig ) . Together , these observations suggest that at least some of parvo/denso-like virus sequences represent expressed EVEs , but also that they are targeted by a piRNA pathway-related mechanism . Unlike for RNA viruses , we were unable to test whether these sequences represent integrations into the host genome , as integrations are indistinguishable from viral genomic ssDNA by PCR , and both +ssDNA and -ssDNA sequences are usually encapsidated by densoviruses . However , Caledonia starfish parvo-like viruses 1 , 2 and 3 are nearly identical to published starfish transcripts , and the two published sequences most similar to Caledonia beadlet anemone parvo-like virus 1 are from an anemone transcriptome and an anemone genome ( S2 Fig ) . In addition , three of the five contigs ( two in the starfish , and one in the anemone ) appear to be ubiquitous in our wild sample . This ubiquitous distribution and close relationship to published sequences support the suggestion ( above ) that some of these sequences may be host integrations . The exceptions are Caledonia starfish parvo-like virus 1 and Millport starfish parvo-like virus 1 , which both had an estimated prevalence of between 4% and 20% in the larger Millport collection . We were able to recover putatively near-complete genomes of 6 . 5 and 5 . 8 Kb , containing the full length structural ( VP1 ) and non-structural ( NS1 ) genes , from Millport starfish parvo-like virus 1 and Caledonia starfish parvo-like virus 1 , respectively ( S2 Table ) . If these sequences are EVEs , as seems very likely for four of the five , then their expression and processing into piRNAs may reflect the location of integration—for example , into or near to a piRNA generating locus [155 , 156] . In contrast , if these sequences are not host EVEs , then the high expression of negative sense transcripts and the presence of primary piRNA-like small RNAs suggests an active Piwi-pathway response targeting DNA viruses in basally-branching animals . These are not mutually exclusive , and it is tempting to speculate that such integrations could provide an active defence against incoming virus infections in basal animals , as suggested for RNA-virus integrations in Aedes mosquitoes [150] . If so , the low-prevalence Millport starfish parvo-like virus 1 sequence , which shared 72% sequence identity with Caledonia starfish parvo-like virus 1 , but displayed positive sense transcripts , positive and negative sense piRNAs and a ‘ping-pong’ signature , is a good candidate to represent an unintegrated infectious virus lineage . The absence of detectable viRNAs in the sponge , sea anemone , or earthworm samples , combined with the presence of 26-29nt ( non-piwi ) viRNAs in the mollusc and 21nt 5'U viRNAs in the brown alga , reinforces the diversity of antiviral RNAi mechanisms in multicellular eukaryotes . Previously , the abundant viRNAs present in plants , fungi , nematodes and arthropods had implied that Dicer-based antiviral RNAi was ancestral to the eukaryotes and likely to be ancestral in animals , with a recent modification [or even loss; 24] in the vertebrates—perhaps associated with the evolution of interferons [157] . Our findings now suggest three alternative hypotheses . First , antiviral RNAi may have been absent from ancestral animals , and re-evolved on at least one occasion—giving rise to the distinctively different viRNA signatures seen in nematodes , arthropods , vertebrates , and now also a mollusc . Second , the ancestral state may have been more similar to current-day mammals , which do not produce abundant easily-detected viRNAs under natural conditions , but may still possess an antiviral RNAi response [31 , 33–35] . In this scenario , antiviral RNAi has been maintained as a defence—possibly since the origin of the eukaryotes—but has diversified substantially to give the distinctive viRNA signatures now seen in each lineage . Third , dsRNA , +ssRNA , -ssRNA , and DNA viruses may be targeted differently by RNAi pathways in divergent animal lineages , but arthropods have recently evolved a defence that gives rise to the same viRNA signature from each class . It is not possible to distinguish among these hypotheses without broader taxonomic sampling and experimental work in key lineages . For example , analyses of the Ago-bound viRNAs of Cnidaria and Porifera could help to distinguish between the first two hypotheses , and an identification of the nucleases and Argonautes and/or Piwis required for the 26-29nt mollusc viRNAs could establish whether this response is derived from a Dicer/Ago pathway or a Zucchini/Piwi like pathway . In each case , the limited taxonomic sampling and a lack of experimental data from these non-model taxa preclude any firm conclusions , and given the alternative possibilities outlined above , our interpretations should be treated as tentative . Nevertheless , the balance of evidence strongly suggests that the well-studied canonical antiviral RNAi responses of Drosophila and nematodes are likely to be derived compared to the ancestral state , and that there is substantial diversity across the antiviral RNAi mechanisms of multicellular eukaryotes . The presence of piRNAs derived from transposable elements in the soma of all of the sampled animals also demonstrates a previously under-appreciated diversity of piRNA-like mechanisms . First , it argues strongly that the predominantly germline expression of the piRNA pathway in key model animals ( vertebrates , Drosophila , and nematodes ) is a derived state , and that “ping-pong” mediated TE-suppression in the soma is likely to be common in other animal phyla , as has been shown for arthropods [17] , and has recently been confirmed in two other molluscs [158] . Second , it suggests that the TE-derived endo-siRNAs seen in Drosophila and mosquitoes [64 , 155 , 159–161] are absent from most phyla , and are therefore a relatively recent innovation . Third , the diversity of piRNA profiles we see among organisms—such as the bimodal length distributions of primary piRNAs in the sponge and in “ping-pong’ piRNAs in the sea anemone—suggests substantial variation among animals in the details of piRNA biogenesis . Finally , the large numbers of primary piRNAs derived from putative endogenous copies of parvo/denso-like viruses in the starfish and sea anemone , and from the putatively endogenous rhabdo-like virus 2 in the dog whelk , suggests that the piRNA processing of endogenous virus copies may be widespread across the animals , perhaps even representing an additional ancient defence mechanism .
We sampled six organisms: The breadcrumb sponge Halichondria panacea ( Porifera: Demospongiae ) , the beadlet anenome Actinia equina ( Cnidaria: Anthozoa ) , the common starfish Asterias rubens ( Echinodermata: Asteroidea ) , the dog whelk Nucella lapillus ( Mollusca: Gastropoda ) , mixed earthworm species ( Amynthas spp . and Lumbricus spp . ; Annelida: Oligochaeta ) , and the brown alga Fucus serratus ( Heterokonta: Phaecophyceae: Fucales ) . Marine species were sampled from rocky shores at Barns Ness ( July 2014; 56 . 00° N , 2 . 45° E ) , and from three sites near Millport on the island of Great Cumbrae ( August 2014; 55 . 77° N , 4 . 92° E ) in Scotland , UK ( S1 Table , S1 Text ) . The terrestrial sample ( mixed earthworms; Lumbricus spp . , and Amythas spp . ) , were collected from The King’s Buildings campus , Edinburgh , UK ( November 2015; 55 . 92° N , 3 . 17° E ) . To maximise the probability of incorporating infected hosts , we included multiple individuals for sequencing ( minimum: 37 sponge colonies; maximum: 164 starfish; see S1 Table for sampling details , numbers ) . Marine organisms were stored separately in sea water at 4°C for up to 72 hours before dissection . After dissection , the selected tissues were immediately frozen in liquid nitrogen , pooled in groups of 5–30 individuals , and ground to a fine powder for RNA extraction under liquid nitrogen ( see S1 Text for details of tissue processing ) . Except for the brown alga Fucus serratus , RNA was extracted using Trizol ( Life Technologies ) and DNase treated ( Turbo DNA-free: Life Technologies ) following manufacturer’s instructions . For Fucus , the extraction protocol was modified from Apt et al . , [162] . Briefly , tissue was lysed in a CTAB extraction buffer , and RNA was repeatedly ( re- ) extracted using chloroform/isoamyl alchohol ( 24:1 ) and phenol-chloroform ( pH 4 . 3 ) , and ( re- ) precipitated using 100% ethanol , 12M LiCl , and 3M NaOAc ( pH 5 . 2 ) . To avoid potential nematode contamination , an aliquot of RNA from each small ( 5–30 individual ) pool was reverse transcribed using M-MLV reverse transcriptase ( Promega ) with random hexamer primers . These were screened by PCR with nematode-specific primers and conditions as described in [163] ( Forward 5'-CGCGAATRGCTCATTACAACAGC; Reverse 5'-GGCGATCAGATACCGCCC ) . We excluded all sample pools that tested positive for nematodes from sequencing , although they were used to infer virus prevalence ( below ) . For each host species , RNA from the nematode-free pools were combined to give final RNA-sequencing pools in which individuals were approximately equally represented . For the sponge , sea anemone , starfish , and dog whelk this pooling was subsequently replicated , using a subset of the original small pools , resulting in sequencing pools ‘A’ and ‘B’ ( S1 and S2 Tables ) . Total RNA was provided to Edinburgh Genomics ( Edinburgh , UK ) for paired-end sequencing using the Illumina platform . Following ribosomal RNA depletion using Ribo-Zero Gold ( Illumina ) , TruSeq stranded total RNAseq libraries ( Illumina ) were prepared using standard barcodes , to be sequenced in three groups , each on a single lane . Lanes were: ( i ) sponge , sea anemone , starfish , and dog whelk ‘A’ libraries ( HiSeq v4; 125nt paired-end reads; a Drosophila suzukii RNAseq library from an unrelated project was also included in this lane ) ; ( ii ) sponge , sea anemone , starfish , and dog whelk ‘B’ libraries ( HiSeq 4000; 150nt paired-end reads ) ; ( iii ) Fucus and Earthworms ( HiSeq 4000; 150nt paired-end reads ) . In total , this resulted in approximately 70M high quality read pairs ( i . e . after trimming and quality control ) from the sponge , 60M from the sea anemone , 70M from the starfish , 70M from the dog whelk , 130M from the earthworms , and 180M from the brown alga ( S3 Table ) . For small RNA sequencing , total RNA was provided to Edinburgh Genomics ( Edinburgh , UK ) for untreated libraries ( A and B ) , or after treatment either with a polyphosphatase ( “A: Polyphosphatase” ) or with sodium periodate ( “B: Oxidised” ) . In the first case , we used a RNA 5' Polyphosphatase ( Epicentre ) treatment to convert 5' triphosphate groups to a single phosphate . This permits the ligation of small RNAs that result from direct synthesis rather than Dicer-mediated cleavage , such as 22G-RNA sRNAs of nematodes . In the second case , we used a sodium periodate ( NaIO4 ) treatment ( S2 Text ) . Oxidation using NaIO4 reduces the relative ligation efficiency of animal miRNAs that lack 3′-Ribose 2′O-methylation , relative to canonical piRNAs and viRNAs . This permits identification of 3′- 2′O-methylated sRNA populations , and is expected to enrich small RNA library for canonical piRNAs and viRNAs . TruSeq stranded total RNAseq libraries ( Illumina ) were prepared from treated RNA by Edinburgh Genomics , and sequenced using the Illumina platform ( HiSeq v4; 50nt single-end reads ) , with all ‘A’ libraries sequenced together in a single lane , and all ‘B’ libraries sequenced together with Fucus and earthworm small RNAs , across four lanes . In total , this resulted in between 46M and 150M adaptor-trimmed small RNAs ( S3 Table ) . Raw reads from RNAseq and small RNA sequencing are available from the NCBI Sequence Read Archive under accession number SRP153010 , within BioProject accession PRJNA394213 . For each organism , paired end RNAseq data were assembled de novo using Trinity 2 . 2 . 0 [95 , 96] as a paired end strand-specific library ( —SS_lib_type RF ) , following automated trimming ( —trimmomatic ) and digital read normalisation ( —normalize_reads ) . Where two RNAseq libraries ( ‘A’ and ‘B’ ) had been sequenced , these were combined for assembly . For the mixed earthworm assembly , which had a large number of reads , high complexity , and a high proportion of ribosomal sequences ( 18% ) , ribosomal sequences were identified by mapping to a preliminary build of rRNA derived from subsampled data and excluded from the subsequent final assembly . To provide a low-resolution overview of the taxonomic diversity in each sample , we used Diamond [97] and BLASTp [164] to search the NCBI nr database using translated contigs , and MEGAN6 [98] ( long reads with the weighted lowest common ancestor assignment algorithm ) to provide taxonomic classification . In addition , for a more sensitive and quantitative analysis of Eukaryotic contamination , we recorded the number of Cytochrome oxidase reads for each reconstructed COI sequence present . To identify cytochrome oxidase 1 ( COI ) sequences , all COI DNA sequences from GenBank nt were used to search all contigs using BLASTn [164] , and the resulting matches examined and manually curated before read mapping . An analogous approach was taken to identify rRNA sequences , but using rRNA from related taxa for a BLASTn search . To identify probable virus and transposable element ( TE ) -like contigs , all long open reading frames from each contig were translated and concatenated to provide a ‘bait’ sequence for similarity searches using Diamond [97] and BLASTp [164] . Only those contigs with an open reading frame of at least 200 codons were retained . To reduce computing time , we used a two-step search . First , a preliminary search was made using translations against a Diamond database comprising all of the virus protein sequences available in NCBI database ‘nr’ ( mode ‘blastp’; e-value 0 . 01; maximum of one match ) . Second , we used the resulting ( potentially virus-like ) contigs to search a Diamond database that combined all virus proteins from NCBI ‘nr’ , with all proteins from NCBI ‘RefSeq_protein’ ( mode ‘blastp’; e-value 0 . 01; no maximum matches ) . Putatively virus-like matches from this search were retained for manual examination and curation ( including assessment of coverage—see below ) , resulting in 85 high-confidence putative virus contigs . A similar ( but single-step ) approach was used to search translated sequences from Repbase [99] , using an e-value of 1x10-10 to identify TE-like contigs . Translated open reading frames from the 85 virus-like contigs were used to search the NCBI ‘RefSeq_protein’ blast database using BLASTp [164] . High confidence open reading frames were manually annotated based on similarity to predicted ( or known ) proteins from related viruses . Where unlinked fragments could be unambiguously associated based on similarity to a related sequence or via PCR ( below ) , they were assigned to the same virus . These contigs were provisionally named based on the collection location , host species , and virus lineage . Where available , the polymerase ( or a polymerase component ) from each putative virus species was selected for phylogenetic analysis . Where the polymerase was not present , sequences for phylogenetic analysis were selected to maximise the number of published virus sequences available . For the Weiviruses , bunya-like viruses , and noda-like viruses , two different proteins were used for phylogenetic inference . Published viral taxa were selected for inclusion based on high sequence similarity ( identifiable by BLASTp ) . Translated protein sequences were aligned using T-Coffee [165] mode ‘m_coffee’ [166] combining a consensus of alignments from ClustalW [167 , 168] , T-coffee [165] , POA [169] , Muscle [170] , Mafft [171] , DIALIGN [172] , PCMA [173] and Probcons [174] . Alignments were examined by eye , and regions of ambiguous alignment at either end were removed . Phylogenetic relationships were inferred by maximum-likelihood using PhyML ( version 20120412 ) ; ( version 20120412; [175] ) with the LG substitution model , empirical amino-acid frequencies , and a four-category gamma distribution of rates with an inferred shape parameter . Searches started from a maximum parsimony tree , and used both nearest-neighbour interchange ( NNI ) and sub-tree prune and re-graft ( SPR ) algorithms , retaining the best result . Support was assessed using the Shimodaira-Hasegawa-like nonparametric version of an approximate likelihood ratio test . All trees are presented mid-point rooted . To estimate virus prevalence in the five animal taxa , we used a PCR survey of the small sample pools ( 5–30 individuals ) for 53 virus-like contigs . There was insufficient RNA to survey prevalence in the brown alga . Aliquots from each sample pool were reverse transcribed using M-MLV reverse transcriptase ( Promega ) with random hexamer primers , and 10-fold diluted cDNA screened by PCR with primers for virus-like contigs designed using Primer3 [176 , 177] . To confirm that primer combinations could successfully amplify the target virus sequences , and to provide robust assays , each of four PCR assays ( employing pairwise combinations of two forward and two reverse primers ) were tested using combined pools of cDNA for each host , with the combination that produced the clearest amplicon band chosen as the optimal assay . We took a single successful PCR amplification to indicate the presence of virus in a pool , whereas absence was confirmed through at least 2 PCRs that produced no product . PCR primers and conditions are provided in S7 Table . Prevalence was inferred by maximum likelihood , and 2 log-likelihood intervals are reported . For 47 of the putative RNA virus contigs , we used PCR to verify that the sequences were not present as DNA in our sample , i . e . were not EVEs . We performed an RT-negative PCR survey of Trizol RNA extractions ( which also contained DNA ) using the primers and conditions provided in S7 Table . Where amplification was successful from cDNA synthesised from a DNAse-treated extraction , but not from 1:10 , 1:100 , or 1:0000-fold diluted RNA samples ( serial dilution was necessary as excessive RNA interfered with PCR ) , we inferred that template DNA was absent . The remaining six ( out of a total of 53 contigs for which designed PCR assays ) were putative parvo/denso-like virus contigs , and were also tested as above . All six DNA virus contigs were detectable as DNA copies . To identify the origin of RNA sequencing reads , quality trimmed forward-orientation RNAseq reads and adaptor-trimmed small-RNA reads between 17nt and 40nt in length ( and trimmed using cutadapt and retaining adaptor triimed reads only; [178] ) were mapped to potential source sequences . To provide approximate counts of rRNA and miRNA reads , reads were mapped to ribosomal contigs from the target host taxa and to all mature miRNA stem-loops represented in miRbase [124] , using Bowtie2 [179] with the ‘—fast’ sensitivity option and retaining only one mapping ( option ‘-k 1’ ) . To identify the number and properties of virus and TE-derived reads , the remaining unmapped reads were then mapped to the 85 curated virus-like contigs , to COI-like contigs , and to 146 selected long TE-like contigs between 2kbp and 7 . 5kbp from out assemblies , using the ‘—sensitive’ option and default reporting ( multiple alignments , report mapping quality ) . For small RNA mapping , the gap-opening and extension costs were set extremely high ( ‘—rdg 20 , 20—rfg 20 , 20’ ) to exclude maps that required an indel . The resulting read mappings were counted and analysed for the distribution of read lengths , base composition , and orientation . In an attempt to identify modified or edited small RNAs , we additionally mapped the small RNA reads to the virus-like and TE-like contigs using high sensitivity local mapping options equivalent to ‘—very-sensitive-local’ but additionally permitting a mismatch in the mapping seed region ( ‘-N 1’ ) and again preventing indels ( ‘—rdg 20 , 20—rfg 20 , 20’ ) . We also reanalysed the data using only perfect ( zero mismatch ) mappings . Neither approach led to substantially different results . | The presence of abundant virus-derived small RNAs in infected plants , fungi , nematodes , and arthropods suggests that Dicer-dependent antiviral RNAi is an ancient and conserved defence . Using metagenomic sequencing from wild-caught organisms we find that antiviral RNAi is variable across animals . We identify a distinctive group of virus-derived small RNAs in a mollusc , which have a piRNA-like length distribution but lack key signatures of piRNA biogenesis . We also report a group of 21U virus-derived small RNAs in a brown alga , which represents an origin of multicellularity separate from that of plants , fungi , and animals . The absence of virus-derived small RNAs from most of our animal samples may suggest either that the majority of animal phyla lack an antiviral RNAi response , or that these phyla possess an antiviral RNAi response resembling that reported for vertebrates , which is not detectable through simple metagenomic sequencing of wild-type individuals . In addition , we report abundant somatic piRNAs across anciently divergent animals suggesting that this is the ancestral state in Bilateria . Our study challenges a widely-held assumption that most invertebrates possess an antiviral RNAi pathway likely similar to that seen in Drosophila , other arthropods , and nematodes . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"sequencing",
"techniques",
"invertebrates",
"rna",
"interference",
"vertebrates",
"animals",
"mammals",
"dogs",
"viruses",
"rna",
"viruses",
"epigenetics",
"molecular",
"biology",
"techniques",
"rna",
"sequencing",
"echinoderms",
"starfish",
"plants",
"research",
"and",
"analysis",
"methods",
"genetic",
"interference",
"gene",
"expression",
"molecular",
"biology",
"algae",
"biochemistry",
"rna",
"eukaryota",
"sea",
"anemones",
"nucleic",
"acids",
"cnidaria",
"genetics",
"biology",
"and",
"life",
"sciences",
"dsrna",
"viruses",
"amniotes",
"organisms"
] | 2018 | Metagenomic sequencing suggests a diversity of RNA interference-like responses to viruses across multicellular eukaryotes |
Bovine tuberculosis ( BTB ) is an endemic zoonosis in Morocco caused by Mycobacterium bovis , which infects many domestic animals and is transmitted to humans through consumption of raw milk or from contact with infected animals . The prevalence of BTB in Moroccan cattle is estimated at 18% , and 33% at the individual and the herd level respectively , but the human M . bovis burden needs further clarification . The current control strategy based on test and slaughter should be improved through local context adaptation taking into account a suitable compensation in order to reduce BTB prevalence in Morocco and decrease the disease burden in humans and animals . We established a simple compartmental deterministic mathematical model for BTB transmission in cattle and humans to provide a general understanding of BTB , in particular regarding transmission to humans . Differential equations were used to model the different pathways between the compartments for cattle and humans . Scenarios of test and slaughter were simulated to determine the effects of varying the proportion of tested animals ( p ) on the time to elimination of BTB ( individual animal prevalence of less than one in a thousand ) in cattle and humans . The time to freedom from disease ranged from 75 years for p = 20% to 12 years for p = 100% . For p > 60% the time to elimination was less than 20 years . The cumulated cost was largely stable: for p values higher than 40% , cost ranged from 1 . 47 to 1 . 60 billion euros with a time frame of 12 to 32 years to reach freedom from disease . The model simulations also suggest that using a 2mm cut off instead of a 4mm cut off in the Single Intradermal Comparative Cervical Tuberculin skin test ( SICCT ) would result in cheaper and quicker elimination programs . This analysis informs Moroccan bovine tuberculosis control policy regarding time frame , range of cost and levels of intervention . However , further research is needed to clarify the national human-bovine tuberculosis ratio in Morocco .
Bovine tuberculosis ( BTB ) is a zoonotic bacterial infection caused by Mycobacterium bovis . It belongs to a group of well-known and newer mycobacteria , together with Mycobacterium tuberculosis , all of which derive from a common ancestor forming the Mycobacterium tuberculosis complex ( MTBC ) [1–4] . M . bovis is capable of infecting a broad range of hosts , including ruminants ( predominantly domestic cattle ) , humans and other primates [5–10] . The wide host range makes BTB highly relevant to conservation projects and difficult to eliminate where wildlife reservoirs are involved , for instance , badgers ( Meles meles ) in the United Kingdom [11 , 12] . Bovine tuberculosis infection in cattle is a chronic disease which first affects the lymph nodes and from weeks to decades later can affect lungs . The disease can also be manifested in other organs such as , mammary tissue , and the gastrointestinal or urinary tract . Since transmission between cattle occurs predominantly through aerosol inhalation [13–17] , the transmission rate is increased by risk factors such as high herd density and intensive breeding [18] . Pseudo-vertical transmission from cows to suckling calves through infected milk has been described [19] . Factors like a long survival period for the microbes in manure and soil and close contact between animals , for example around water sources , also contribute to an increased risk of infection [20 , 21] . In humans , contaminated dairy products are considered to be the main source of BTB infection , usually resulting in extra-pulmonary infection such as lymphadenitis [22–24] . These patients are missed by thoracic radiographic screening and the resulting diagnostic cascade [25] . Aerosol cattle-to-human transmission can occur during close contact with infected animals , posing an occupational risk , especially for pastoralists and farmers [2 , 26] . Infection risks linked to local cultural practices , for instance consumption of fresh blood , are reviewed by Daborn [27] . There is evidence that human patients can transmit BTB to animals , and human to human transmission occurs [28 , 29] . There is a bottleneck in detecting human BTB cases because the routine diagnostic protocols were developed for patients with pulmonary tuberculosis , as caused by M . tuberculosis . Tuberculosis ( TB ) and BTB cannot be distinguished on the basis of clinical symptoms , radiography or histopathology [30] . Glycerol-containing Löwenstein-Jensen medium , the long-time gold standard for TB culture , inhibits the growth of M . bovis , thereby increasing the number of undetected cases [31] . New molecular diagnostic tools , for example spoligotyping , and even whole genome sequencing have been developed for M . bovis detection [32 , 33] . Although they require enhanced laboratory infrastructure and personnel training which are not currently available in some developing countries , these new techniques offer promise for epidemiological research , control and adequate treatment , particularly since M . bovis is resistant to pyrazinamide , one of the first-line antibiotics for TB treatment [34] . Morocco is transitioning from extensive pastoralist livestock and dairy production to more intensified production due to increasing demands for dietary protein by a growing human population [35] . The shift in agricultural practice and increased use of high-producing Holstein cattle in place of local breeds may have an impact on BTB epidemiology and contribute to a higher prevalence [36] . The official national control program in Morocco is currently based on a test and slaughter scheme . However , large-scale application remains challenging because testing is not mandatory , and the proposed compensation , ranging from 470 euros for local breeds to 980 euros for improved breeds , is considered lower than market value . In Morocco , BTB is an endemic zoonosis in livestock . Even though the predominant livestock species in Morocco are sheep and goats , cattle remain of major importance . The most recent national survey , conducted in 2004 , showed an individual cattle prevalence of 18% and a herd prevalence of 33% [37] . This prevalence remained similar in the individual level ( 20% ) , while the herd prevalence increased ( 58% ) in a 2012 pilot study of 1 , 200 cattle using the tuberculin skin test [38] . Since 2000 , the health risk of tuberculosis in Morocco has been addressed through a national TB program funded by the Ministry of Health in collaboration with the World Health Organization ( WHO ) . In 2014 , TB caused 2’800 deaths in Morocco [39] , and human tuberculosis had a relatively high incidence , with 36’000 new cases ( 106 cases per 100 , 000 inhabitants ) [40] . These data do not appear to differentiate between M . tuberculosis and M . bovis infection . A recent meta-analysis estimated the median proportion of human BTB among all TB cases in 13 African countries at 2 . 8% , with a range of 0–37 . 7% [41] . National prevalence data from a range of countries worldwide were summarized in a 2014 review; in Mexico up to 13% of all TB cases are reportedly due to BTB , while in the United States it is only 1 . 4% [42] . In Morocco , Bendadda et al . reported M . bovis prevalence of 17 . 8% among drug resistant TB isolates from 200 human sputum samples [43] . In the early 20th century , the prevalence in German cattle herds was 90% [44] , with 25–80% in other European states and only 2–10% in the US [45] . In most industrialized countries , the health risk and economic loss from M . bovis were considerably reduced or eliminated through strict test-and-slaughter and meat inspection protocols for cattle , along with the implementation of milk pasteurization and financial compensation of farmers [45] . Using a similar control strategy , Switzerland eradicated BTB in 1960 [46] . In most developing countries where the disease is endemic , such measures are not feasible due to financial constraints , particularly for farmer compensation , and inadequate veterinary services [47] . Alternatives for BTB endemic countries to reduce the health and economic risk related to the disease must be sought . Although cost estimation is difficult due to immense local variability in production parameters and prices [48] , the global economic loss due to BTB is thought to be about 3 billion USD annually [49] . In cattle , the disease has a significant economic impact , through an increased death rate and decreased milk and meat production , draft power and fertility [50] . Modelling approaches have been used , mainly for developing countries , to estimate different parameters and factors related to BTB . Previous publications on the economics of BTB focus mainly on the cost of disease and control efforts . Analyses of the profitability of control efforts are very scarce [51] . A study on the economics of BTB in Africa showed higher losses in intensive dairy systems in peri-urban areas of Ethiopia when compared to extensive pastoral production systems in rural areas but did not include the cost to public health [48] . Zinsstag et al presented a simplified framework for a model of animal to human transmission in which transmission between cattle and from cattle to humans is considered [51] . This model allows the simulation of different scenarios over 5–10 years , with and without intervention , where the measurable outcome is prevalence in humans and in cattle . The economic analysis further delineates broader issues such as inter-sectorial contributions from agricultural and public health or private households affected by BTB . Interventions to reduce health and economic risks , such as those related to BTB , are non-linear processes . Although , statistical analyses of data have been used for many years to analyze different types of health interventions , mathematical modeling represents an alternative approach which provides a broader understanding , especially regarding disease transmission to humans . Several mathematical BTB models have been developed to study the transmission dynamics and to assess the effectiveness of control measures mostly for wildlife ( badgers and possums ) but also for cattle [52] . In Italy , a compartmental stochastic model of within and between-farm BTB dynamics in cattle was developed using a monte-Carlo simulation of BTB epidemics following the random introduction of infected individuals in the network . Consequently , slaughterhouse inspection has been found to be the most effective surveillance component [53] . The same methodology was used in Great Britain in order to study BTB transmission dynamics and to assess the currently used control measures [54] . A model composed of two sub-models for buffalo and cattle populations in South Africa showed that BTB infection is only sustained in cattle and buffaloes when all transmission routes are involved [55] . In France , a compartmental stochastic model was used to assess within-herd spread of BTB operating in discrete time in order to design , calibrate and validate a model of spread of M . bovis within a cattle herd . Various herd management practices as well as control programs were parameterized into the model . Therefore , the median effective reproductive ratio was estimated to be 2 . 2 and 1 . 7 respectively in beef and dairy herds [56] . This paper presents an ordinary differential equation ( ODE ) mathematical model of BTB transmission from cattle to humans in order to estimate the disease cost and simulate potential interventions in Moroccan cattle .
Annual data on cattle numbers are estimated using data routinely collected by the Ministry of Agriculture and reported to veterinary services . Our model considered these data from 1995 to 2013 . In the transmission model , the cattle population was not stratified by age and sex . We used a tuberculin prevalence of 18% for cattle , as reported in the most recent national survey ( 2003 ) [33] . A similar prevalence was noted in a smaller study performed in 2012 in Sidi Kacem , Morocco [38] . A schematic diagram of the model is depicted in Fig 1 and the variables and parameters are described in Tables 1 and 2 , respectively . The cattle population is divided into three mutually exclusive compartments consisting of susceptible cattle ( SC ) , exposed cattle with latent BTB which are positive to the tuberculin test , without showing gross visible lesions ( EC ) and infected cattle with active BTB showing tuberculosis lesions ( IC ) . Those parameters were estimated based on Ngandolo et al 2009 [57] . The total cattle population ( NC ) at time t is: NC ( t ) = SC ( t ) + EC ( t ) + IC ( t ) ( 1 ) The human population consists of four mutually exclusive compartments: susceptible humans ( SH ) , exposed humans with latent BTB reacting to the Mantoux test ( EH ) , infected humans with active BTB ( IH ) and humans recovered from BTB with temporary immunity ( RH ) . The total human population ( NH ) at time t is: NH ( t ) = SH ( t ) + EH ( t ) + IH ( t ) + RH ( t ) ( 2 ) The susceptible cattle population ( SC ( t ) ) increases through birth ( at a rate bC ) and decreases through exposure to BTB ( at a rate βC ) and mortality ( at a rate μC ) . Exposure to BTB is assumed to be frequency dependent . According to Bernues [58] , the rate bC decreases by 5% for exposed cattle with latent BTB ( EC ( t ) ) and for infected cattle with active BTB ( IC ( t ) ) , such that: dSCdt=bCSC+ ( 0 . 95×bC ( EC+IC ) ) −βCICSCNC−μCSC− ( 1−sp ) pSc ( 3 ) The population of exposed cattle with latent BTB ( EC ( t ) ) is generated through infection of susceptible cattle with BTB ( at a rate βC ) and decreased through the development of active BTB ( at a rate αC ) and through mortality ( at a rate μC ) . Consequently: dECdt=βCICSCNC−αCEC−μCEC−se*pEc ( 4 ) Similarly , the infected cattle population with active BTB ( IC ( t ) ) is generated through the development of active BTB among exposed cattle with latent BTB ( at a rate αC ) and decreased through mortality ( at a rate μC ) : dICdt=αCEC−μCIC−se*pIc ( 5 ) For simplicity , no additional mortality rate due to BTB is assumed for the cattle and human populations in the model . The susceptible human population ( SH ( t ) ) increases through birth ( at a rate bH ) and through recovered humans becoming susceptible again ( at a rate δC ) . The population decreases through exposure to BTB from cattle with active BTB ( at a rate βH ) and through natural mortality ( at a rate μH ) . For the susceptible human population , the exposure ( both direct ( aerosol ) and indirect ( milk ) transmission ) to BTB from cattle with active BTB is assumed to be frequency dependent and proportional to the number of infected cattle ( IC ) : dSHdt=bHSH−βHICSHNC−μHSH ( 6 ) Human to human transmission is assumed to be negligible . The population of exposed humans with latent BTB ( EH ( t ) ) is generated through infection of susceptible humans with BTB ( at a rate βH ) and decreased through the development of active BTB ( at a rate αH ) and through natural mortality ( at a rate μH ) : dEHdt=βCICSHNC−αHEH−μHEH ( 7 ) The infected human population with active BTB ( IH ( t ) ) is generated by the development of active BTB among exposed humans with latent BTB ( at a rate αH ) and decreased by recovery of humans with active BTB due to treatment ( at a rate γH ) and by natural mortality ( at a rate μH ) : dIHdt=αHEH−γHIH−μHIH ( 8 ) The population of humans recovered from BTB and temporarily immune due to treatment is generated through the recovery of humans with active BTB ( at a rate γH ) and decreased through humans becoming susceptible to BTB again after the end of the prophylactic period of the drugs ( at a rate δH ) and through natural mortality ( at a rate μH ) , so that: dRHdt=γHIH−δHRH−μHRH ( 9 ) A sensitivity analysis of the model recalculated the change of prevalence if individual parameters varied from baseline over 30 years . Although Morocco has a test and slaughter policy for the control of BTB , it is currently not effectively implemented . The BTB transmission model was used to estimate the effect of the proportion of tested and slaughtered tuberculin positive animals on the duration to reach freedom from disease , achieving an individual animal prevalence of less than one in a thousand tested animals ( <1/1000 ) according to the standards of the World Organization for Animal Health ( OIE ) [65] . The proportion of tested and slaughtered animals was simulated by removing 10–100% of the exposed ( Ec ) and infectious cattle ( Ic ) per year from the herd . The control reproductive number Rc including the test and slaughter intervention as proportion p with a test of sensitivity se was computed as: Rc=α*β ( α+μ+se*p ) ( μ+se*p ) ( 18 ) The associated costs were estimated in a summaric way , assuming an incremental cost of comparative intradermal or interferon gamma ( BOVIGAM ) testing of 3 euros per animal . The cost of compensation at 80% of the market value varies from 470 euros for local breeds to 970 euros for improved breeds [66] . For the current estimation of the cost of BTB elimination in Morocco , we used a single value of 500 euros of compensation per slaughtered animal . Models run with and without interventions were simulated using data from 2013 onwards . The OIE recommended cut-off for SICCT interpretation is 4 mm , however , many studies showed that a severe cut-off of 2mm increased the sensitivity of the test [67 , 68] , without affecting the specificity compared to the recommended cut-off [69] . Consequently , we decided to consider both 2mm and 4mm cut-offs in the model , and to compare the respective results . The present model considered both options of SICTT cut-off at 2mm and 4mm .
The reproductive ratio of the cattle to cattle transmission of bovine tuberculosis without intervention was 1 . 325 . For the total cost , the birth rate of cattle bc was the most sensitive parameter influencing the dynamics of BTB transmission ( Fig 2 ) . High birth rate values lead to an increased cattle population yielding higher costs for elimination . For the time to elimination , the sensitivity of the test was the most sensitive parameter . Low test sensitivity ( i . e . with cut-off at 4mm ) leads to low detection of infected animals and therefore less culling of infectious cattle , which leads to a longer time to elimination . The simulation of a test and slaughter intervention led to a decline in BTB prevalence depending on the proportion p of testing ( Fig 3 ) . The time to elimination , i . e . the time to reach an individual animal prevalence of less than one in a thousand , ranged from 75 years for p = 20% to 12 years for p = 100% . For values of p > 60% , the time to elimination was below 20 years ( Fig 3 ) . The reproductive number decreased rapidly below one with an increasing proportion test and slaughter p ( Fig 4 ) . With 60% testing and culling , the prevalence of exposed and active human BTB cases decreased from 3 . 5 per 1 , 000 . 000 to less than 1 in 1 , 000 , 000 at the time of freedom from disease after 20 years ( Fig 5 ) . The cost of test and slaughter depends on the percentage p of test and slaughter ( Table 5 ) . Lower p results in lower cumulated costs but longer time until elimination . The cumulated cost is remarkably stable for p values higher than 0 . 2 , ranging between 1 . 47 to 1 . 87 billion Euros within a time range of 12 to 75 years to reach freedom from disease . The cumulated cost of BTB test and slaughter intervention and the time to elimination were lower using 2mm cut-off of SICTT compared to the 4mm cut-off ( Fig 6 ) .
This manuscript presents the first cattle to cattle and cattle to human compartmental deterministic mathematical model . Differential equations were used to describe different pathways within and between compartments at human and animal level . Sensitivity analysis of the model has been used to determine the most sensitive parameter . Additionally , different scenarios of test and culling interventions were simulated by considering different proportions of tested population per year ( p ) .
BTB prevalence was found to reach less than one per thousand in less than 20 years when the proportion of tested cattle was above 60% . The annual cost for this potential intervention was nearly 77 million euros . Intervention in cattle was found to impact the prevalence of human TB due to M . bovis , which decreased from 5 per 10 , 000 to 1 per 10 , 000 after 17 years . The economic assessment presented here is preliminary , and a detailed cost and cost-effectiveness analysis will be published separately . However , our analysis informs Moroccan bovine tuberculosis control policy on the time horizon , range of cost and optimal levels of intervention . An effective control program will depend on the human resources and technical and logistical capacity of the veterinary services to implement testing and slaughtering of animals . If the proportion of cattle subjected to test and slaughter was greater than 60% , freedom from disease would be reached in less than 20 years . The simulation results suggest that switching from a 4mm cut off to a 2mm cut off would be likely to result in significantly shorter durations of elimination programs and much cheaper elimination campaigns . Our model simulates the removal of individual animals rather than whole herds . Past experience in Europe has shown that whole herd removal is critical for effective elimination in low prevalence situations [46] . In addition , a herd based model of the Moroccan cattle population could potentially lead to a lower intervention cost , as it is more realistic . A recalculation of the intervention cost taking into account stratification by breed , sex and age should be undertaken , as it could lead to a different cost estimation of BTB control strategy . Shortage of human resources should be considered for intervention planning , a maximum of 40% cull rate might be feasible; however , it would be costly in view of current Moroccan economic situation . One may think that test and slaughter implementation would lead to a reduction in cattle population and its by-product . But on the other hand , the increased import of cattle from other countries , with enhanced control measures , could maintain the current population density . In the meantime , as dairy products are provided mostly from highly controlled farms where BTB has a very low prevalence , we could argue that milk production would not be significantly affected . The WHO includes BTB amongst the seven neglected zoonoses which are perceived to be severe threats to public health [1] . Further molecular epidemiology investigations in Morocco are needed in order to clarify local and national human BTB/TB ratios . To reach this goal , closer collaboration , at the national and international level , between the human and animal health sectors through a One Health approach is highly recommended . Operations in these two sectors remain largely independent in Morocco . Communication must be enhanced to establish a One Health approach , which has proven efficacy in health service delivery and potential for economic savings in zoonosis control [74 , 75] . | Tuberculosis is a disease of humans and animals which mainly affects the lungs but can also manifest in other organs . A variety of tuberculosis bacteria cause the disease and are usually transmitted through air , i . e . inhalation of aerosols . Bovine tuberculosis ( BTB ) occurs predominantly among domestic cattle , although wild animals are an important reservoir for transmission . Humans are usually infected with BTB through contaminated dairy products or close contact with cattle . While BTB has been eliminated in cattle and human populations of most high-income countries , it is still a major health threat in low- and middle-income countries . In Morocco , the disease frequently occurs in cattle and poses a health risk for humans . An effective intervention to reduce BTB among domestic cattle and reduce the risk to humans is slaughter of cattle which test positive for the disease . We simulated BTB in the Moroccan cattle and human populations using a disease transmission model . We assessed effects of test and slaughter in regard to elimination of disease in cattle and humans and estimated the associated costs with the model . The time to elimination of disease depended on the number of cattle tested . Our model suggests that the disease might be eliminated in cattle within 32 years if 40% of Moroccan cattle are tested annually and infected individuals are slaughtered or within 13 years if at least 90% of the cattle population are tested . The estimated total costs for the time periods until elimination ranged from 1 . 55 billion euros for testing and slaughter at 50% to 1 . 48 billion euros at 90% . These results can be used as a guide for planning BTB control policy in Morocco with regard to time frames and associated costs . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion",
"and",
"conclusions",
"Model",
"properties"
] | [
"death",
"rates",
"livestock",
"morocco",
"medicine",
"and",
"health",
"sciences",
"ruminants",
"demography",
"geographical",
"locations",
"tropical",
"diseases",
"vertebrates",
"animals",
"mammals",
"animal",
"slaughter",
"bacterial",
"diseases",
"animal",
"management",
"infectious",
"disease",
"control",
"africa",
"veterinary",
"science",
"infectious",
"diseases",
"veterinary",
"diseases",
"tuberculosis",
"agriculture",
"people",
"and",
"places",
"bovine",
"tuberculosis",
"in",
"humans",
"biology",
"and",
"life",
"sciences",
"cattle",
"amniotes",
"bovines",
"organisms"
] | 2017 | Transmission dynamics and elimination potential of zoonotic tuberculosis in morocco |
Erns is an essential virion glycoprotein with RNase activity that suppresses host cellular innate immune responses upon being partially secreted from the infected cells . Its unusual C-terminus plays multiple roles , as the amphiphilic helix acts as a membrane anchor , as a signal peptidase cleavage site , and as a retention/secretion signal . We analyzed the structure and membrane binding properties of this sequence to gain a better understanding of the underlying mechanisms . CD spectroscopy in different setups , as well as Monte Carlo and molecular dynamics simulations confirmed the helical folding and showed that the helix is accommodated in the amphiphilic region of the lipid bilayer with a slight tilt rather than lying parallel to the surface . This model was confirmed by NMR analyses that also identified a central stretch of 15 residues within the helix that is fully shielded from the aqueous layer , which is C-terminally followed by a putative hairpin structure . These findings explain the strong membrane binding of the protein and provide clues to establishing the Erns membrane contact , processing and secretion .
The genus Pestivirus belongs to the family Flaviviridae , together with the genera Hepacivirus , Flavivirus and Pegivirus . It represents a group of economically important animal viruses like classical swine fever virus ( CSFV ) and bovine viral diarrhea virus ( BVDV ) [1] . The economical damage caused by these viruses is not only due to the acute infection of the animals , but stems also from their ability to cause persistent infection of the fetus after infection of a pregnant animal [2] , [3] . If an infection of the fetus by BVDV is established , a persistently infected calf is born . This calf often shows no signs of disease , but sheds huge amounts of infectious virus particles throughout its whole life , which leads to an efficient spreading of the virus . The genome organisation and basic molecular features of pestiviruses are more similar to the hepacivirus HCV ( human hepatitis C virus ) than to the other members of the Flaviviridae . The positive sense single stranded RNA genome consists of about 12 , 300 nucleotides and contains one single open reading frame that codes for a single polyprotein precursor of ∼4000 amino acids [1] . This precursor is co- and posttranslationally processed by cellular and viral proteases to release the viral proteins . Compared to the HCV RNA , the pestivirus genome codes for two additional proteins: the non-structural protein Npro , and the structural protein Erns [1] . Both proteins interfere with the immune response of the infected animal and are important for establishing a persistent infection [4] . The two proteins exert two different functions . Npro is an autoprotease and leads to the degradation of IRF3 ( interferon regulatory factor 3 ) in the infected cell via the proteasome [5] , [6] , [7] , [8] , [9] , [10] , and it also interferes with IRF7 dependent pathways [11] . This results in the deactivation of the innate immune response of the infected cell . In contrast , Erns is a viral glycoprotein that forms disulfide-linked homodimers and can be found together with the glycoproteins E1 and E2 on the surface of the enveloped virus particle [3] , [12] , [13] . Erns consists of 227 amino acids , has a molecular weight of 42–48 kDa , being heavily glycosylated except for its C-terminal region [3] , [14] , [15] . It is not only involved in the formation of infectious virus particles , but it is also secreted from the infected cells , and up to 50 ng/ml of the protein can be detected in the blood of infected animals [16] . In addition , Erns has an intrinsic RNase activity , which is very unusual for an RNA virus protein [17] , [18] , [19] . The structure of the RNase domain of Erns was recently determined , with the finding that it has a T2-RNase like fold [20] that confirmed data from sequence analysis studies [18] . T2-RNase represents a very old and unique RNase family whose members are broadly distributed in nature , but the functions of these RNases are unknown . The enzymatic activity of Erns is necessary for its activity as a virulence factor . It could be shown that the deactivation of the RNase activity by deletion of one amino acid in the active site of the protein caused attenuation of the virus in its natural host [21] , [22] . Interestingly , attenuation was also observed in an RNase positive virus , in which Erns dimerization was prevented by mutation of the Cys residue that forms the intermolecular disulfide bond in the wt protein [23] . Moreover , an abrogation of the Erns RNase activity together with a deletion of the Npro coding sequence prevented the establishment of a persistent BVDV infection [4] . We and others postulated that the secreted version of Erns - and not the protein bound to the virus particle - represents the key player that interferes with the immune system . The secretion of Erns is controlled by the C-terminal end of the protein . In previous studies we have determined several parameters in the C-terminal region of Erns that are necessary for the retention/secretion of the protein [24] , [25] , [26] . The free C-terminus of Erns is formed upon cleavage of the Erns/E1 precursor protein by the cellular signal peptidase [27] . This cleavage site is a very unusual substrate for the signal peptidase , because the characteristic signal activating this peptidase is normally composed of a transmembrane helix followed by a so-called von Heijne sequence [28] , [29] . The cleavage site between Erns and E1 contains such a von Heijne sequence , but the Erns C-terminus lacks a transmembrane helix and contains an amphipathic helix instead . Nevertheless , the C-terminus of Erns obviously has to fold in a certain conformation that is accepted as a substrate by the signal peptidase . The C-terminus of Erns governs not only protein cleavage and secretion , but it is also important for the membrane binding of the protein . Sequence analysis predicts a helical fold for the C-terminus , which would bestow it with a marked amphipathic character [24] , [25] . Fig . 1 shows the 2D flat projection of the 3D structure of the Erns anchor ( Lys167 – Ala227 ) assuming a continuous α-helical conformation . The hydrophilic and hydrophobic faces are maintained throughout the entire length of the helix , which implies that it could bind flat onto the membrane surface . This amphipathic structure would explain the membrane anchoring , but not the action of the signal peptidase , nor the regulation of secretion . As an alternative arrangement , it has been recently suggested that the Erns membrane anchor might fold as a helical hairpin by forming a long ladder of salt bridges between its N-terminal and C-terminal helical segments , a so-called electrostatic ‘charge zipper’ [30] . The resulting amphiphilic hairpin would in principle have an appropriate length to span the lipid bilayer in a transmembrane alignment and could thus serve as a substrate for the signal peptidase . To better understand the role of Erns and its mechanism of membrane anchoring and secretion , we have recently obtained some initial data on short peptide fragments from this region of the protein when bound to lipid bilayers [26] , but the non-continuous nature of these sequences precluded any firm interpretation . Here , we show that the complete C-terminal anchor of Erns indeed adopts a continuous α-helical fold when bound to the membrane . In contrast to the simplistic pictures of a surface-bound helix or a transmembrane hairpin , however , our results show that the Erns C-terminus is slightly tilted with regard to the membrane surface , and a substantial stretch of residues is shielded from the aqueous phase by the hydrophobic environment , while the rest is located at the water/membrane interphase . This refined model represents the first example of a viral structural protein that is bound to a membrane via an amphipathic helix .
In a previous analysis we had examined three overlapping peptide fragments corresponding to the Erns anchor sequences of CSFV strain Alfort/Tübingen [26] and BVDV strain CP7 ( data not shown ) . Their CD analysis revealed a strong tendency for the middle and the C-terminal part of the anchor sequence , represented by the two corresponding peptides , to fold as a helix . To verify these results for the entire Erns anchor ( Lys167 – Ala227 ) , we determined the secondary structure of this 61-residue domain in different environments by CD , as shown in Fig . 2 . We first used a solution of 50% TFE in phosphate buffer ( PB ) pH 6 . 5 ( Fig . 2 A , dashed line ) , which promotes intramolecular hydrogen bonds . This CD spectrum of the Erns anchor showed an overall helical fold , and the secondary structure deconvolution revealed a helix content of 80% for these data ( Fig . 2 A , bar diagram ) . Thus , almost the complete Erns C-terminus is in principle able to adopt a helical conformation . To test whether this secondary structure is also present in pure phosphate buffer ( PB ) , we measured the Erns anchor in PB pH 6 . 5 ( Fig . 2 A , solid line ) and pH 3 ( Fig . 2 A , dotted line ) . However , the Erns anchor was only partially soluble at pH 6 . 5 , leading to protein aggregation and turbidity in the aqueous sample . Therefore , the CD lineshape shows spectral artefacts caused by absorption flattening and differential scattering at wavelengths <215 nm , and the secondary structure analysis yielded only a very low degree of helix content . At pH 3 , on the other hand , the protein is completely soluble in PB , and the secondary structure calculation revealed a helix content of about 50% . To analyse the secondary structure of the Erns anchor in a membrane-like environment , we first used detergent micelles in low salt buffer . As the Erns anchor is positively charged we compared micelles of zwitterionic DPC and negatively charged SDS . Both systems supported a strong helical folding , and the secondary structure deconvolution revealed a helical fraction of over 60% ( Fig . 2 A , bar diagram ) . We then used zwitterionic DMPC lipid vesicles , and a 1∶1 mixture of DMPC with anionic DMPG . Both systems induced a high degree of helical folding ( ∼70–80% ) ( Fig . 2 A , bar diagram ) , similar to 50% TFE . These results indicate that the Erns C-terminus has a helix content of up to 80% in a membrane ( -mimicking ) environment , and may therefore be present as a continuous amphipathic helix , as implicated in Fig . 1 . To examine the orientation of the helical sequence relative to the membrane surface we used oriented CD ( OCD ) . In macroscopically oriented membrane samples it is possible to estimate the tilt angle of a membrane bound helix from the intensity of the negative band at 208 nm in the OCD spectrum , which is polarized parallel to the helix axis [31] , [32] , [33] . If the helix adopts an alignment parallel to the membrane surface , as expected for Erns , the minimum at 208 nm has a stronger intensity than the minimum around 223 nm . In contrast , a transmembrane helix that lies perpendicularly to the membrane surface shows no negative band at 208 nm , or even some positive ellipticity ( Supporting Information Fig . S1 ) . The OCD spectra of the Erns anchor , recorded in oriented DMPC or DMPC/DMPG ( 1∶1 ) ( Fig . 2 B ) , show a pronounced minimum at 208 nm with nearly the same intensity as the one at 223 nm . A transmembrane alignment of the Erns anchor , as had been speculated for the Erns/E1 precursor [30] , can thus be excluded for the mature processed Erns under these conditions . However , because the intensity of the 208 nm minimum is less than the intensity of the 223 nm band , the helix does not seem to be aligned completely parallel to the membrane surface . Without taking any minor secondary structure elements into account ( e . g . disordered regions , which have a minimum at 198 nm ) , the Erns anchor appears to be slightly tilted within the membrane . Many amphipathic helices , especially antimicrobial peptides , are known to interact with each other in a concentration dependent manner , which often results in changes in their membrane alignment . At low concentrations these peptides exhibit a mostly parallel orientation with regard to the membrane surface , but at higher concentrations the self-assembly of these molecules leads to a re-alignment and the formation of a transmembrane pore [33] , [34] , [35] , [36] , [37] , [38] . In the case of Erns , the possibility of self-assembly via an intermolecular charge zipper had been suggested [30] , or helix-helix interactions via a GxxxG motif might be conceivable . To test for a putative concentration-dependent change in the secondary structure ( e . g . aggregation or oligomerization ) or in the orientation of the helical segment of Erns , we compared several samples with different lipid/protein ratios . The CD spectra of all four tested concentrations ( P/L ratios ) in the DMPC/DMPG 1∶1 lipid mixture were essentially identical ( Fig . 2C ) . Neither did the OCD spectra at the same ratios of 1∶20 , 1∶50 and 1∶100 , as displayed in Fig . 2D , show any change in the alignment of the helical segment . ( The OCD sample at 1∶200 did not yield a reliable spectrum due to technical problems resulting from the very low protein concentration . ) In summary , these CD and OCD data demonstrate that the Erns anchor does not show a concentration-dependent change in its global secondary structure nor in its orientation in the membrane . Most importantly , the data recorded here give no indication at all of a transmembrane alignment at any concentration . The observed peripheral location of the C-terminal part of the Erns membrane anchor is highly intriguing , because this architecture represents a so far unknown type of substrate for the cellular signal peptidase , which usually requires a transmembrane helix upstream of the cleavage site . To investigate the structure of the C-terminal fragment in more detail , we expressed a truncated construct named ErnsΔN , which represents the far C-terminal 34 amino acids of the anchor sequence ( Arg194 – Ala227 ) . To analyse the secondary structure of this C-terminal Erns anchor fragment ( Fig . 2 E ) , we used 50% TFE ( dashed line ) , DPC micelles ( straight line ) , and DMPC vesicles ( not shown ) . Unfortunately , the CD spectrum of ErnsΔN in DMPC vesicles showed scattering artefacts which lead to an error-prone CD spectrum and prevented an exact analysis . To allow comparison with the subsequent NMR analysis , we examined the C-terminal region also in small lipid bicelles with DHPC/DMPC ( 4∶1 ) ( dotted line ) . The corresponding CD spectra of ErnsΔN showed a high percentage of helical folding ( ∼70–90% ) in all three systems ( Fig . 2 E , bar diagram ) . This means that the Erns anchor can fold as a long amphipathic helix not only in detergent micelles and lipid vesicle suspensions , but also in bicelles . The OCD spectrum of the C-terminal fragment in Fig . 2F shows a less intensive band at 208 nm than the complete Erns anchor at a P/L ratio of 1∶50 . This means that either the N-terminal region of the Erns anchor is less tilted in the membrane than the C-terminal region , or the N-terminal elongation pulls the C-terminal region into a more parallel orientation . Importantly , the band at 208 nm still has a distinct negative signal amplitude , which excludes the possibility that this region of the Erns anchor could be inserted into the membrane in a transmembrane alignment . Further information on the orientation of the Erns anchor in the membrane , the positions of individual residues , and the water accessibility of individual NH groups was obtained from liquid-state NMR spectroscopy . Although the full-length Erns anchor yielded good quality 1H15N-HSQC spectra in DHPC/DMPC bicelles , as well as in DPC and SDS micelles , the corresponding 3D-15N-HSQC-NOESY and TOCSY spectra suffered from line broadening that impeded sequential assignment ( Supporting Information Figure S2 ) . The stretch Trp203-Gly212 could be tentatively assigned based on characteristic proton shifts , but the assignment could only be safely confirmed later by comparison with the spectra of ErnsΔN . Therefore , the detailed NMR analysis was done for the truncated ErnsΔN corresponding to the 34 C-terminal residues of Erns . Backbone assignment ( 1H , 15N , 13Cα , 13Cβ ) was achieved for all residues , except for the first two in DHPC/DMPC ( 4∶1 ) bicelles . We sought to determine whether parts of ErnsΔN are protected from the solvent , by titration with the paramagnetic agent Gd-DOTA and by dissolution of a lyophilized sample of ErnsΔN/DHPC/DMPC in D2O . Supporting Information Figure S3 shows that residue protection factors of ErnsΔN with Gd DOTA at 0 . 5 mM are uniform along the entire sequence . D2O exchange revealed no stably protected residues either ( data not shown ) , given that already a single exchange event leads to a disappearance of the signal . Similarly , titration with paramagnetic agents shows a protection only when the residues are deeply inserted in the hydrophobic interior of the bicelle [39] . We therefore determined also the water accessibility of the NH groups with the more sensitive CLEANEX-PM pulse sequence [40] , which measures proton exchange rates . The CLEANEX pulse sequence applies a water-selective excitation pulse prior to a chemical exchange sequence in which magnetization transfer from water protons to 15N-bound exchangeable protons occurs , followed by a 1H15N-HSQC-type experiment . The chemical exchange efficiency depends on the accessibility of the 15N-bound proton , and thus reports on its location in the bicelle or its stable participation in a hydrogen bond . Here , a single encounter with a water molecule may already lead to a measurable magnetization , meaning that for partially solvent-accessible amide protons this method is more sensitive than the other two , because the protein exchange time is limited to 100 ms . The assigned 1H15N-HSQC of ErnsΔN is shown in Fig . 3A . Judging from the nuclear Overhauser effect ( NOE ) and chemical shift anisotropy ( CSI ) patterns ( Fig . 4A , middle part ) , the helix extends from Leu215 or Glu216 , corresponding to 50% helical content . These findings support the CD data above , that showed a high degree of helical folding in the same membrane-mimetic environment ( data above ) . The remaining residues in the C-terminus show some HN-HN contacts as well as NOEs between Trp222 , Phe223 and Tyr226 ( Fig . 4B ) , which suggests a loop-like structure , but the quality of the spectra did not yield enough NOEs for a full 3D structure determination . Nonetheless , the HN-HN contacts together with the loop-like contacts in the C-terminus ( Fig . 4A , upper part ) and the water accessibility measurements argue in favour of a dynamic loop-like conformation of the far C-terminus rather than a firmly folded α-helical structure . A comparison of the 15N-HSQC and the CLEANEX spectra ( Fig . 3B ) revealed that only 13 amide protons are engaged in exchange ( i . e . they give cross-peaks in the assigned CLEANEX spectrum ) . We may thus conclude that the remaining 20 amide groups are protected from exchange and are therefore assume a stable secondary structure within a more hydrophobic environment characterized by a low dielectric constant . A comparison with the water-HN NOEs in a 1H15N-NOESY experiment confirmed our analysis . To obtain the water accessibility of each NH group , we calculated the normalized proton exchange rate by dividing the intensity of each CLEANEX peak by the intensity of the corresponding 1H-15N HSQC peak . Fig . 3C shows that a long stretch of 15 amino acids in the middle of ErnsΔN does not have any water contact at all and should therefore be protected within the bicelle . This stretch comprises not only hydrophobic amino acids , but also contains two Thr , one Arg , two Lys and one Gln . Most remarkably , the side chains of Gln207 and Trp203 , which are positioned on the edge of the hydrophobic face of the helix , are not in contact with water either , thus again supporting the model of an amphiphilic helix that is embedded in and protected by the membrane . In contrast , the N-terminal region of ErnsΔN up to Leu200 shows a high normalized proton exchange rate that is characteristic of a water-exposed surface-location . Surprisingly , most of the C-terminal residues , including the very last amino acid Ala227 , are again shielded from water . 15N-NMR relaxation analysis ( Fig . 4A , lower part ) suggests that the residues comprising the helical part are fairly rigid . The T1 , T2 and hetNOE values show a tendency to gradually change from residue 201 towards the N-terminus and from residue 215 or 216 towards the C-terminus , indicating increasing flexibility towards the ends . Besides the immediate N-terminus up to residue 196 ( in accordance with the lack of data for the first residues ) , the last two residues of the C-terminus are highly flexible , which is evident from their high T2 values and low to negative hetNOEs . Together with the fact that these two and the preceding residues at the utmost C-terminal end are protected from contact with water and show an interconnecting network of NOEs , a picture emerges where the C-terminus forms a stable but semi-flexible structure within the bicelle , whereas the N-terminus is unstructured and exposed on the surface of the bicelle . Earlier cell culture studies had reported that the C-terminus of Erns ( BVDV Strain CP7 ) has a major influence on the retention/secretion of the protein . The deletion of the last five C-terminal amino acids ( FGAYA ) led to a dramatic increase in the amount of secreted Erns [24] . A recently conducted cell culture experiment of the Erns protein from CSFV strain Alford/Tübingen showed that even the loss of the four C-terminal residues ( GAYA ) has a major impact on protein secretion . In these experiments the protein secretion rate increased from below 10% ( wt protein ) to 35% ( truncated version ) ( unpublished data ) . Since the major part of the membrane binding of the Erns anchor is supposedly contributed by the central region of the amphipathic helix , we wondered whether the C-terminus may have any influence on the conformation of the membrane anchor , thereby modulating its membrane association . To answer this question , we expressed and analysed yet another , C-terminally trunctated protein ErnsΔNΔC , a variant of ErnsΔN lacking the last 6 amino acids ( WFGAYA ) . The assignment of ErnsΔN could be easily transferred to the 1H-15N HSQC spectrum of this shortened version ( Fig . 5A ) . The corresponding CLEANEX spectrum ( Fig . 5B ) and the calculated normalized proton exchange rates ( Fig . 5C ) indicated that in ErnsΔNΔC only 8 amino acids located in the middle of the sequence were shielded from water . In addition , two further amino acids , Thr201 on the N-terminal side and Lys214 on the C-terminal side of the shielded sequence showed no water contact . Comparison of the normalized proton exchange rates of ErnsΔN and ErnsΔNΔC in Fig . 6A reveals several differences resulting from the deletion of the 6 C-terminal amino acids . For the N-terminal amino acids , ErnsΔN and ErnsΔNΔC yielded nearly the same values ( except for the peptide bond NH group of Gln195 ) , hence the conformation and water-exposed location of these residues should be equivalent in both proteins . However , residues Thr201 , Gly212 and Lys213 , which in the case of ErnsΔN are embedded in the bicelle , now seem to be located somewhat closer to the lipid/water interphase in ErnsΔNΔC , as seen by the small but significant change in the proton exchange rate . Especially the peptide bond of Trp203 shows a high proton exchange rate in ErnsΔNΔC and should therefore in the truncated fragment be located at the membrane surface where it is exposed to water . Interestingly , the NH side group of Trp203 did not show a change in its water accessibility upon truncation , but is still located in a hydrophobic environment . Most notably , the water accessibility of the C-terminal amino acids is increased in ErnsΔNΔC . The peptide NH group of the four C-terminal amino acids Glu216 , Asn217 , Lys218 and Ser219 all exhibit an enhanced proton exchange rate once the last six amino acids are deleted . Moreover , the Thr202 NH group in ErnsΔNΔC shows a proton exchange rate that is comparable to the values determined for the aforementioned four amino acids . Interestingly , the side chain NH of Asn217 showed nearly the same low proton exchange rate in both proteins . This indicates that only the peptide bond has changed its position in ErnsΔNΔC whereas the side chain is still protected from water . The results of the water accessibility experiments with ErnsΔN and ErnsΔNΔC are summarized in Fig . 6B . It is obvious that the deletion of the last six amino acids results in a change in water accessibility of upstream sequences and increases the number of NH bonds that are engaged in fast exchange . This occurs especially in the formerly water shielded central region of the Erns protein . Interestingly enough , the flanking amino acids Thr201 and Lys214 did not gain direct access to water , but the accessibility of the peptide NH groups increased at both sides of the formerly shielded domain upon C-terminal truncation . In both deletion constructs , there remains a continuous stretch of 8 amino acids showing no water accessibility , including the side chain of Gln207 . This finding of a discontinuity in the protection pattern at residues 216–219 does not support the picture of a perfectly straight amphipathic helix in the C-terminal region . To elucidate the degree of helicity of membrane-bound Erns by an independent method , we performed all-atom Monte Carlo ( MC ) simulations with an all-atom intramolecular force field [41] that was recently used to describe the reversible folding of various proteins [42] , [43] . To speed up the simulation , we employed an implicit membrane model with three layers providing discrete dielectric environments [44] . ErnsΔN ( Fig . 7B ) as well as the entire Erns anchor ( Fig . 7A ) were simulated at temperatures ranging from 220K to 400K , starting from completely helical conformations that were placed in the proximity of the membrane surface . The region from Thr172 to Lys220 remained helical , while the N- and C-terminal regions showed partial unfolding with increasing temperature ( see Supporting Information Fig . S4A ) . A similar picture emerged for ErnsΔN , as illustrated in Fig . 7B , which remained helical from Leu200 to Lys214 . The last six C-terminal residues are unfolded for both peptides ( Supporting Information Fig . S4A/C ) . Both peptides remained bound to the membrane surface in all simulations , with the hydrophobic residues facing inward . Moreover , the helix was slightly tilted ( Supporting Information Fig . S5 ) . The Monte Carlo simulations can give a good prediction of the local secondary structure of ErnsΔN , but the implicit membrane model is not suitable to gain an impression of the depth of insertion or of the exact tilt angle of the protein within the membrane . We therefore conducted all-atom molecular dynamics ( MD ) simulations in an explicit lipid bilayer composed of 512 DMPC molecules , in order to obtain further information on the membrane insertion of ErnsΔN ( Fig . 7C ) . The resulting MD model shows a pronounced helical conformation of ErnsΔN and a slightly tilted orientation in the membrane with an angle of 15° relative to the membrane surface . The peptide maintained a stable helical conformation throughout all simulations , but showed a kink around residue 202 when pulled into the membrane . Besides using the conventional MD approach of allowing the peptide to approach the membrane from the aqueous phase , the simulation in Fig . 7C was started from a position deep within the membrane . This should be a more relevant physiological starting position , since at least part of the structure should be located within the membrane when the C-terminus is generated by signal peptidase cleavage of the E/E1 precursor . The peptide maintained a stable helical conformation throughout the simulation , while the membrane initially bent to compensate for the peptide translocation when the protein was pulled into the membrane core . However , after 35 ns of free MD , the peptide was back to its original position near the membrane surface and assumed its stable orientation there .
Erns represents one of the four known pestiviral structure proteins . The protein is crucial for the formation of infectious viral particles and plays a major role also during the infection of cells . But aside from this elementary function , Erns represents also a virulence factor of pestiviruses . In this context it is important that Erns has an intrinsic RNase activity . RNases are only very rarely found in RNA viruses [45] , and to our knowledge the pestivirus Erns is the only viral structural protein displaying such enzymatic activity . This RNase activity is crucial for the virulence of the virus . It could be shown that pestiviruses containing an RNase negative Erns protein are viable and able to replicate to nearly wild type virus titers , but are attenuated in their natural hosts [21] , [22] . Most importantly , the RNase activity is a major factor for establishing a persistent infection upon transplacental infection of the fetus in a pregnant host animal [4] . Erns is not only concentrated within the cell at the site of virus budding , but it is also secreted to some extent from an infected cell and is thus found in the blood of infected animals [15] , [16] . According to the working hypotheses put forward so far , the virulence factor activity of the Erns RNase is supposedly linked directly to its secretion from infected cells . Accordingly , an analysis of Erns membrane anchoring and the mechanisms underlying its partial secretion are of major importance for understanding its activity . It had been proposed earlier that Erns should be bound to the host membranes and to the virion via interaction with the E2 protein that contains a typical transmembrane region [46] . Regardless of the question whether an Erns/E2 heterodimer is formed in all pestiviruses , the results presented here and data published before [24] , [25] , [26] clearly show that the Erns C-terminus represents an intrinsic membrane anchor per se . The amphiphilic region serves to attach the protein to lipid bilayers and enables its intracellular retention in the absence of any other viral protein . Accordingly , the hypothesis of indirect membrane anchoring of Erns achieved via interaction with E2 - as put forward by Lazar et al . [46] - is refuted by the data available now . The structure of the Erns N-terminal RNase domain was successfully determined by X-ray crystallography , yielding important clues to its enzymatic functions [20] . Unfortunately , the full-length Erns protein containing its C-terminal membrane anchor region could not be crystallized , so that structural data on this functionally important domain are missing . We have therefore investigated the structure of the C-terminal domain using several complementary methods . First , we could demonstrate in cell culture experiments that this domain does indeed serve as the membrane anchor [24] , [25] . However , in those experiments the membrane anchor did not bind to membranes as tightly as a typical transmembrane helix . Sequence prediction and model building showed that the Erns membrane anchor could be structured as a long amphipathic helix . Mutations interfering with the proposed amphipathic character of the C-terminal region affected the membrane anchoring as well as the secretion of the protein [25] , [26] . Peripheral membrane binding via amphipathic helices is quite common for both cellular and viral proteins . Cellular amphipathic helices are typically located at the N-terminus and consist of only 3–4 helical turns . Such proteins are found in the cytoplasm or the ER and have a broad range of important biological functions . For example , such structures are used in the cell for measuring membrane curvature [47] , for generating membrane vesicles [48] , [49] , or for establishing protein-protein interactions [50] . The amphipathic helices are generally oriented parallel to the membrane surface and act , in the case of proteins involved in the budding of membrane vesicles , as a membrane anchor of these proteins . The local curvature of the membrane involved in vesicle budding is usually not induced by the amphipathic helix itself , but e . g . by the bent form of the proteins [49] or by protein-protein crowding [51] . The considerable length of the Erns amphipathic helix ( up to about 50 amino acids in TFE ) is more reminiscent of amphipathic helices found in another group of proteins , such as cytolytic peptides . These peptides usually consist of an amphipathic helix composed of 23–31 amino acids [52] . Melittin , a toxic component of bee venom , for example , binds to cellular membranes and forms a membrane pore by oligomerization , which leads to the damage of the target cell . Similarly , the antimicrobial peptide PGLa is a straight amphiphilic α-helix , which binds flat to the membrane surface , whereupon it can start to tilt and eventually assemble transiently as an oligomeric transmembrane pore [34] , [36] , [38] . The melittin and PGLa helices can be slightly tilted into the membrane like the Erns anchor [35] , [37] , [53] , but the Erns anchor most likely lacks the ability of a concentration depending structural re-orientation according to our experiments . Several viral proteins are known to contain short amphipathic helices , such as the Brome-Mosaic-Virus protein 1a [54] , or the NS3 , NS4A , NS-4B and NS-5A proteins of hepatitis C virus [55] , [56] . Also for pestiviruses membrane binding of NS5A was demonstrated to occur via an amphipathic helix similar to the closely related hepatitis C virus [57] . But all these systems represent non-structural proteins , and the amphipathic helix is only used for membrane attachment and oligomerization to build up the replication complexes on the cytoplasmic membrane surface [58] . To our knowledge Erns is the only structural viral protein that is anchored via an amphipathic helix . This amphipathic helix is unusually long and located in the C-terminal region . It combines several functions , as it is not only important for the membrane binding of the protein , but also for the control of its secretion/retention , its intracellular location [26] , and for cleavage of the glycoprotein Erns/E1precursor by the cellular signal peptidase [27] . All these different demands have to be supported by the specific conformation and particular sequence of the Erns C-terminal region . Our analyses revealed that the Erns anchor is predominantly helical when bound to the membrane , but that the stability of the helix varies considerably over the sequence , with 15 residues representing a core helix that can be extended towards both sides . Monte Carlo ( MC ) simulations support the experimental data , as they yielded a similar extent of helicity and could identify the stretch Leu200 – Lys214 as the most stable helical region , in perfect agreement with CD and NMR data . Both the OCD analyses and the simulations revealed a slight tilt of the helix with respect to the membrane surface ( Fig . 2F and Supporting Information Fig . S5 , respectively ) . Since the MC analyses do not provide information on the location of the anchor in the membrane , molecular dynamics ( MD ) simulations were conducted . The MD analyses also revealed a helical conformation with a slight tilt . The helix was , embedded just below the lipid headgroup region of the membrane . Unlike the other data , the helix in the MD simulations extended virtually over the complete length of the peptide . Thus , the helix content of this simulated structure is significantly higher than what was observed experimentally in the CD , NMR and MC analyses , which had shown unwinding of the N- and C-terminal residues . This discrepancy is most likely due to an overestimation of the H-bond energy in the force field . As a general consensus of the CD , OCD , NMR , and MC simulation data , and in part also MD simulation results , at least part of the helix should be inserted more deeply in the membrane than the surrounding amino acids . This conclusion is in agreement with the data obtained by the NMR CLEANEX experiments showing that the central region of the amphipathic helix does not exchange the NH protons with water protons . Likewise , the side chain protons of Trp203 and Gln207 are protected and thus seem to be located in the hydrophobic interior of the bicelle . Nonetheless , the immersion within the bicelle is either not very deep or not perfectly stable , because all residues experience the influence of the paramagnetic agent gadolinium ( Supporting Information Fig . S3 ) . This ion induces strong relaxation that leads to a disappearance of the signals within a minimum radius of 5 Å [59] . The fact that the NMR signals of the full-length Erns anchor show dynamic exchange most probably with the free unstructured protein also indicates that the interaction with the membrane-mimicking bicelles is only of moderate stability . Interestingly , the N-terminal end of ErnsΔN is located on the bicelle surface or protrudes into the solvent , given its high water accessibility , while the amino acids of the far C-terminal end are shielded from water including the last C-terminal residue Ala227 . Together with the observed NOE pattern , this result suggests that the C-terminal end of the sequence is at least temporarily located within the membrane , being quite flexible with regard to its secondary structure , as most clearly seen for Ala227 , the very last C-terminal residue of the protein . Ala227 does not show any water contact , which argues for a membrane immersed location , but the relaxation and Het-NOE measurements revealed a flexible conformation . Any particular stable C-terminal conformation , such as a putative turn close to the water/lipid interphase was not reproduced in the MD simulations , possibly due to an overestimation of the helical hydrogen-bonding . However , the MC simulation revealed a significant probability for a turn-like structure at the C-terminus ( Supporting Information Fig . S4D ) , and the same pattern was also revealed by MC simulations of the full-length Erns anchor ( Supporting Information Fig . S4B ) . This model is supported by the pronounced influence that the native Erns C-terminus has on the central helix , as seen from the increased water accessibility in the truncated construct ErnsΔNΔC . This effect was not restricted to the C-terminal region of the immersed part , but was also seen for residues located upstream at the N-terminal end of the inserted helix . Thus , the C-terminal end of the Erns anchor has an effect on the immersion of the whole inserted segment . On the other hand , the N-terminal region of the protein did not show such an interesting effect , as all amide groups had strong water contacts . The structural model of the Erns C-terminus presented here does not answer all open questions about the biological mode of action of the Erns anchor , because it does not allow any conclusions about the mechanism governing the equilibrium of Erns retention and secretion . Nevertheless , it gives a first hint on how the interaction of the protein with the membrane occurs . Although the Erns protein lacks a transmembrane domain or a GPI anchor , which are usually responsible for tight membrane association , its unusually long amphipathic helix confers strong membrane binding . This anchor is not only attached to the membrane via the hydrophobic face of the amphiphilic helix , but because of the tilt part of the helix including its far C-terminal end is inserted into the membrane . This may facilitate the arrangement of lipids around the inserted peptide and may lower the free energy of the system to stabilize the membrane/protein interaction . It thus appears feasible that the Erns anchor could bind more strongly to membranes than a typical amphipathic helix on the bilayer surface , thereby providing the firm membrane association that is crucial for a viral surface protein . Nevertheless , the membrane affinity of this anchor is considerably lower than that conferred by a transmembrane helix [24] , which seems to be an important prerequisite for the observed secretion of Erns . Erns represents the first membrane protein for which anchoring via an amphipathic helix is described , and this unusual type of membrane attachment can be hypothesized to be of functional importance . The structure adopted by this anchor in a membrane could contribute to the known equilibrium between retention and secretion by adjusting the binding force at an appropriate level . This hypothesis is in agreement with the observation that C-terminally truncated Erns is more efficiently secreted [24] , [26] , because the observed increased water accessibility of the central part of this truncated helix suggests that this region becomes less deeply immersed and consequently has a decreased binding affinity . Whether this proposed decrease of binding ( alone ) is responsible for increased secretion of the mutant protein cannot be decided yet . Alternatively , a different orientation of the protein in the membrane could interfere with the interaction between Erns and other proteins important for its membrane association . Further investigation is necessary to answer the question whether other mutations enhancing secretion [25] , [26] also alter the immersion and binding affinity of the helix . In a rather speculative working hypothesis , one might propose that either a slight truncation of the C-terminus or its unfolding as a result of ( changing the ) interaction with a partner molecule could destabilize lipid binding of the anchor and lead to secretion . So far , nobody has achieved a detailed analysis of the primary structure of the secreted protein , and any data concerning putative interaction partners of the anchor sequence are missing , so that both possibilities have to be investigated in future work . For the time being , the present data support a model that provides strong membrane binding , which , however , can be modulated by changes affecting the far C-terminal end or the overall structure of the anchor . The structure model presented here displays the monomeric form of the protein but in vivo , a considerable amount of Erns is found as a homodimer covalently linked via a disulfide bond between the Cys residues at position 171 of the protein [23] , [60] , [61] , [62] . Nevertheless , we had no indication of dimer formation when analyzing the full-length anchor containing Cys171 . The mass spectra of the purified protein did not reveal the presence of the dimeric form and the NMR analyses proved furthermore that every nucleus was found in only one defined surrounding . This observation implies that only one conformation - whether dimer or monomer - was present in the samples . For a dimer , this result could only be obtained in the case of ideal symmetry . Moreover , we conducted NMR analyses on the Erns anchor both with and without the addition of DTT and were not able to identify any differences between these spectra ( not shown ) . Most importantly , the NMR analyses leading to the structural model were conducted with a C-terminal part of the anchor lacking Cys171 . Fraom biochemical analyses of the Erns protein it is known that the ability to form dimers is massively weakened when the disulfide linkage is prevented by mutation of Cys171 [23] . Taken together , these results strongly support the notion that the Erns anchor or fragments thereof analyzed in our experiments were monomers . Erns dimerization is important in vivo , however , and the dimeric form is found in virions , infected cells , and also in the supernatant of infected cells . Knocking out Erns dimer formation by mutation of Cys171 doesn't interfere with virus viability but leads to virus attenuation in the natural host [23] , so dimer formation plays a role for Erns function . Even though we do not have any data on the structure of a membrane anchored Erns dimer , it is tempting to speculate on the structure of such molecule . We can conclude from the data on a peptide corresponding to the N-terminal part of the CSFV Alfort/Tübingen Erns membrane anchor [26] , as well as from the crystal structure of residues 1 to 165 of the BVDV NCP7 Erns [20] that Cys171 is located in a rather flexible region of the protein , linking the enzymatically active N-terminal domain to the C-terminal membrane anchor . Hence , contact of the two monomers in this region should not be sterically hindered . In plane binding of the membrane anchor helices of both monomers parallel to the bilayer surface according to our structural model would also not interfere with protein/protein interaction between the regions containing Cys171 . The angle between the two anchor helices in a two dimensional projection should be flexible , since there is no reason to postulate any particular fixed arrangement . Furthermore , as pointed out above , we have no indication that two or more anchor molecules could engage in a parallel or antiparallel alignment of the two amphipathic helices of an Erns dimer . Whether the two regions around Cys171 of the two monomers form a parallel stem–like structure or whether they cross each other cannot be answered yet and also the membrane topology of this contact region is unclear at the moment . Further experimental work would be necessary to get an idea of the structure of the Erns membrane anchor in a dimeric state . One important step towards elucidation of this point would be to identify residues that are part of the dimerization interface in the region around Cys171 . The model established here for the Erns C-terminus describes only the folding of the fully processed protein . However , Erns is not translated as a single protein but rather as a part of the pestivirus polyprotein that is posttranslationally cleaved into the different viral proteins . The N-terminus of Erns is generated upon cleavage by the cellular signal peptidase [27] , which is also responsible for several other steps of pestivirus polyprotein processing [15] . The signal peptidase usually cleaves after a so called von Heijne sequence that is preceded by a transmembrane helix [28] , [29] . It has long been puzzling why and how the C-terminal end of Erns can be generated by signal peptidase , given that this sequence lacks a transmembrane element [27] . The structural model for the Erns C-terminus presented here cannot explain this conundrum , as we found no indication for a transmembrane orientation of the C-terminal segment . Therefore , the structure of the uncleaved Erns C-terminus should differ from the final form of the processed C-terminus to allow the cleavage by the cellular signal peptidase . It seems likely that the Erns C-terminus adopts an alternative transient structure during processing , while being tethered to the membrane via the transmembrane region of the E1 protein that follows downstream in the polyprotein . This sequence of events could explain why the cleavage between Erns and E1 is delayed compared to other polyprotein processing steps [15] . Recently , a new structural motif named “charge zipper” was identified in the membrane binding domains of various proteins [30] . The contact between two amphipathic helices is stabilized through electrostatic interactions between charged amino acids on the two apposed segments , such that a long ladder of salt bridges is formed between them , either leading to an intramolecular hairpin , or resulting in intermolecular oligomerization . Such a charge zipper motif was also identified in the C-terminal membrane anchor of Erns . Folding of this region into a helical hairpin would allow a transmembrane insertion of the anchor , since the charged amino acids would be shielded within the structure , while hydrophobic residues would be present on the outer face and could be easily inserted into the membrane . The resulting structure could provide a ( transient ) transmembrane helical hairpin structure upstream of the Erns/E1 processing site , which might explain its cleavage by the signal peptidase . After cleavage , a structural rearrangement of the released Erns C-terminus could occur , resulting in the identified structure of the mature Erns C-terminus that allows membrane anchoring of Erns and control of secretion . This charge zipper hypothesis is highly interesting but has to be analyzed in detail with further experiments . The Erns membrane anchor thus combines several very different functions in a rather short stretch of the sequence . These functions are important for the life cycle of pestiviruses . They rely on a specific structure that is probably established through different transition states , and which promotes membrane binding and signal peptidase processing . The final structure of the mature Erns C-terminus presented here as a model describes for the first time a new way of anchoring a viral surface protein via an unusually long amphipathic helix to a membrane . This arrangement represents a new perception of protein/membrane interaction that might also be relevant for other peripheral membrane proteins . Importantly , the structural model of the Erns C-terminus reveals not only a new possible fold of an amphipathic membrane anchor , but it also provides new insight into the biochemical requirements of the Erns protein and paves the way towards a better understanding of the virulence function of this fascinating system .
Plasmid pd29G was used as a starting point for all expression constructs . It consists of the plasmid pETZ2-1 ( kindly provided by Gunter Stier [63] ) , which codes for a Z2 domain to enhance the water solubility of the fusion protein , a C-terminal His-TAG for protein purification and a N-terminal TEV protease cleavage site to allow removal of the complete TAG from the desired expression product . The insertion in pd29G consists of a cDNA coding for part of the viral polyprotein from BVDV strain CP7 Erns ( Lys167 - Ala227 ) after four cloning derived amino acids ( GAMA ) . The cDNA sequence was optimized for the bacterial expression of the protein . Plasmid pd29G-1 , containing the sequence information for the protein ErnsΔN ( Arg194 - Ala227 ) , was generated by a QuikChange ( QC ) PCR using ‘CCTGTATTTTCAGGGCCGTCAGGGGACCGC’ as forward and ‘GCGGTCCCCTGACGGCCCTGAAAATACAGG’ as reverse primer . QC PCR was conducted according to the protocol by Strategene ( Heidelberg , Germany ) . The expression plasmid of ErnsΔNΔC ( Arg194 – Thr221; plasmid pd29G-1-2 ) was established from pd29G-1 via QuikChange mutagenesis with primers ‘GAAAACAAAAGCAAAACCTAATAACTCGAGCACCACCACC’ and ‘GGTGGTGGTGCTCGAGTTATTAGGTTTTGCTTTTGTTTTC’ . The established constructs were all verified by nucleotide sequencing with the BigDye terminator cycle sequencing Kit ( PE Applied Biosystems , Weiterstadt , Germany ) . Sequence analysis and alignments were done with MultiAlin [64] . Cloning was done using standard procedures [65] . The expression of proteins was done with E . coli strain BL21 ( DE3 ) in standard LB-Medium for the CD measurement , or in minimal medium containing 15N ( 15NH4Cl , ISOTEC , Sigma-Aldrich , USA ) and 13C ( 13C6 D-Glucose , Cambridge Isotope Laboratories , Inc , USA ) for the NMR analysis . 1 l of medium was inoculated with an overnight culture until the OD600 of the mixture was between 0 . 05 and 0 . 1 . The bacterial growth at 37°C and 220 rpm was observed until an OD600 of 0 . 8 was reached . At this point protein expression was induced by addition of IPTG ( final concentration 0 . 5 mM ) . For expression in minimal medium , 2 . 5 ml 20% 13C6-Glucose per liter of culture was added to the medium . The bacteria were incubated at 20°C and 220 rpm and harvested after 3 h . After centrifugation for 10 min at 5000 g and 4°C the bacteria were resuspended in 15 ml Lysisbuffer [50 mM NaH2PO4 pH 8 . 0 , 300 mM NaCl , 30 mM Imidazol , Lysozym , 6% TritonX-100 , 1 tab . Roche Complete Protease inhibitor without EDTA ( Roche , Mannheim , Germany ) ] per liter of minimal medium . After 10 min incubation at room temperature the bacteria were lysed by three freeze-thaw cycles performed with liquid nitrogen and warm water and sonification for 6×30 sec on ice ( Branson Sonifier B15 , level 7 , cycle 80% ) . The insoluble debris was removed by centrifugation ( Beckman JA17 rotor , 30 min , 31000 g ) at 4°C . The purification was started with a 5 ml Ni-NTA column ( Protino Ni-NTA Columns , Macherey-Nagel , Germany ) on an FPLC system ( LKB GradiFrac , Pharmacia Biotech , Freiburg ) with a flow rate of 3 ml/min . The UV absorbance at 280 nm was measured with a connected absorbance recorder ( LKB Optical Unit , LKB REC102 , Pharmacia Biotech , Freiburg ) to identify protein containing fractions . A step gradient of 50 mM and 100 mM imidazole was used to prevent unspecific protein binding , and elution was accomplished with 300 mM imidazole . Afterwards , the protein containing fractions were pooled and ultrafiltrated ( Amicon Ultra-15 , Millipore ) . The retentate was diluted in TEV-Buffer ( Invitrogen , USA ) and again ultrafiltrated until the NaCl concentration was less than 5 mM . The protein concentration was determined by measuring the absorbance at 280 nm to calculate the amount of AcTEV-Protease ( Invitrogen , USA ) that was needed to cleave off the N-terminal Z2-TAG . It was assumed that 10 U AcTEV-Protease could cleave 2 . 16 mg substrate . To prevent oxidation , the solution was overlayed with CO2 and incubated for several days at room temperature . Each day a 1 µl sample of the solution was collected and analyzed to check the cleavage process . After cleavage was completed , the cleavage products were separated with a reverse phase HPLC ( BT9200 Titan HPLC pump , Eppendorf , Hamburg ) using a C4-Reprosil 300 column ( Dr . Maisch GmbH , Ammerbuch-Entringen , Germany ) and a gradient from 20% to 60% acetonitrile with 0 . 05% trifluoroacetic acid . The absorbance at 280 nm was recorded ( BT9520 IN UV/Vis detector , LKB REC101 , Pharmacia Biotech , Freiburg ) and the eluent was collected in 2 ml fractions ( Foxy Jr . , ISCO ) and analyzed . Positive fractions were pooled and lyophilized for several days . Afterwards , the molecular mass of the protein was measured using a MALDI-MS ( Ultraflex I , Bruker ) to determine the achieved isotope labeling , the purity and the absence of oxidation products . Phosphate buffer ( PB ) , 2 , 2 , 2-trifluoroethanol ( TFE ) and sodium dodecyl sulfate ( SDS ) were obtained from VWR ( Darmstadt , Germany ) . The detergent n-dodecylphosphocholine ( DPC ) and the phospholipids DMPC 1 , 2-dimyristoyl-sn-glycero-3-phosphatidylcholine ( DMPC ) , 1 , 2-dimyristoyl-sn-glycero-3-phosphatidylglycerol ( DMPG ) and 1 , 2-dihexanoyl-sn-glycero-3-phosphatidylcholine ( DHPC ) used for vesicle and bicelle preparation were purchased from Avanti Polar Lipids ( Alabaster , AL , USA ) . A weighed amount of lyophilized Erns protein was dissolved in deionized water for preparing a 50 µM stock solution . SDS and DPC were used in a concentration of about 10 mM in 10 mM PB pH 6 . 5 with a protein/detergent ratio of 1∶600 . The lipid powders of DMPC and DMPG were dissolved in 50∶50 chloroform/methanol ( v/v ) to get lipid stock solutions of ∼7 mM . Aliquots of these stock solutions were mixed in a glass vial and thoroughly vortexed to obtain the DMPC/DMPG mixture ( 1∶1 molar ratio ) . Subsequently , the organic solvent was removed under a gentle stream of nitrogen , followed by overnight incubation under vacuum . The DMPC or DMPC/DMPG lipid film that had formed in the vial was dispersed by the addition of 200 µl PB and homogenized by vigorous vortexing for 7×1 min and by 7 freeze-thaw cycles . Afterwards , small unilamellar vesicles were formed by sonication of the multilamellar vesicles for 4 min in a strong ultrasonic bath ( UTR 200 , Hielscher , Germany ) . The sonication procedure was repeated 3 times ( with intermittent cooling of the water in the ultrasonic bath to room temperature with ice , to avoid overheating the samples ) . To prepare bicelles , a weighed amount of DHPC was first dissolved in 10 mM PB by sonification . An aliquot of this solution was used to dissolve a weighed amount of DMPC . Afterwards , the bicelle dispersion was homogenized by vortexing and freeze-thaw cycles as described above . Due to the significant technical challenges of examining bicelles in optical spectroscopy , which are caused by much stronger light dispersion compared to sonicated small unilamellar vesicles , the protein concentration of the sample was calculated by the UV-VIS absorption of a stock solution , from which the corresponding dilution factor could be determined . To prepare the samples for CD analysis , an aliquot of the Erns protein stock solution was added to PB , to a 50∶50 mixture of TFE/PB ( v/v ) , to SDS or DPC micelles , or to the corresponding lipid dispersions . The final protein concentration in PB , TFE/PB and in micellar environment was 15 µM . In the lipid vesicle samples the protein concentration was adjusted in the range from 13–27 µM and the lipid concentration between 0 . 7–1 . 8 mM , resulting in peptide-to-lipid ( P/L ) ratios of ∼1∶20 , 1∶50 , 1∶100 and 1∶200 . A 20 µM protein and 2 mM ( total ) lipid concentration and a P/L ratio of 1∶100 was set up in the bicelle samples by dissolving the lyophilized ErnsΔN protein directly in the bicelle dispersion . CD spectra were recorded on a J-815 spectropolarimeter ( Jasco , Groß-Umstadt , Germany ) in rectangular quartz glass cells of 1-mm path length ( Suprasil; Hellma , Müllheim , Germany ) between 260 and 185 nm at 0 . 1-nm intervals . The temperature was set to 25°C for the peptide solutions in PB , the 50% TFE mixture , and the micellar solutions , and at 30°C for the vesicle or bicelle suspensions ( i . e . , well above the lipid phase transition temperature of 23°C for DMPC and DMPG ) using a water thermostat-regulated cell holder . Three repeat scans at a scan rate of 10 nm/min , 8 s response time , and 1 nm bandwidth were averaged for each sample and for the baseline of the respective peptide-free sample . After subtracting the baseline spectra from the sample spectra , CD data were processed with the adaptive smoothing method in the Jasco Spectra Analysis software . To calculate the mean residue ellipticities required for quantitative secondary structure estimation , the concentration of the peptide stock solutions was determined from the UV absorbance of the respective peptide at 280 nm . For better comparison of the spectra of the different samples , the calculated mean residue ellipticity ( MRE ) is shown in the graphs . Secondary structure analyses were performed using the CDSSTR program [66] , [67] with the implemented singular value decomposition ( SVD ) algorithm; by the CONTIN-LL [68] , [69] program , which is based on the ridge regression algorithm; and by the SELCON-3 [70] , [71] program , which incorporates the self-consistent method together with the SVD algorithm to assign protein secondary structure . The three algorithms are provided by the DICHROWEB online server [72] , [73] . The quality of the fit between experimental and back-calculated spectrum according to the secondary structure fractions was assessed from the normalized root mean square deviation ( NRMSD ) , with a value <0 . 1 considered as a good fit [72] . Oriented protein-lipid samples for OCD measurements were prepared by depositing the proteolipid vesicles ( with DMPC and DMPC/DMPG , as described above ) on a planar quartz glass substrate . Each sample was generated by spotting a 60–80 µl aliquot of the vesicle sample onto a 20 mm diameter quartz glass plate ( SUPRASIL , Hellma , Jena ) as a ∼12 mm central circular spot , and dried under a gentle stream of air . Afterwards , the sample was rehydrated for 15 h at 30°C and 97% relative humidity in an OCD sample cell using a reservoir of saturated K2SO4 solution . The in-house built OCD cell can be integrated in a J-810 spectropolarimeter as an accessory , and further details on the OCD sample preparation and measurements have been described [33] , [74] . The thin oriented bilayers formed during hydration of the sample minimize the possibility of undesired spectral artifacts caused by linear dichroism or absorption flattening . The OCD spectra were recorded as an average of 8 scans with a 45° rotation of the cell after each scan to further reduce spectral artifacts due to linear dichroism arising from imperfections in the sample , strain in the quartz glass windows , or imperfect alignment of the window . For OCD measurements the same data acquisition parameters were used as in the conventional CD experiments described above . Background spectra of pure lipid bilayers ( without protein ) were subtracted from all OCD spectra . In order to compare the different OCD spectra in a better way all spectra were normalized to match their ellipticity around the minimum at ∼220 nm . The NMR analysis was carried out in a lipid bicelle system with a total lipid concentration of 200 mM in PBS ( 50 mM KH2PO4 pH 6 . 8 , 50 mM NaCl ) , using DHPC and DMPC ( Avanti Polar Lipids , Alabaster , AL , USA ) at a ratio of 4∶1 . For sample preparation , the weighed amount of DH/MPC was dissolved in 450 µl PBS by vortexing and sonification until the solution was clear . Afterwards , the required amount of DH/MPC was added and the mixture was again sonified . Thereafter , 0 . 5 µmol of the lyophilized protein and 50 µl D2O were added . After another round of sonification the insoluble material was removed by centrifugation for 20 sec . at 320 rcf to get a clear solution . The protein/lipid ratio obtained by this procedure was 1∶222 . NMR experiments were performed on a Bruker Avance I 600 MHz spectrometer with a broadband triple resonance probe , or on a Bruker Avance III 600 MHz spectrometer with a cryo probe and a Z-gradient . The 15N-HSQC and CLEANEX experiments [40] , and titration with Gd 1 , 4 , 7 , 10-tetraazacyclododecane-1 , 4 , 7 , 10-tetraacetic acid ( Gd-DOTA , Sigma-Aldrich ) were typically acquired with 2- 4 scans and a total of 128 increments in the indirect dimension , between 23 and 50°C . Titration with Gd-DOTA was performed on 0 . 5 mM sample of ErnsΔN in either DHPC/DMPC or DPC micelles . A CLEANEX spin-lock field of 4 . 8 kHz was applied for a mixing time of 100 ms . 3D 15N-HSQC NOESY and 15N-HSQC-TOCSY experiments were performed at 23°C with a mixing time of 120 ms ( NOESY ) and 60 ms ( TOCSY ) . 200–250 increments in the 1H dimension and 55 increments in the 15N dimension were acquired . A 3D 13C-HMQC NOESY was acquired in 200 mM deuterated DPC ( Avanti Polar Lipids , Alabaster , AL , USA ) . 1H 15N 13C triple resonance experiments ( HNCACB , HNCA , CBCA ( CO ) NH ) at 23°C [75] were used to assign the 1HN , 13C and 15N resonances . 15N longitudinal ( R1 ) and transversal ( R2 ) relaxation as well as the heteronuclear NOE ( het-NOE ) were measured at 27°C . Relaxation delays varied between 10 . 8 and 3466 . 8 ms for R1 , and 14 . 4 to 259 . 2 ms for R2 [76] . One duplicate point was included to test for instabilities . The heteronuclear NOE was determined as the signal intensity ratio of 1HN/N crosspeaks with and without 1H saturation . All experiments were recorded in an interleaved manner [77] . The water signal was suppressed with a combination of the water-flip-back and the WATERGATE scheme in all cases . Spectra were processed using the nmrPipe software package [78] and analyzed with NMRView [79] . The normalized proton exchange rate was calculated by the intensity of the peak in the CLEANEX spectra divided by the intensity of the correlated peak in the 15N-HSQC spectra Using the Monte Carlo simulation package SIMONA [80] we performed a total of 1 . 2×109 steps , each of which changed a single dihedral , for both Erns systems in the all-atom AMBER99SB-ILDN force field [41] , in combination with an accurate implicit description of the solvent and membrane interactions [44] . Starting from an ideal all-helical conformation in the proximity of the outermost membrane layer , all simulations showed a quick attachment to the membrane surface . Depending on the simulation temperature ( Erns - 220K , 270K , 320K , 360K , 370K , 380K , 400K; ErnsΔN - 300K , 320K , 340K , 360K , 370K , 380K , 400K ) , the secondary structure varies over the sequence . We determined the standard deviation for the per-residue secondary structure information from five distinct populations per temperature , using the secondary structure assignment package DSSP [81] . For each population we averaged the fraction of secondary structure elements over every 10 , 000th step , neglecting the initial 8×105 steps to allow for equilibration . MD simulations were conducted using the molecular simulation package GROMACS 4 . 5 . 5 [82] . The AMBER99SB-ILDN force field [41] was used for the peptide , together with the SLIPID force field for the DMPC bilayer . ErnsΔN ( Arg194 – Ala227 ) was constructed as an ideal helix with an acetylated N-terminus using the program xleap from the AmberTools package [83] . The peptide membrane complex was formed during an unrestrained membrane binding simulation of 10 ns length , by placing the peptide molecule parallel to a pre-equilibrated lipid bilayer at a distance of 1 . 8 nm above the lipid headgroups , at an elevated temperature of 480 K to speed up insertion . During the binding simulation , hydrogen bonds within the peptide were restrained to prevent unfolding . After cooling down , the system was equilibrated at 303 K with position restraints of 1000 kJ/ ( mol nm2 ) on the peptide for 500 ps . After this , an unrestrained MD simulation of 500 ns length was conducted using a Nose-Hover thermostat [84] and Parrinello-Rahman barostat [85] , with semiisotropic pressure coupling . A time step of 2 fs was used together with the LINCS algorithm [86] to constrain bonds involving hydrogen atoms . Long range electrostatics were treated via PME combined with a 1 . 4 nm direct space cutoff for vdW and Coulomb interactions . For another set of pulling simulations , the GROMACS pull code was used . The peptide was pulled into the membrane with a force constant of 10000 kJ/ ( mol nm2 ) and a pull rate of 0 . 2 pm/ps along the membrane normal for 10 ns . Then , the system with the peptide in the center of the bilayer was equilibrated for 500 ps using position restraints of 1000 kJ/ ( mol nm2 ) on the peptide . After that , another unrestrained MD simulation of 75 ns length was conducted . | The Erns protein ( envelope protein , RNase , secreted ) of pestiviruses represents one of the most fascinating proteins in virology . Erns is not only an essential structural component of the virus particle but also an unspecific RNase . The latter activity is dispensable for pestivirus replication but represents a virulence factor involved in the establishment of lifelong persistent infection . These functions of Erns are connected with its repressive activity on the type I interferon response of the infected host probably depending on secretion of part of the protein synthesized within the infected cell followed by its distribution with the blood stream . To understand the mechanisms leading to an equilibrium between intracellular retention ( for production of virus particles ) and secretion ( for repression of the innate immune response ) the principles of Erns membrane binding need to be better understood . The recently published Erns crystal structure , however , is lacking the relevant carboxyterminal membrane anchor region . We report here structure analyses of the Erns membrane anchor bound to model membranes . This work was based on circular dichroism , nuclear magnetic resonance spectroscopy , and structure simulations , and revealed a new type of membrane anchor for a surface protein . These data will help to explain the unusual functions of Erns . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"medicine",
"chemistry",
"veterinary",
"science",
"biology"
] | 2014 | Structure of the Membrane Anchor of Pestivirus Glycoprotein Erns, a Long Tilted Amphipathic Helix |
The unique ability of intrinsically disordered proteins ( IDPs ) to fold upon binding to partner molecules makes them functionally well-suited for cellular communication networks . For example , the folding-binding of different IDP sequences onto the same surface of an ordered protein provides a mechanism for signaling in a many-to-one manner . Here , we study the molecular details of this signaling mechanism by applying both Molecular Dynamics and Monte Carlo methods to S100B , a calcium-modulated homodimeric protein , and two of its IDP targets , p53 and TRTK-12 . Despite adopting somewhat different conformations in complex with S100B and showing no apparent sequence similarity , the two IDP targets associate in virtually the same manner . As free chains , both target sequences remain flexible and sample their respective bound , natively -helical states to a small extent . Association occurs through an intermediate state in the periphery of the S100B binding pocket , stabilized by nonnative interactions which are either hydrophobic or electrostatic in nature . Our results highlight the importance of overall physical properties of IDP segments , such as net charge or presence of strongly hydrophobic amino acids , for molecular recognition via coupled folding-binding .
It has become clear that many functional proteins do not fold into unique three-dimensional structures , as expected from the classical view of proteins , but remain highly conformationally dynamic under native conditions . These intrinsically disordered proteins ( IDPs ) are now estimated to represent a significant fraction of many genomes . For example , roughly half of the proteins in mammals are predicted to contain disordered segments of more than 30 amino acids in length , and a fourth to be fully disordered [1] . Interestingly , the prevalence of disorder is far from uniform among different functional classes , suggesting IDPs carry special and biologically relevant properties . In particular , the prevalence is high among proteins performing important regulatory and signaling functions [2] . The perhaps most striking IDP property is their ability to undergo a conformational disorder-order transition upon contact with a target molecule . Various biological advantages have been suggested to stem directly from this coupled folding-binding process [3]–[6] , including the ability to bind specifically to multiple and structurally diverse partners [7] . Indeed , intrinsic disorder has been found to be a common feature of “hub” proteins with especially large numbers of links in protein interaction networks [8] , [9] . An important example is given by the tumor suppressor p53 , a transcription factor and regulatory protein that can induce cell cycle arrest and apoptosis [10] . The C terminal domain of p53 ( henceforth simply p53 ) binds to at least four different globular proteins adopting four different conformations in the process [11]–[13] . How this binding diversity is achieved at the molecular level is not well understood and , in particular , cannot be rationalized by the classical lock-and-key model of protein-protein interactions . The binding diversity of IDPs can in principle lead to two different types of signaling modes , one-to-many , in which one disordered region binds to multiple folded proteins , and many-to-one , in which many disordered segments bind to the same folded protein [14] . An example of the latter is given by S100B , a -modulated homodimeric protein with several disordered protein partners [15]–[17] , one of which is p53 . Each S100B monomer consists of two EF-hand motifs which upon -binding reorients one -helix ( helix III ) , thereby exposing a hydrophobic binding pocket . In order to better understand this multiple specificity phenomenon , we study the coupled folding-binding of two disordered peptide sequences to S100B . As target peptides we choose p53 and a fragment from the protein CapZ , denoted TRTK-12 , which both have available experimental structural data for the complexes [16] , [18]–[20] . Molecular Dynamics ( MD ) simulation studies have previously been performed on the disordered N and C terminal regions of p53 , with either positional constraints on the peptide [21] or by focusing on the characteristics of the bound and free states separately [22]–[24] . There has , however , not been a study comparing the processes of different IDP segments binding to the same ordered target . MD simulations of biomolecular systems are typically hampered by limitations in conformational sampling . Large dynamical transitions such as coupled folding-binding of proteins is particularly challenging . We circumvent the problem by using a computationally efficient Monte Carlo-based implicit-solvent model , which has been extensively tested on the folding of peptides and small proteins [25]–[27] . Our basic approach has previously been tested on PDZ domain-peptide binding , with a variation of the potential energy function [28] , [29] . One advantage of this approach over recent protein-peptide docking methods [30]–[32] is that an equilibrium picture of the association is obtained . Additionally , we perform explicit-water MD simulations on IDP segments as free chains and in complex with S100B to further validate the MC result . Our work suggests striking similarities in the association mechanism of p53 and TRTK-12 , despite their different amino acid sequences . In particular , we find that both sequences populate a transient intermediate state which is stabilized by either hydrophobic or electrostatic interactions .
We start by examining the conformational behavior of our two peptide sequences as free chains using both MC and MD simulations , respectively ( see Methods ) . Despite no discernable sequence similarity between the p53 and TRTK-12 peptides , we find that the two chains behave qualitatively similarly , as seen from Figure 1 . Both peptides remain rather flexible at temperatures K and sample a wide range of conformations . A slight preference for -helical states can be seen . Consequently , the conformational fluctuations of the free peptide chains include structures resembling the bound state . These bound-like populations are nonetheless quite low ( see the low value tails of the RMSD distributions ) , in agreement with previous MD simulations on the p53 peptide [21] . We also find that the results of the two different simulation approaches are in good qualitative agreement . This is important as a validation of the full-scale protein-peptide simulations to be described below . The largest deviation appears for the p53 sequence , for which the MD results produce a peak in the RMSD probability distribution at around 6 Å . By inspection , we see that this peak represents peptide chains folded into a -hairpin-like structure . The MC simulations of p53 also produce a small population of -hairpins , albeit typically more fully formed , resulting in a shift of the peak to around 7 . 5 Å ( see Figure 1A ) . We note that the tendency for the free p53 peptide to form -hairpin-like structures was also observed in the MD study of Allen et al . [22] . Overall , we find that both the free p53 and TRTK-12 peptides sample multiple conformations , including transient -helical structures . We turn now to the binding processes of S100B and the p53 and TRTK-12 peptides , respectively . To this end , we rely on the computational convenience of the MC approach . The basic procedure follows earlier work [28] , [29] and operates in the following way . The S100B dimer structure is maintained in its native state by constraints , allowing side-chain and small backbone motions , while the peptide chains are free to explore the protein surface . Simulations are performed at fixed temperatures and are long enough to produce multiple binding and unbinding events in each trajectory such that an equilibrium picture of the interaction is obtained . Representative MC trajectories are shown in Figure S1 in Supporting Information . To monitor the progress of binding we use two different observables , and . The distance is taken between two points , the center-of-mass of the simulated peptide and the center-of-mass obtained from the peptide coordinates in the experimental complex structure . Hence , a small value indicates binding close to the experimentally determined binding site . The other observable , , is the number of amino acid contacts between the peptide chain and the S100B peptide binding pocket , which we have defined as a set of amino acid surface positions ( see Methods ) . Because of the symmetry of the S100B homodimer , with its two identical binding pockets , we determine the center-of-mass distance using ( 1 ) where and are obtained for the two binding pockets , respectively . For , we count the total contacts made with either binding pocket residues , thus assuming that a peptide can not contact both sites at the same time . This way , high and low is a signature of a peptide bound tightly at either of the two binding pockets . The overall features of the obtained folding and binding processes can be seen from the free energy surfaces in Figure 2 . Both sequences exhibit major free energy minima representing bound states , characterized by Å and 20–40 . This indicates that binding occurs mainly at the experimentally determined S100B peptide binding pockets . We emphasize that this result is obtained in unbiased simulations , in which the peptides are left free to explore the entire protein surface . We find additional , more detailed similarities between the binding processes of the p53 and TRTK-12 sequences . In the bound state , both p53 and TRTK-12 exhibit fluctuating helical contents . Hence , the peptides display a significant population of bound but not yet folded structures . This population is apparent in the one-dimensional free energy profiles in Figure 3 , which reveal a plateau-like behavior in representing an intermediate state between the fully bound and free states . The term “encounter complex” has been used to indicate such a metastable state in protein-protein interactions [33] . For coupled folding-binding of disordered chains , there are two extreme mechanistic possibilities for association corresponding to either folding-before-binding ( conformational selection ) or binding-before-folding ( induced folding ) of the peptide [34] . The significant population of “non-native” bound peptide conformations we observe for p53 and TRTK-12 indicate a binding mechanism dominated by induced-folding for both sequences . Similar observations have been made in the studies of Chen [21] and Pirolli et al . [23] for the p53 sequence , as well as for associations involving other IDPs [35]–[37] . In previous studies on PDZ domain-peptide interactions we found a two-state-like binding mechanism [28] , [29] , i . e . , where the bound state was reached via a single well-defined binding transition state . The difference with the current systems may be a length effect , as PDZ domains typically bind much shorter peptides ( around 4–6 amino acids ) . These short chains may not have sufficient chain flexibility to promote the formation of intermediate states [38] . X-ray and NMR structural studies on p53 [16] and TRTK-12 [19] , [20] show that both peptides bind the hydrophobic surface on S100B , exposed upon -binding , but with somewhat different configurations . In complex with S100B , the p53 peptide is almost fully helical and docked parallel with the S100B helix III while TRTK-12 has only a single -helical turn and is oriented perpendicular to helix III . The two minimum-energy conformations obtained from our MC binding simulations of p53 and TRTK-12 , respectively , manage to capture many of these characteristics as shown in Figure 4 . The TRTK-12 peptide in particular has in effect assumed the correct structure , being only partly helical and its tryptophan ( Trp271 ) sidechain directed into the hydrophobic binding pocket . The min- conformation of p53 deviates slightly from the experimental structure . As depicted in Figure 4A , its phenylalanine ( Phe385 ) is contacting the hydrophobic pocket rather than being solvent exposed , as in the experimental structure [16] . This hinders the p53 peptide from being fully -helical . However , the overall chain orientation relative to helix III is essentially correct . It is clear from the binding free energy surfaces in Figure 2 that both peptides exhibit a significant structural diversity in the bound state . Minimum-energy structures may therefore not always provide a complete picture . The p53 peptide has a single , smooth bound state free-energy minimum spanning 2–10 Å . It turns out that this state is structurally quite well represented by the p53 min- conformation in Figure 4A , with contacts between the S100B binding pocket and the p53 Phe385 as a prominent feature . By contrast , TRTK-12 has two distinct bound state minima characterized by 2–5 Å and 10 Å , respectively . The underlying structures of the small- ( global ) free energy minimum are similar to the min- conformation . Interestingly , the local minimum at 10 Å exhibited by TRTK-12 corresponds to an -helix parallel to the S100B helix III , reminiscent of the experimental binding pose of p53 [16] . To make a more detailed comparison with experimental data , we define the bound state ( BS ) by and for the p53 and TRTK-12 systems , respectively , based on the free energy profiles in Figure 3A . The populations of all possible amino acid contacts in the BS between the peptide and the S100B binding pocket are illustrated in Figure 5 . p53 mainly binds to the binding pocket via its hydrophobic residues Leu383 , Met384 and Phe385 , in contrast to TRTK-12 which interacts with all its amino acids . This difference aside , the simulations suggest that the largest hydrophobic residue in each peptide fulfills a similar function . For p53 , Phe385 is involved in all of the most populated contacts , equivalent to Trp271 in TRTK-12 . These residues interact mainly with the hydrophobic amino acids of the binding pocket , especially with Ile36 , Leu44 , Val56 and Phe76 . Our simulations suggest more involvement of the p53 Phe385 than what at first glance is apparent from the NMR structure . However , the importance of this position for binding is implied in the study of Rustandi et al . [39] , who found that mutation of the Phe to a Trp , a larger hydrophobic amino acid , increases the binding affinity by 3–4 fold . As demonstrated in Figure 6 , most S100B binding-pocket residues make on average the same number of contacts with p53 as with TRTK-12 . There are however some positions , in particular Lys55 and Cys84 , which are mostly involved in binding TRTK-12 . These two residues indeed form contacts with TRTK-12 in the experimental structures [19] , [20] , which we further find to be stable also in MD simulations of the S100B-TRTK-12 complex , performed as described below ( see Figure S2 ) . While there is an overlap between the obtained contact maps in Figure 5 and the set of interactions derived from experiments , our BS ensembles do not perfectly fit with the experimental NMR structures . Some residue-residue contacts are only present in the simulations and , conversely , some contacts are found in the experimental structures but not in the simulations . The latter mainly emerges in the p53 simulations , where the two inter-molecular salt bridges ( Arg379–Glu45 and Lys386–Glu86 ) , as well as contacts involving the N-terminal part of the p53 chain , are mostly absent in our simulations . The reason is probably the rather frequent interaction of Phe385 with the hydrophobic S100B binding pocket , which lead to a disrupted folding of the p53 -helix . Our BS conformational ensembles are also more diverse than the set of NMR derived model structures . This can be seen by calculating the average RMSD between pairs of structures ij in the NMR and BS ensembles , respectively . For p53 , we obtain Å for the NMR ensemble of S100B-p53 structures and Å for the simulation-derived BS ensemble . The corresponding values for TRTK-12 are Å and Å , respectively . It is possible therefore that our model is not able to capture the folding of the peptides in full , at least for p53 . There is some uncertainty , however , in the interpretation of the nuclear Overhauser enhancements ( NOEs ) intensities which underly NMR structure calculations , especially with regard to the diversity of the protein ensemble [40] , . The differences may also be due to the relatively weak binding conditions used in our simulations , where the BS is populated only at around . At lower temperatures , the BS ensemble will likely become structurally more well-defined . In the analysis of our binding simulations we rely on available experimental structures of the protein-peptide complexes . For S100B and TRTK-12 , three structures have been proposed [18]–[20] . One of these involves a novel coil-like conformation of TRTK-12 ( PDB ID 1MQ1 ) [18] while the NMR and X-ray structures of Weber et al . both indicate a partially -helical conformation ( PDB ID 1MWN and 3IQQ ) [19] , [20] . Our MC results are in line with the presence of a short -helix in TRTK-12 , as demonstrated by the minimum-energy conformation in Figure 4B . To further inspect the experimental disagreement in TRTK-12 structure , we employed our MD approach to test the stability of two of the proposed structures , 1MQ1 [18] and 1MWN [19] . Three simulations of 50 ns were performed for each complex , both with and without loaded in S100B , as shown in Figure 7 . The calculations consistently indicate a higher stability for 1MWN than for 1MQ1 , at least in terms of RMSD fluctuation measured over the TRTK-12 chain . The peptide secondary structure in the 1MWN trajectories remains intact over the full MD trajectories ( data not shown ) . These MD trajectories thus also support the partly -helical structural models of Weber et al . [19] , [20] and , moreover , provide further support for an agreement between our two different computational approaches . How do the peptide chains reach their respective bound states or , in other words , what is the peptide binding mechanism ? Key to answering this question is partly bound peptide conformations populated transiently during the binding process . Although we cannot directly assess kinetic aspects of the binding process with our MC approach , we can examine the structural characteristics of peptide conformations which are neither entirely bound nor in the unbound state . As shown above , both sequences exhibit similar plateau-like regions in the free energy profiles ( see Figure 3 ) . Many of these conformations must be populated during binding and it is therefore of interest to characterize and compare these states . We define this intermediate state , IS , as conformations which are neither unbound ( ) nor part of the bound state , BS . Contrary to the BS ensembles , the IS of the two peptide sequences exhibit strong similarities . To show this , we performed a clustering procedure on all IS conformations with the aim of visualizing typical structures ( see Methods for details ) . The four largest clusters for p53 and TRTK-12 , respectively , are displayed in Figure 8 , representing more than half of the IS ensemble in both cases . As depicted in Figures 8A and C , the IS peptide conformations are located primarily in two different regions in the periphery of the S100B binding pocket . In the first group , represented by the violet and the green cluster centroid structures in Figure 8 , the peptides are located in the vicinity of a hydrophobic surface composed primarily of Phe43 and Phe87 on the S100B domain . Contacts between these amino acids and the peptides are also evident from the IS contact probabilities in Figure 5 , which moreover indicate that they involve mainly the peptide residues Phe385 and Trp271 on p53 and TRTK-12 , respectively . Interestingly , while these contacts are among the most frequent in IS , they are much less populated in the BS . In the second IS group , indicated by the red and the yellow centroids in Figure 8 , the peptides are situated between helix III and helix IV of S100B . Here the stabilizing interactions are instead primarily electrostatic . The possible existence of binding spots in the periphery of the binding pocket is in part supported by the work of Arendt et al . [42] , who found that small molecules from a fragment library mainly bound to two sites of the S100B surface , one of which overlaps with the first region described here . The similar binding mechanism of p53 and TRTK-12 is especially interesting in light of their low sequence similarity and the significant structural differences in their bound states . The S100B interaction with disordered peptides can be compared with so-called peptide recognition domains ( PRDs ) , which are common modules in signaling proteins and include 14-3-3 , PDZ , and SH3 [43] . These domains typically bind sets of peptide sequences conforming to simple linear sequence motifs . For example , SH3 domains bind peptides comprising a Pro-X-X-Pro motif , where Pro is proline and X any amino acid type . A structural analysis of the many-to-one binding exhibited by a particular 14-3-3 domain revealed a relatively high structural similarity in the protein-peptide complexes involving five different target peptides [14] . Similarly , PDZ domains typically bind their peptides in a structurally specific way involving the peptide C terminus [44] . The multiple specificity exhibited by S100B thus differ from PRDs in at least two ways . First , the peptide sequences do not conform to a simple linear motif and , second , their bound structures differ significantly . A possible biological benefit of these two properties is that cross-reactivity among different S100B target peptides may be minimized . In other words , the relatively high sequence disparity among the IDP targets keep them from acquiring each others general functions . For example , the difference in sequence between CapZ/TRTK-12 and p53 may keep CapZ/TRTK-12 from binding to p53 partner domains other than S100B . Despite the lack of a simple sequence motif for known S100B targets , we note some interesting common characteristics . Both p53 and TRTK-12 include a single aromatic amino acid ( Phe385 and Trp271 , respectively ) , one or more aliphatic amino acids , as well as several Lys and Arg resulting in a net positive charge of the peptide . These characteristics holds also for a third disordered S100B target for which a structure of the complex is available [45] , taken from the N terminal regulatory domain of NDR kinase . It is possible that this basic compositional similarity of target peptides underpins the similarity in the binding mechanism , such that interactions made early in the binding process , being diverse and nonspecific in nature , are largely determined by the overall physical properties of the chain segment . An issue which remains to be addressed is to what extent a common binding mechanism is a general characteristics of many-to-one protein signaling . For this comprehensive additional simulations beyond the scope of the present investigation are required . However , we note that preliminary simulations for the interaction between S100B and NDR , performed using our all-atom MC approach , produced a minimum-energy conformation where the ( sole ) aromatic residue of the NDR peptide is involved in binding in a similar way as for p53 and TRTK-12 ( see Figure S3 ) . We have used a combination of MD and MC all-atom simulations to understand the coupled folding and binding of two disordered peptides , p53 and TRTK-12 , to the -loaded form of S100B , as an example of a many-to-one signaling mechanism . Despite the significant difference in sequence , we find remarkable and unexpected similarities in their association behavior with S100B . First , as free peptides , the two sequences have similar predispositions towards helical conformations , but overall remain rather flexible chains . Only to a small extent do they sample their respective S100B-bound structures . Second , both p53 and TRTK-12 populate an intermediate , metastable state during the folding-binding process , which may serve as an initial encounter complex of the interaction . The intermediate state is divided structurally between two different binding surfaces on the periphery of the binding pocket where stabilization occur by either hydrophobic or electrostatic interactions and includes a significant fraction of -helical structure ( see Figure 8 ) . Interestingly , major disparities between the two S100B targets become apparent only in the final bound state , where the patterns of contacts with the S100B surface are significantly dissimilar .
MD simulations of the free p53 and TRTK-12 peptides were started from random conformations which were obtained from high temperature ( 400 K ) MD simulations . 20 simulations were performed for p53 and TRTK-12 , respectively . The duration for each simulation was 200 ns . The simulations of the –S100B–TRTK-12 complex were initialized from the two available NMR structures ( PDB IDs 1MWN and 1MQ1 ) . For each version three runs of 50 ns , with different initial velocity distributions , were performed . In 1MQ1 , the coordinates for the four ions are missing . They were therefore added by superimposing the structure of 1MWN on 1MQ1 and transferring the coordinates of the ions from 1MWN to 1MQ1 . A final manual adjustment of the ion positions was made in order to avoid steric repulsions with the proteins . We also performed three simulations of each complex variant without the ions present . In order to maintain the structure of S100B , position restraints on the backbone atoms were added . All simulations were performed using the program GROMACS 4 . 07 [46] , [47] with the OPLS-AA/L force field [48] and the SPC/E water model [49] . The starting conformation was placed in the center of a cubic water box with at least 10 Å from the box edge . Periodic boundary conditions were used and counter ions ( Na+ or Cl− ) were added to neutralize the net charges . The long-range electrostatic interactions were treated with the particle mesh Ewald method [50] . The cutoff distances were set to 10 Å for short-range Coulomb and van der Waals interactions . The bond lengths were fixed by the LINCS algorithm [51] , and a time step of 2 fs was used . Each system was first relaxed by 1000 steps of steepest-descent energy minimization . After the minimization , the system was equilibrated at 298 K for 100 ps under an NVT ensemble and further equilibrated for 200 ps at constant pressure ( 1 bar ) . V-rescale [52] and Parrinello-Rahman [53] were used to couple the system to the simulation temperature and pressure with coupling constants of 0 . 1 ps and 2 . 0 ps , respectively . Production simulations were performed at constant temperature ( 298 K ) and pressure ( 1 bar ) . Coordinates were saved every 5 ps . For the free peptides , 10 simulations were performed with the PROFASI program package [25]–[27] using a simulated tempering algorithm [54] , [55] . Each run consisted of MC steps and visited 6 temperatures in the range 280–330 K for p53 and 201–301 K for TRTK-12 . The temperature intervals were chosen to envelop the protein–peptide simulation temperatures . 10 simulations were also performed for each S100B-peptide system , where every run consisted of MC steps . All binding simulations were performed in a cube with side length 150 Å and periodic boundary conditions . The temperature was held fixed , chosen such that the peptide would be bound roughly half of the time ( K for p53 and 270 K for TRTK-12 ) . The protein had flexible backbone and sidechains , but was kept close to its native bound conformation ( taken from the NMR structures with PDB IDs 1DT7 and 1MWN , where the best representative conformers were chosen ) . This was done by using a total energy function , where is the original physical energy function defined in Ref . [27] . Sampling of protein conformations close to the S100B native state was ensured by the non-optimized root-mean-square deviation penalty term given by ( 2 ) where and are the positions of the Cα atom of amino acid i in the simulated and experimental protein structures , respectively , and is the number of amino acids in the protein . The additional energy term constitutes typically less than 25% of the total energy in the system . Four flexible N-terminal residues , Phe88–Glu91 , of each S100B monomer were removed . The average amount of secondary structure in the free peptides was determined using STRIDE assignments [56] . The similarity to the natively bound structure was analyzed by calculations of the peptide backbone RMSD ( translationally and rotationally optimized ) with respect to the best representative conformer of the NMR structures ( 1DT7 for p53 and 1MWN for TRTK-12 ) . As an indication of whether the peptide is bound to the correct binding site , we used the observables and . is the distance between the centers-of-mass of the simulated and the experimental peptide in the complex , respectively . is the number of contacts between the peptide and the protein binding pocket , where two residues are defined to be in contact if they exhibit at least two pairs of heavy atoms separated by less than 4 . 5 Å . The residues defining the protein binding pocket are listed in Figure 5 . In order to characterize the most populated structural states of the intermediate , we used a Complete Linkage Hierarchical clustering method [57] . The aim of this procedure is to group protein-peptide conformations that are similar to each other into clusters . To determine similarity we used the standard root-mean-square-deviation measure , RMSD . The cutoff value in the clustering procedure was taken to be 6 Å and 8 Å for TRTK-12 and p53 , respectively , which resulted in roughly the same number of clusters for both sequences . The clustering procedure guarantees that within each cluster , the RMSD between all pairs of conformations is less than the chosen cutoff value . | A substantial fraction of our proteins are believed to be partly or completely disordered , meaning that they contain regions that lack a stable folded structure under typical physiological conditions . This is a feature which plays a key role in their functions . For example , it allows them to have many structurally different binding partners which in turn permits the construction of the intricate signaling and regulatory networks necessary to sustain complex biological organisms such as ourselves . Whereas measuring the binding strengths of associations involving disordered proteins is routine , the binding process itself is today still not fully understood . We use two different computational models to study the interactions of a folded protein , S100B , which can bind various disordered peptides . In particular , we compare two peptides whose structures are known when in complex with S100B . Our results suggest that , although the peptides assume different structures in the bound state , there are similarities in how they associate with S100B . The possibility to computationally model the interplay between proteins is an important complement to experiments , by identifying crucial steps in the binding process . This is essential to understand , e . g . , how single mutations sometimes lead to serious diseases . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Methods"
] | [
"biophysic",
"al",
"simulations",
"protein",
"folding",
"biology",
"computational",
"biology",
"biophysics",
"simulations",
"biophysics"
] | 2012 | Binding of Two Intrinsically Disordered Peptides to a Multi-Specific Protein: A Combined Monte Carlo and Molecular Dynamics Study |
Meiotic recombination is a major factor of genome evolution , deeply characterized in only a few model species , notably the yeast Saccharomyces cerevisiae . Consequently , little is known about variations of its properties across species . In this respect , we explored the recombination landscape of Lachancea kluyveri , a protoploid yeast species that diverged from the Saccharomyces genus more than 100 million years ago and we found striking differences with S . cerevisiae . These variations include a lower recombination rate , a higher frequency of chromosomes segregating without any crossover and the absence of recombination on the chromosome arm containing the sex locus . In addition , although well conserved within the Saccharomyces clade , the S . cerevisiae recombination hotspots are not conserved over a broader evolutionary distance . Finally and strikingly , we found evidence of frequent reversal of commitment to meiosis , resulting in return to mitotic growth after allele shuffling . Identification of this major but underestimated evolutionary phenomenon illustrates the relevance of exploring non-model species .
Accurate chromosome segregation at the first meiotic division often requires crossovers ( COs ) between homologous chromosomes [1] . Meiotic COs result from the repair of programmed DNA double strand breaks ( DSBs ) by homologous recombination , that can yield gene conversions through non reciprocal transfer of genetic variations ( GCs ) [2] . Only a fraction of DSBs yields CO-associated GCs , the remaining yielding GCs without associated COs , called non-crossovers ( NCOs ) . Overall , meiotic recombination provides a large source of genetic diversity by rearranging allelic combinations and therefore plays a key role in evolution . The budding yeast S . cerevisiae has long been a model to dissect recombination mechanisms . Meiotic recombination is initiated by Spo11-catalyzed DSBs [3] within nucleosome depleted regions enriched in gene promoters [4] in the context of chromatin loops organized around the synaptonemal complex [5] . Meiotic DSBs and COs occur within well-characterized hotspots across the genome . The mapping of 4 , 163 COs and 2 , 126 NCOs from 46 S . cerevisiae tetrads [6] showed that the frequency of meiotic recombination events correlates with previously localized DSBs [7 , 8] . It was also highlighted that the frequency of segregating chromosome without a CO is low in S . cerevisiae ( 1 out of 46 tetrads ) , supporting the importance of at least one CO per chromosome for accurate segregation [9 , 10 , 11] . Meiotic DSBs and associated recombination hotspots , driven by a sequence determinant , have long been expected to be extremely unstable across evolution because of their supposed inherent self-destructive nature [12] . Indeed , being preferentially cut , the strongest hotspots are expected to be preferentially converted with the weakest ones . This is the case in some mammalian species where DSB hotspots are determined by the sequence specific DNA binding of the fast evolving PRDM9 methyl transferase protein [13] . Remarkably , meiotic recombination initiation hotspots show a good conservation within Saccharomyces species despite a sequence divergence almost 10-fold higher than between human and chimpanzees that completely lack hotspot conservation [14–17] . The same observation was made for two Schizosaccharomyces species [18] . This likely results from the fact that DSB hotspots in yeasts lie in gene promoters , which are conserved functional elements and do not depend on sequence specific elements [15 , 19] . However , the Saccharomyces species have almost completely collinear genomes , which could favor hotspot conservation . Here , we assessed the properties and variations of the meiotic recombination landscape by taking advantage of the broad evolutionary range of the Saccharomycotina species [20] . We explored the recombination landscape of the Lachancea kluyveri yeast species ( formerly Saccharomyces kluyveri ) that shares the same life cycle as S . cerevisiae but is a distantly related species , which diverged from it prior to a whole-genome duplication event for more than ~100 million years ( i . e . protoploid species ) [21 , 22] .
To explore the L . kluyveri meiotic recombination landscape genome-wide we generated a diploid hybrid by mating two polymorphic haploid isolates , NBRC10955 ( MATa ) and 67–588 ( MATα ) [23] . The NBRC10955/67-588 hybrid contains approximately 85 , 705 polymorphic sites distributed across the eight homologous chromosomes , resulting in a genetic divergence of ~0 . 7% . It sporulates efficiently , with a spore viability of 73% ( S1 Fig ) . Overall , the divergence is in the same order of magnitude to those of the S288c/SK1 and S288c/YJM789 hybrids used for meiotic tetrad analysis in S . cerevisiae [6 , 24] . By whole genome sequencing , we genotyped all four viable spores from 49 meioses of the NBRC10955/67-588 hybrid . We selected a set of 56 , 612 reliable SNPs as genetic markers to define the allelic origin along the chromosomes ( see material and methods and S2 and S3A Figs ) . This resulted in a median distance of 196 bp between consecutive markers , with only 27 inter-marker distances higher than 5 kb . These markers are well distributed across the genome , with the exception of the rDNA array on chromosome H and the subtelomeric regions . Subtelomeric regions carried low quality sequences and translocation events , which were manually determined . To avoid spurious identification of recombination events at the ends of the chromosomes , we disregarded some subtelomeric regions , adapting case by case their size according to the presence of translocations and/or to the poor sequence quality ( S4 Fig ) . Altogether these regions range from 2 to 32 kb in size and represent 2% of the entire genome ( 251 kb ) . From these data , we performed a detailed analysis of individual recombination events using the CrossOver program [25] . Across the 49 meioses , we identified 274 NCOs and 875 single COs ( Fig 1A and S1 Table ) , 541 of which were associated with a detectable GC tract . Small GC tracks associated or not to COs might not overlap with enough genetic markers and the corresponding number could be underestimated . Surprisingly , we observed a high frequency of double COs , which correspond to two COs at the same location involving the four chromatids ( 138 events , 16% of the total COs ) ( Fig 1A ) as well as a high frequency of regions with a 4:0 allelic segregation ( 505 regions ) ( Fig 1A ) . Previous S . cerevisiae studies reported mostly single COs and GC tracts with 3:1 allelic segregation [6 , 24 , 25] . These types of events involving the four chromatids simultaneously are usually expected to reflect mitotic recombination events preceding meiotic entry and are generally disregarded [6 , 24] . In our case , their quantitative importance prompted us to examine them more carefully . We identified a total of 505 regions with a 4:0 segregation . The size distribution of these tracts follows a bimodal pattern with an almost equal partitioning of events in both clusters ( Fig 1B ) . Based on this distribution , we defined two categories: short 4:0 regions ( median = 2 . 3 kb , N = 213 ) and large 4:0 regions ( median = 224 kb , N = 292 ) with a size cut-off of 10 kb ( Fig 1B ) . The size distributions of the 213 short tracts smaller than 10 kb associated with COs and NCOs mimic the size distributions of the 3:1 conversion tracts associated with COs and NCOs ( Fig 1C ) . The 292 large regions with a 4:0 segregation bear all the characteristics for being the results of mitotic G2 COs prior entry into meiosis that lead to large regions of loss of heterozygosity ( LOH ) , ranging from 11 to 1 , 882 kb . These 4:0 segregating regions and the double COs were found within 38 tetrads . In these tetrads , the number of large 4:0 regions varies between 2 to 18 and their cumulative size ranges from 5% to 63% of the genome ( Fig 2A and 2B ) . All these large LOH regions are unique to each tetrad and the conserved allelic version seems random leading to an equal frequency of each parental allele ( Fig 2B ) . However , these large LOH events are not equally distributed along the genome . Centromere-proximal regions show a very low rate of LOH events . Every chromosome carries LOH events with the exception of the 1-Mb introgressed region ( Sakl0C-left ) carrying the MAT locus that displays a systematic 2:2 segregation ( Fig 2 ) . Overall , the frequency and repartition of 4:0 segregating regions and double COs support a high proportion of mitotic events prior meiotic entry . Such a mitotic genomic instability is unexpected , except from some mutants of genes involved in genome integrity maintenance for instance . To test the mitotic stability of the NBRC10955/67-588 hybrid , we induced an accumulation of mitoses by generating two independent mutation accumulation lines . They went through 12 single cell bottlenecks on rich medium representing ~ 240 generations . The sequence analysis of the two evolved diploids as well as the initial parental diploid ( only ~ 40 generations ) revealed no LOH event ( S5 Fig ) . In conclusion , the frequency of mitotic recombination on rich medium of the NBRC10955/67-588 hybrid is too low to explain the high amount of LOH events in the dissected tetrads , and therefore that these events most likely occurred during meiosis . The only alternative explanation left is that the high frequencies of observed LOH , double CO and 4:0 GC events are due to the so-called “return to mitotic growth” ( RTG ) , alternatives such as very active meiotic recombination hotspots but with peculiar recombination properties being unlikely . RTG is a meiotic reversion where cells enter into meiotic prophase , initiate inter-homolog meiotic recombination , but return to vegetative growth before meiosis I [26 , 27] . The reshuffled diploid obtained post RTG will go through a second meiosis resulting to the studied tetrad ( Fig 3 ) . The recombination events induced by the abortive meiosis occurring before the RTG are indistinguishable from inter-homolog mitotic G2 events ( Fig 3 ) . In support of a meiotic origin , we found a lack of LOH events near the centromeres , similar to canonical meiotic recombination events . However , two out of the 38 tetrads display LOH regions including all of the eight centromeres . A possible explanation for these two cases would be that RTG occurred after meiotic anaphase I . Such events would have escaped the meiotic commitment thought to be irreversible after meiotic prophase in S . cerevisiae through a positive feedback expression of the Ndt80 meiotic transcription factor [28] . Finally , RTG implies at least some residual growth on sporulation medium . We therefore isolated individual NBRC10955/67-588 hybrid cells on sporulation medium , and after two days , we did observe small colonies of about 0 . 5 mm of diameter , corresponding approximately to ten generations . This indicates that potassium acetate used as sporulation medium and cellular nutrients storage are enough to ensure residual growth , before starting sporulation . A possible explanation is that potassium acetate does not induce a strong enough starvation stress , due to the more respiratory lifestyle of this species [29 , 30] . Since RTG was highlighted , it has been suspected to happen in natural conditions and induce phenotypic diversity by generating new allelic combinations [31] . The ease with which L . kluyveri undergoes RTG leads us to consider that its frequency in natural conditions is high . The resulting LOH regions allow the expression of recessive alleles and new combination of interactive alleles . A recent study also demonstrated that COs induced during RTG can be used to generate an allelic reshuffled population allowing genetic linkage analysis [31] . Interestingly , RTG is initiated in stress conditions and the resulting allelic shuffling might increase the fitness to environment , immediately assuring cell survival . Therefore , the fact that RTG is easily induced in some yeast species can reveal a new mechanism of adaptation to stress , the generality of which remains to be determined . Our recombination map revealed that the main hotspot associated with RTG events , i . e . double COs , 4:0 NCOs and LOH regions corresponds to the VMA1 gene located on chromosome F ( hotspot 5 , Fig 2C and S6 Fig ) . This gene encodes a subunit of a vacuolar ATPase and the PI-Sce1 endonuclease ( also called Vde1 ) in S . cerevisiae as well as in L . kluyveri [32 , 33] . This is a site-specific endonuclease active during meiosis that promotes a homing reaction ( S7 Fig ) . In a heterozygous diploid where only one of the VMA1 alleles carries PI-Sce1 , the endonuclease cleaves the VMA1 sequence lacking the endonuclease-coding portion to initiate homing , which consists of a recombination reaction associated or not with a CO ( S7 Fig ) . Using de novo genome assembly , we defined which version of VMA1 was carried by each parental strain as well as in the other 26 L . kluyveri natural genomes already sequenced [23] . In total , nine strains harbor a VMA1 gene with PI- Sce1 , including our parental strain NRBC10955 , and nineteen isolates harbor an empty VMA1 gene , including the other parental strain 67–588 . Therefore , meiosis specific homing of PI-Sce1 explains the high recombinogenic activity at the VMA1 locus observed in our dataset that includes five COs , five double COs , nine short 4:0 NCOs and ten LOH events ( S6 Fig ) . Note that in a single meiosis , PI-Sce1 can cut one or two chromatids generating 3:1 and 4:0 NCOs as well as single and double COs , respectively . Therefore , in this specific case , only borders of long tracts of LOH can be unambiguously attributed to RTG events . To estimate the number of recombination events per meiosis in L . kluyveri , we excluded double COs , 4:0 NCOs and large LOHs that are considered to result from RTG events . This results in 334 COs without detected GC , 541 single COs with 3:1 GC tracts and 274 NCOs in 49 meioses ( Fig 1 and S1 Table ) . We also estimated that at least 100 single COs occurred within LOH regions ( see below ) . This corresponds to 19 . 9 COs and 5 . 6 NCOs per meiosis on average . RTG events comprise 138 double COs , 178 large LOH tracts within chromosomes ( 100 + 55 + 23 , Fig 1 ) , 114 large LOH tracts encompassing one telomere and 132 4:0 NCOs . Each double CO corresponds to one CO per RTG . Large internal LOH tracts account for two COs per RTG . Telomere proximal LOH tracts account for one CO per RTG . Altogether , we observed 608 COs ( 138 double COs + 470 LOHips , Fig 1 ) and 132 NCOs from RTG within 38 meioses , corresponding to 16 COs and 3 . 5 NCOs per RTG on average considering that only one RTG occurred . The detected recombination events per RTG correspond to an underestimation of the actual number of events since we analyzed only one of the two daughter cells after RTG ( Fig 3 ) . Considering that not all of the recombination events occurring during RTG are detected ( as describe in [31] ) , we estimate that the average number of COs per RTG is similar to those per meiosis ( S8 Fig ) . The presence of LOH regions also leads to an underestimation of meiotic recombination events that occur in such regions . Even numbers of COs within a LOH region involving only two chromosomes are missed as well as all NCOs . Considering the global recombination rate , we estimate that about 190 COs are expected within LOH regions , 90 of which being not undetectable ( see Methods ) . Both COs and NCOs are significantly less abundant than in S . cerevisiae that has a similar genome size , and where 73 to 90 COs and 27 to 46 NCOs have been detected per meiosis , depending on the studied hybrid [6 , 24] . We used a stringent threshold for the detection of conversions ( see Methods ) . This may have led us to miss some short conversions . Indeed , the median size of L . kluyveri conversion tracts is 2 . 9 kb for those associated with COs , and 2 . 2 kb for NCO conversion tracts ( Fig 1B ) compared to 2 kb and 1 . 8 kb in S . cerevisiae , respectively [6] . However , our strategy should reveal virtually all COs . Therefore , we can conclude that the genetic map of the L . kluyveri hybrid used is three to four times smaller than that of S . cerevisiae . As previously observed in S . cerevisiae and in other species , the average number of COs per chromosome is correlated to chromosome length in L . kluyveri . Remarkably , the linear relation between chromosome size and number of COs displays a positive intercept of 1 . 0 CO for a virtual chromosome size of 0 bp both in S . cerevisiae and L . kluyveri ( Fig 4A ) . This suggests that in these species the meiotic program is set up in such a way that at least one CO will form per chromosome independently of its length . S . cerevisiae and L . kluyveri have a similar genome size of approximately 12 Mb and therefore the recombination rate is estimated to ~1 . 6 COs/Mb and ~6 COs/Mb for L . kluyveri and S . cerevisiae , respectively . This difference underlines the particularly high recombination rate of S . cerevisiae that is rather uncommon in other eukaryotes ( Fig 4A ) . Indeed , the fission yeast Schizosaccharomyces pombe , which also harbors a 12 . 6 Mb genome size , displays a number of COs ranging from 34 to 44 per meiosis , depending on the hybrid studied , leading to approximately 3 COs/Mb per meiosis [34 , 35] . Recombination analyses were also performed in other more distant fungi species such as Agaricus bisporus and Gibberella zeae , where the recombination was estimated to be ~0 . 78 COs/Mb and ~0 . 65 COs/Mb per meiosis , respectively [36–38] . The size of the smallest chromosome might explain part of the difference in recombination rates [39] . Indeed , species carrying small chromosomes tend to display high recombination rates to stochastically ensure at least one CO on the smallest chromosome . This is illustrated by a negative correlation between the size of the smallest chromosome and the overall recombination rate across species [39] . We compiled meiotic data from 30 species to complete the one obtained in L . kluyveri and compared for all of them the recombination rate with the size of the smallest chromosome . As expected , we observed a negative correlation between these two parameters ( R2 = 0 . 79 , S9 Fig ) and L . kluyveri recombination data are consistent with this correlation . Whether the chromosome number , 8 vs . 16 in L . kluyveri and S . cerevisiae , respectively , impacts the global meiotic recombination rate is unknown . Mechanistically , several hypotheses could explain the reduced rate of COs in L . kluyveri , such as less initiating meiotic DSBs than in S . cerevisiae , or a more frequent use of the sister chromatid as a template for DSB repair . We therefore compared the DSB level between S . cerevisiae and L . kluyveri on chromosomes of similar length by pulse-field gel electrophoresis ( PFGE ) . We performed this experiment in the homozygous diploid CBS10367 isolate that has been independently chosen for its fast and efficient sporulation properties , although it shows a two-hour delay in meiotic entry compared to the S . cerevisiae SK1 isolate ( S10A Fig ) . At comparable time points after meiotic induction , i . e . four and six hours , respectively , we found similar DSB levels , suggesting that the three to four-fold lower level of recombination in L . kluyveri does not come from a corresponding decrease in initiating DSBs ( S10B Fig ) . Chromosome C stands as an outlier with significantly less COs per meiosis compared to the other chromosomes ( Fig 4A ) . This specifically comes from the low recombination activity of the 1-Mb left arm of the chromosome C ( Sakl0C-left ) where no CO was observed and only one NCO in the PAC10 gene was detected ( S6 Fig ) . De novo genome assemblies of the two parental genomes using mate pair sequencing data ( see Methods ) show that the two chromosomes C are collinear to the chromosome C of the CBS3082 reference strain ( S11 Fig ) . Consequently , the absence of CO on Sakl0C-left is not due to the presence of chromosomal rearrangements such as inversions that could prevent recombination . This low recombination activity of the Sakl0C-left region is unexpected and adds up to the other peculiarities of this genomic region . We recently performed a L . kluyveri population genomic survey and showed that this specific region is a relic of an introgression event , which occurred in the last common ancestor [23] . Our study also revealed that the Sakl0C-left region underwent a molecular evolution pattern different from the rest of the genome . It is characterized by a higher GC content , a higher sequence diversity within isolates ( π = 0 . 019 vs . 0 . 017 ) as well as a dramatically elevated rate of A:T → G:C substitutions . Therefore , the higher density of mutations on Sakl0C-left between our parental strains compared to the rest of the genome ( 1 . 13% vs . 0 . 69% ) could prevent the use of the homologous chromatid in DSB repair because of the anti-recombination activity of the mismatch repair machinery [24 , 40] . Another possibility is that fewer DSBs are induced in this region , this being potentially related to the high GC content . To test this latter possibility , we monitored the DSB level on several chromosomes including chromosome C by Southern blot after PFGE of samples from meiotic time courses using the CBS10367 isolate ( Fig 5 ) . We observed DSB hotspot signals all along chromosomes D and E . Only one centromere proximal hotspot was detected on Sakl0C-left . Finer mapping of this DSB hotspot showed it is located in the promoter region of GPI18 , matching the CO hotspot number 1 ( S12 Fig and see below ) . This result is remarkable since it reveals a 1Mb region deprived of detectable DSBs . Interestingly , even if the GPI18 hotspot is located in the Sakl0C-left , it lies close to the centromere outside but flanking the GC rich region . The next unanswered challenge is to determine what makes such a large region refractory to Spo11 cleavage . Intriguingly , we previously observed that the Sakl0C-left region has an increased ancestral recombination rate estimate ρ ( based on linkage disequilibrium decay ) compared to the rest of the genome ( ρ = 8 . 5 vs . 3 . 2 Morgans/kb , respectively ) [23] . This observation stands in contrast to the actual recombination rate we determined here . An evolution of the recombination rate over time , among other causes , could explain this discrepancy [23 , 41] . Interestingly , the MAT locus is in this region and the unusual absence of recombination revealed here reminds of its suppression associated to the MAT locus in fungi species such as Neurospora tetrasperma or Cryptococcus spp . [42 , 43] , or more generally to sex chromosomes such as the chromosome Y in mammals . The reason of this suppression is unclear but the lack of recombination might result in a long-term degeneration of genes [44] . A consequence of the absence of recombination in the Sakl0C-left is a global deficit of COs on chromosome C ( Fig 4A ) . Strikingly , 25 out of 49 tetrads show no CO on chromosome C but correct chromosome segregation , with four viable spores . In addition , more than half of the L . kluyveri tetrads on average contained another chromosome without any CO ( E0 , for 0 exchange ) while it happened in only one tetrad out of 46 in S . cerevisiae [6] ( Fig 4C and S2 Table ) . Since the presence of E0 chromosomes threatens accurate chromosome segregation at meiosis I , this result suggests either inaccurate chromosome segregation in L . kluyveri , notably for chromosome C , and therefore elevated spore death , or the existence of an efficient compensatory mechanism such as distributive segregation [45] . We could not determine the spore viability of the parental strains of our hybrid because no corresponding homozygous diploid was available . However , we determined that the spore viability of the L . kluyveri reference strain CBS3082 is close to 100% showing that at least this isolate undergoes proper homolog segregation at meiosis I . Assuming the high frequency of E0 is a general property of this species , this tends to support the existence of an efficient compensatory mechanism for meiotic chromosome segregation in L . kluyveri . By analyzing inter COs distances , we could determine if L . kluyveri COs interfere . CO interference is a phenomenon that tends to ensure an even distribution of COs along chromosomes and that may promote efficient chromosome segregation at meiosis I . This feature is not conserved across species as S . pombe does not display any CO interference . While the mechanism of CO interference is still unclear , the so-called ZMM group of proteins Zip1 , 2 , 3 , 4 , Spo16 , Msh4 , 5 and Mer3 promote CO interference in S . cerevisiae [46] . By contrast to the other Lachancea species , L . kluyveri conserved all these genes [47] , which led us to anticipate the presence of CO interference . We looked for CO interference by measuring inter-CO distances and fitting them to a gamma distribution function characterized by a shape ( k , estimating interference ) and scale ( θ , estimating the inverse of CO rate ) parameter [25 , 48 , 49] . We first confirmed that L . kluyveri displayed CO interference by comparing inter-CO distances distribution to a gamma distribution with no interference ( shape parameter ( k ) = 1 , Kolmogorov-Smirnov test P = 0 . 0072 ) . However , the optimal estimation of the strength of interference in L . kluyveri is significantly lower than that of S . cerevisiae ( k = 1 . 47 vs . 1 . 96 , Kolmogorov-Smirnov test P = 0 . 015 ) . Additionally , because the COs frequency is low , the scale parameter in L . kluyveri is higher than in S . cerevisiae ( θ = 187 vs . 61 . 7 ) ( S13A Fig ) . Remarkably , the distribution of size of LOH regions follows approximately the same gamma law ( k = 1 . 45 and θ = 213 , S13B Fig ) , which is in agreement with their meiotic origin . To identify CO hotspots ( Fig 4B and S6 Fig ) , we determined the density of COs along the genome using a 5 kb window , and defined as hotspots the regions displaying seven or more COs per 5 kb ( P < 0 . 02 based on 105 randomly generated COs positions ) . In addition to the VMA1 hotspot discussed previously , seven other CO hotspots were identified across the 49 meioses and 38 RTGs . They are located in the GPI18 , RAS1 , PIS1 , PRM2 , OST6 , VMA10 , and ALT2 orthologs ( hotspot 1 , 2 , 3 , 4 , 6 , 7 , and 8 , respectively ) ( Fig 4B ) . As mentioned previously , the GPI18 CO hotspot is on the left arm of chromosome C at approximately 9 kb from the centromere ( hotspot 1 , Fig 4B ) . This corresponds exactly to the end of the 1-Mb introgressed region ( Sakl0C-left ) and carries relics of transposable elements in both parental strains ( S3B Fig ) . Whether these transposable elements influence DSB formation or not remains to be established , since the presence of such elements is not predictive of the local recombination activity [50] . Another hypothesis is that we enriched the instance of COs happening on the chromosome C by selecting only full viable tetrads , as the absence of recombination on the Sakl0C-left region might lead to a reduction of the spore viability . In any case , this CO hotspot correlates with the DSB hotspot we observed in the promoter region of GPI18 from the CBS10367 isolate that also has relics of transposable elements ( S12 Fig ) . Among the six remaining CO hotspots , four are associated with genes with an expression level above the median during growth in rich medium ( S3 Table ) [41] . As for GPI18 , we observed corresponding DSB hotspots in the promoters of RAS1 ( chromosome E ) and PIS1 ( chromosome F ) ( S12 Fig ) . The comparison of the recombination hotspots between S . cerevisiae and L . kluyveri is statistically limited by the low recombination rate in the protoploid species . However , none of the well-known recombination hotspots in S . cerevisiae were significantly detected in L . kluyveri ( e . g . IMG1 , HIS4 , CYS3 ) [6 , 8] . In addition , among the eight hotspots detected in L . kluyveri , only one is significantly enriched for COs in S . cerevisiae considering the meiotic recombination dataset from Mancera et al . [6] . We tested the significance of recombination rate difference between the two species in a 10 kb windows ( gene-proximal COs ) around each pair of orthologs involved in a hotspot of L . kluyveri . We confirmed the difference for all the recombination hotspots with the exception of the one located in the syntenic block MSC7-VMA10-BCD1 ( Fisher exact test: P < 0 . 01 ) . At a larger scale , we compared between the two species the number of gene-proximal COs for all pairs of orthologs and observed a poor correlation ( R2 < 0 . 01 , Fig 6 ) . The VMA10 hotspot appears to be the only one conserved between the two species . Interestingly , the Saccharomyces species ( e . g . S . cerevisiae , Saccharomyces paradoxus ) , which are closely related , display similar or conserved recombination patterns , as shown by using population genomics data and comparing genome-wide recombination initiation maps [15 , 17] . We investigated a potential link between the location of the recombination events and different local properties of the genome such as GC content , gene density and marker density . No significant correlation was found . However , this might be partially due to the small number of meiotic events analyzed ( S3 Fig ) . Here , our results demonstrate that the distribution of meiotic recombination is not persistent across distantly related yeast species covering a large evolutionary scale , exceeding the one of the Saccharomyces genus . A possible factor explaining the conservation of repartition of recombination across this genus is the shared synteny and chromosomal organization .
Genome evolution is strongly tied to the meiotic behavior and recombination landscape . With the exploration of the meiotic genetic exchanges of a protoploid yeast species and its comparison to S . cerevisiae as well as other fungi ( Fig 4A ) , we unveiled characteristics that brought a new light on the variation of the recombination landscape and its potential impact on evolution . The lack of recombination hotspot conservation between L . kluyveri and S . cerevisiae might result from synteny breakage that would modify the cis interactions between recombination hotspots previously distant from one another . Short- and long-range cis interactions between recombination initiation sites have been previously observed and discussed in S . cerevisiae [51] . They include short-range competition for DSB factors [52] and Tel1-mediated DSB interference in cis over distances spanning several meiotic chromatin loops [53] . Our data could therefore support a direct link between synteny breakage on recombination hotspots evolution . However , we cannot exclude that recombination hotspots also evolve independently of synteny breakage in yeast as in Drosophila that do not have PRDM9 like elements [54] , but that would be at a much slower speed since recombination initiation hotspots are conserved over the entire Saccharomyces clade . Exploration of the recombination landscape across a larger number of yeast species spanning a broad evolutionary distance will be important to understand better the effect of synteny breakage on recombination hotspots evolution . A striking feature of L . kluyveri is the almost complete absence of recombination in a 1-Mb region of the chromosome C , resulting in the absence of CO on this whole chromosome for about half of the analyzed tetrads . We showed that this results from a lack of Spo11-induced DSB in the corresponding region . What prevents Spo11 from making DSBs over such a long region and the consequences of this meiotic defect remain to be determined . Interestingly , the absence of recombination in this introgressed region that determines mating-types has the consequence to genetically link all the genes and therefore all phenotypic characteristics their alleles might induce , which potentially hindered species diversification and adaptation . The capacity for L . kluyveri hybrids to undergo RTG during stress , resulting in a high number of LOH regions , is an unexpected source of genetic and therefore phenotypic diversity . Meiotic reversions are generally not taken into account in the exploration of species evolution . They have been mainly described in S . cerevisiae and little is known about their occurrence across any other yeast species . Moreover , it is difficult to estimate the frequency of meiotic reversions in natural conditions . For instance , LOHs detected in natural isolates could be the results of mitotic recombination or inbreeding . Here , our data show that the control of commitment into meiosis might be weaker in some yeast species , increasing allelic shuffling during stress . The significance of this evolutive advantage needs to be further investigated as it has been underestimated so far . Our unexpected observations obtained by the deep exploration of recombination landscape of a non-conventional yeast species also demonstrate that yet many mechanisms and variations affecting evolution still remain to be described and explored .
Both parental strains NBRC10955a ( MATa ) and 67–588 ( MATα ) had their CHS3 gene encoding for chitin synthase inactivated to facilitate tetrad dissection [23 , 55] . Cells were electroporated with a DNA fragment containing the drug resistance marker KanMX , produced by fusion PCR [29] . Gene replacement was validated by PCR . All primers are listed in S4 Table . Growth and cross were performed on YPD ( yeast extract 1% , peptone 2% , glucose 2% ) at 30°C . Sporulation was induced by plating cells on sporulation medium ( agar 2% , potassium acetate 1% ) and incubating 2 to 7 days at 30°C . The asci of the tetrads were digested using zymolyase ( 0 . 5 mg/mL MP Biomedicals MT ImmunO 20T ) and the spores were isolated using the MSM 400 dissection microscope ( Singer instrument ) . The whole genome of each of the 196 spores from the 49 full viable tetrads selected was sequenced . Genomic DNA was extracted using the MasterPure Yeast DNA Purification Kit ( tebu-bio ) and sequenced using Illumina HiSeq 2000 technology . We used paired-end libraries and 100 bp per read . The average reads coverage is 204x and 100x for NBRC10955a and 67–588 , respectively , and around 70x for each of the 196 spores . The reads obtained were cleaned , trimmed and aligned to the reference genome using BWA ( -n 8 -o 2 option ) . Raw data are available on the European Nucleotide Archive ( http://www . ebi . ac . uk/ena ) under accession number PRJEB13706 . To select SNPs to be used as genotyping markers , the reads of the parental strains NBRC10955a and 67–588 were mapped against the reference genome , in which repetitive and transposable elements were masked with RepeatMasker . SAMtools was used to call variants , and a set of high quality differential SNPs between parental strains was generated . To be integrated in this set , differential positions must be , for both parents , ( I ) covered between 50x and 300x , ( ii ) have a mapping quality score higher or equal to 100 , ( iii ) have an associated allele frequency ( AF tag in VCF file ) of 0 or 1 , and ( iv ) should not be located in close proximity ( 10 base pairs ) to an indel ( insertion/deletion ) . The final set consists of 58 , 256 markers used to genotype the progeny . The reads associated with each spore were mapped and variants called with the same strategy as for the parental strains . At each position of the set , a parental origin was assigned if the residue ( i ) corresponds to one parental version , ( ii ) shows a coverage higher than 25% of the average coverage of the spore ( 20x in most cases ) , ( iii ) has a mapping quality higher than 100 , and ( iv ) displays an allele frequency of 0 or 1 . If not , a NA flag was assigned to the position . From this matrix ( 58 , 256 × 196 genotypes ) , we discarded markers flagged NA across 60 spores ( 30% ) or more , markers displaying unbalanced genotype origin in the population ( fraction of parental origins above 70% / under 30% ) as well as subtelomeric regions carrying translocated and duplicated regions . Finally 56 , 612 markers corresponding to high quality differential SNPs between parents and for which most of the spores displayed a reliable allelic origin were used . The matrix of genotypes is available ( see Supplementary data ) . To identify the localization and size of recombination events , including GCs with or without COs and LOH , we adapted the CrossOver python scripts from the Recombine pipeline ( http://sourceforge . net/projects/recombine/ ) [25] to fit L . kluyveri genomic characteristics . For each tetrad , we disregarded markers for which at least one spore displayed a NA flag before running CrossOver . In order to conserve only reliable information , a GC event was considered only when at least three markers were involved , which corresponds to 63% of the detected GCs . We defined a LOH as a 4:0 GC event larger than 10 kb ( Fig 1 ) . LOH regions detected were validated in one tetrad by looking at the sequence data directly . To do so , pooled reads of all four spores from the selected tetrad were aligned to the reference genome ( BWA ) , and the polymorphic positions were identified ( SAMtools ) . Regions without LOH would present both parental allelic versions and thus appear to be heterozygous , whereas LOH regions would be homozygous . The detected homozygous regions coincide with LOH regions identified using CrossOver ( S4 Fig ) . All detected meiotic recombination events are provided in S5 Table and a global summary of the number of events detected per chromosome and per tetrad in S2 Table . The number of COs occurring within LOH regions was estimated based on the recombination rate . A total of 875 COs has been detected in the genomes of 49 meioses , excluding the Sakl0C-left region and LOH regions ( 416 . 4 Mb ) . The cumulative size of all the LOH regions was 90 . 4 Mb . With the hypothesis that the recombination rate is homogeneous , we estimated that 190 COs occurred within LOH regions of our meiotic sample . Not all of the recombination events occurring during RTG can be observed in the final tetrad as only two of the four RTG chromatids are conserved ( Fig 3 ) . A previous study estimated that the underestimation of the RTG is about 2/3 but it depends on the recombination rate as well as the chromosome size [31] . S . cerevisiae SK1 [53] and L . kluyveri CBS10367 [23] diploid strains homozygous for the sae2 null allele were used to detect the level of cut chromosomes during meiotic time courses . Procedures relative to S . cerevisae , Southern blot analysis and chromosome preparation in agarose plugs are as described in [53] . Oligonucleotide sequences for probe synthesis are in S4 Table . L . kluyveri cells were inoculated in pre-sporulation medium ( 0 . 5% yeast extract , 1% peptone , 0 . 17% yeast nitrogen base , 1% potassium acetate , 1% ammonium sulfate , 0 . 5% potassium phthalate , pH 5 . 5 ) over night at 30°C to reach an OD600 of 1 , washed once with water and resuspended in 1% potassium acetate at 30°C . Cells were harvested at different time points , washed with water and frozen . Pulse field gel electrophoreses were performed using a Bio Rad CHEF DRIII apparatus at 14°C , initial switch time 60 seconds , final switch time 120 seconds , 6 V/cm , angle 120° for 20h . SOAPdenovo ( v2 . 04 , [56] ) was used to construct de novo assemblies for both parental strains using a combination of paired-end and mate-pair Illumina reads ( European Nucleotide Archive: PRJEB13706 ) . We tested a range of k-mer size ( from 45 to 83 ) and compared the obtained assemblies at the contiguity level . We found k = 63 and k = 71 to be optimal for the 67–588 and NBRC10955 parental strains , respectively . The selected assemblies were compared to the reference genome with nucmer [57] and results were plotted with mummerplot ( S11 Fig ) . | Meiotic recombination promotes accurate chromosome segregation and genetic diversity . To date , the mechanisms and rules lying behind recombination were dissected using model organisms such as the budding yeast Saccharomyces cerevisiae . To assess the conservation and variation of this process over a broad evolutionary distance , we explored the meiotic recombination landscape in Lachancea kluyveri , a budding yeast species that diverged from S . cerevisiae more than 100 million years ago . The meiotic recombination map we generated revealed that the meiotic recombination landscape and properties significantly vary across distantly related yeast species , raising the yet to confirm possibility that recombination hotspots conservation across yeast species depends on synteny conservation . Finally , the frequent meiotic reversions we observed led us to re-evaluate their evolutionary importance . | [
"Abstract",
"Introduction",
"Results",
"and",
"discussion",
"Conclusion",
"Materials",
"and",
"methods"
] | [
"fungal",
"spores",
"meiosis",
"cell",
"cycle",
"and",
"cell",
"division",
"cell",
"processes",
"fungi",
"model",
"organisms",
"fungal",
"evolution",
"experimental",
"organism",
"systems",
"dna",
"fungal",
"reproduction",
"homologous",
"recombination",
"saccharomyces",
"research",
"and",
"analysis",
"methods",
"mycology",
"chromosome",
"biology",
"evolutionary",
"genetics",
"yeast",
"biochemistry",
"cell",
"biology",
"nucleic",
"acids",
"genetics",
"biology",
"and",
"life",
"sciences",
"yeast",
"and",
"fungal",
"models",
"saccharomyces",
"cerevisiae",
"dna",
"recombination",
"evolutionary",
"biology",
"organisms",
"chromosomes"
] | 2017 | Variation of the meiotic recombination landscape and properties over a broad evolutionary distance in yeasts |
The Southeast Asian liver fluke ( Opisthorchis viverrini ) chronically infects and affects tens of millions of people in regions of Asia , leading to chronic illness and , importantly , inducing malignant cancer ( = cholangiocarcinoma ) . In spite of this , little is known , at the molecular level , about the parasite itself , its interplay with its hosts or the mechanisms of disease and/or carcinogenesis . Here , we generated extensive RNA-Seq data ( Illumina ) representing adult and juvenile stages of O . viverrini , and combined these sequences with previously published transcriptomic data ( 454 technology ) for this species , yielding a combined assembly of significantly increased quality and allowing quantitative assessment of transcription in the juvenile and adult stage . This enhanced assembly reveals that , despite the substantial biological similarities between the human liver flukes , O . viverinni and Clonorchis sinensis , there are previously unrecognized differences in major aspects of their molecular biology . Most notable are differences among the C13 and cathepsin L-like cysteine peptidases , which play key roles in tissue migration , immune evasion and feeding , and , thus , represent potential drug and/or vaccine targets . Furthermore , these data indicate that major lineages of cysteine peptidases of socioeconomically important trematodes have evolved through a process of gene loss rather than independent radiation , contrasting previous proposals .
Parasitic worms of humans and other animals cause diseases of major socio-economic importance around the world . In spite of their significance , many of them have been substantially neglected in terms of research and their control [1] . Among parasitic flukes ( i . e . , trematodes ) , the foodborne trematodes , including the major human liver flukes , Opisthorchis viverrini and Clonorchis sinensis , are particularly understudied . In parts of Southeast Asia , including Cambodia , People's Democratic Republic of Laos , Thailand and Vietnam , O . viverrini is estimated to infect ∼9 million people [2] , with ∼67 million being at risk of infection [3] . The life-cycle of this parasite is complex , involving multiple intermediate hosts and a prey-to-predator transmission cycle [4] . Briefly , embryonated eggs are shed into the environment in the faeces from the infected definitive host ( mainly humans , dogs and cats ) . After the eggs are shed into water ( usually via untreated sewage ) , they are ingested by freshwater snails ( Bithynia spp . ) and then hatch in the gut , releasing the motile embryo ( = miracidium ) , which develops into a sporocyst . Asexual reproduction within the sporocyst gives rise to rediae and then cercariae . The motile cercariae are released from the snail into the aquatic environment . Thereafter , these larvae undergo host finding and must penetrate the skin of a cyprinoid fishes ( e . g . , Puntius spp . ) and encyst as metacercariae within the skin and/or musculature to survive . The piscivorous definitive hosts become infected upon ingestion of such fish in a raw or undercooked state [5] . Following gastric passage , the metacerariae excyst in the duodenum , and the juvenile flukes migrate via the ampulla of Vater and common bile duct to the intra-hepatic bile and/or sometimes into the pancreatic ducts ( ∼14 days ) , whereupon they develop into reproductively-active , hermaphroditic adults ( ∼4 weeks ) , which release embryonated eggs via bile/pancreatic fluid into chyme and then via host faeces into the freshwater environment , thus completing the life cycle [4] . Although infection is often asymptomatic , signs or symptoms associated with opisthorchiasis can include eosinophilia and , in intense infections , diarrhoea , epigastric pain , anorexia , pyrexia , jaundice and/or ascites [6] . Chronic opisthorchiasis often leads to cholangitis , periductal fibrosis , cholecystitis and/or cholelithiasis , and , in up to 71% of infected humans in endemic areas , can induce malignant cancer ( cholangiocarcinoma ) [7] . Hence , O . viverrini has been classified as a Group I carcinogen [8] . In endemic regions , sanitation infrastructure is often limited , and cyprinoid fish is consumed in a variety of traditional dishes . For cultural reasons , this fish is often eaten raw , and there is a resistance to recommendations to cook fish to prevent the transmission of opisthorchiasis . Therefore , the only practical measure to reduce the prevalence of cholangiocarcinoma is the treatment of O . viverrini infection with praziquantel [5] . However , the reliance on this sole treatment carries a significant risk that drug resistance will develop against this compound , as has been observed for trematocidal drugs in other flukes [9] . Clearly , understanding the intricacies of the biology of O . viverrini and opisthorchiasis is central to developing new and urgently needed intervention strategies . Yet , in spite of our knowledge of the morphological changes that occur in the parasite throughout its life cycle and its paramount importance as a carcinogen , we know very little about the molecular basis of the developmental biology of O . viverrini , its interactions with its hosts and the pathogenesis of disease , particularly carcinogenesis . The advent of next-generation sequencing and bioinformatic technologies [10] , [11] now provides unprecedented opportunities to address some of these key areas . Recently , Young and coworkers [12] characterised the transcriptome of O . viverrini using 454 sequencing technology , which provided a solid basis to explore , for the first time , the transcriptional profiles of different developmental stages of this parasite . Logically extending this work , we now characterize differential transcription between adult and juvenile stages of O . viverrini using the method of RNA-Seq ( Illumina technology ) [13] , and identify key molecules inferred to be associated with development , reproduction , feeding and survival of this neglected parasite .
Hamsters used in this study were maintained at the animal research facilities at the Faculty of Medicine , Khon Kaen University , Thailand . All work was conducted in accordance with protocols approved by the Animal Ethics Committee of Khon Kaen University ( reference number #0514 . 1 . 12 . 2/23 ) in accordance with the ‘Animal for Scientific Purposes Act’ of the National Research Council of Thailand . Using established methods [14] , metacercariae were collected from infected cyprinoid fishes from the Khon Kaen province of Thailand and used to orally infect eight helminth-free , inbred Syrian ( golden ) hamsters ( Mesocricetus auratus ) in Khon Kaen University . Hamsters were euthanized 14 and 42 days following inoculation with metacercariae , in order to collect juvenile and adult stages of O . viverrini , respectively . Worms ( all stages ) were expressed from the intra- and extra-hepatic bile ducts and cultured in vitro for 2 h , to stimulate the regurgitation of their caecal contents [12] prior to being washed extensively in physiological saline ( 25°C ) , snap-frozen in liquid nitrogen and then stored at -80°C . The specific identity of the flukes was verified by first isolating genomic DNA [15] and then carrying out PCR-coupled , bidirectional sequencing ( ABI 3730×l DNA analyzer , Applied Biosystems , California , USA ) of the second internal transcribed spacer ( ITS-2 ) of the nuclear ribosomal DNA [16] . These sequence data were compared by pairwise alignment to published sequence data in the National Center for Biotechnology Information ( NCBI ) GenBank archive ( accessible via www . ncbi . nlm . nih . gov:AY584735 ) . Total RNA was isolated each from pools of adult ( n = 15 ) or juvenile ( n = 40 ) O . viverrini using the TriPure ( Roche ) reagent [17] and treated with DNase I ( TurboDNA-free , Ambion ) according to the manufacturer's instructions . In order to control for variation in transcription due to host-related factors ( e . g . , immunological/genetic differences ) , these pools were constructed from individuals isolated from worms collected from each of the infected host animals . Total RNA concentrations were estimated spectrophotometrically , and RNA integrity was verified by agarose gel electrophoresis and using a 2100 BioAnalyzer ( Agilent ) . Polyadenylated ( polyA+ ) RNA was purified from 10 µg of total RNA using Sera-Mag oligo ( dT ) beads , fragmented to a length of 100–500 nucleotides , reverse-transcribed using random hexamers , end-repaired and adaptor-ligated , according to a recommended protocol ( Illumina ) . Ligated products of 200 bp were excised from an agarose gel , PCR-amplified for 15 cycles , purified over MinElute column ( Qiagen ) and sequenced ( single-end ) on a Genome Analyzer II ( Illumina ) . Following sequencing , the quality of all RNA-Seq data was assessed; only sequences with a PHRED score of ≥30 and a length of ≥40 nucleotides ( nt ) were retained . RNA-Seq data from the adult and juvenile libraries were combined with 454 data generated from adult O . viverrini in a previous study [12] and assembled using the program OASES v . 0 . 1 . 21 [18] . The k-mer and coverage cut-off were optimized to achieve the best assembly with the greatest mean length of contiguous sequences and the fewest incomplete transcripts . Reads predicted to represent mitochondrial , ribosomal , host or microbial sequences , based on a BLASTn comparisons with sequences in the NCBI non-redundant database ( accessible via www . ncbi . nlh . nih . gov ) , were removed from the dataset prior to subsequent analysis . Each transcript was conceptually translated in six frames , using customized PERL scripts , with the longest opening reading frame ( ORF ) being used to define the coding domain and its inferred amino acid sequence . Following data assembly , the combined 454/Illumina transcripts were subjected to analysis/annotation using an established semi-automated , bioinformatic pipeline [12] , [19] and compared ( by BLASTx analysis at an E-value cut-off of <10−05 , 10−15 and 10−30 ) with conceptually translated proteins from previously published transcriptomic data for O . viverrini [12] , C . sinensis [12] , Fasciola spp . [19] , [20] , as well as coding domains predicted from the genomes of Schistosoma spp . [17] , [21] , [22] . Functional annotation of each predicted peptide was inferred using multiple methods , including BLASTx comparison ( cut-off: E-value: <10−5 ) with the non-redundant sequence database ( March , 2012 ) available via GenBank ( NCBI; http://www . ncbi . nlm . nih . gov/est/ ) and the UniProt database [23] , prediction of conserved protein domains using InterProScan [24] , allowing assignments of parental gene ontology ( GO ) terms ( http://www . geneontology . org/ ) and homology-based mapping to conserved biochemical pathways in the Kyoto Encyclopaedia of Genes and Genomes ( KEGG ) using the KEGG orthology-based annotation system ( KOBAS ) [25] . Excretory/secretory ( ES ) proteins were predicted on the basis of the presence of a signal peptide at the N-terminus and absence of a transmembrane domain using the program PHOBIUS [26] and/or the identification of close homologues ( BLASTp analysis: E-value cutoff <10−05 ) in the signal peptide database ( SPD ) [27] or a custom-built ES database containing published proteomic data for nematodes [28] and trematodes [17] . Peptidases and their inhibitors were predicted by BLASTx comparison with the MEROPS database ( January 2011 ) [29] . The additional annotation of specific , key protein classes was achieved by BLASTx comparison with the KS-Sarfari and GPCR-Sarfari ( http://www . sarfari . org ) and the Transporter Classification ( TCDB ) [30] databases . The transcriptome assembled here using both 454 and Illumina data was compared qualitatively and quantitatively with the 454-based transcriptome published previously for the adult stage of this species [12] . Both assemblies were compared based on common assembly metrics ( i . e . , number of contigs , N50 and N90 metrics [28] , largest contigs , number of annotated proteins , mean predicted peptide length and largest predicted peptide ) using established bioinformatic approaches [28] . One-to-one orthologous transcripts ( transcript = a ‘true’ mRNA represented by the contigs constructed during the assembly of the 454 and/or 454+ Illumina data ) were identified in the ‘old’ ( i . e . , 454 ) and ‘new’ ( i . e . , 454+ Illumina ) assemblies using the reciprocal best-hit method [31] based on the BLASTn algorithm . Using this approach , we identified transcripts common to both datasets and unique to each , and then assessed the functional annotation data available for each of these transcript groups ( i . e . , common to both , unique to the ‘old’ assembly and novel to the ‘new’ assembly ) . Where possible , we attempted to assess the support for these BLASTn comparisons by mapping all Illumina reads to the contigs representing these sequences using the Burrows-Wheeler Aligner ( BWA ) program [32] . Because homopolymer errors ( i . e . , indels ) represent a known limitation of 454 sequencing [33] , and had been predicted in the previously published O . viverrini transcriptome [12] , we explored the extent to which the addition of Illumina RNA-Seq data included here was able to correct these errors and the frequency with which such repairs improved/restored the predicted ORF of each contig . To do this , we conducted reciprocal pairwise Smith Waterman alignments of each contig in each assembly using the BWA-SW command in the BWA program [32] . These pairwise alignments were filtered for the best ( i . e . , most similar ) alignment for each contig pair , and interrogated for insertions or deletions associated with homopolymers of ≥4 nucleotides ( i . e . , indels in the 454 only assembly that were corrected through the addition of the Illumina data ) . Following the optimization of the ‘new’ transcriptome for O . viverrini , we explored differential transcription between the adult and juvenile stages of this species using the RNA-Seq data for each stage . To account for large differences in the numbers of reads generated for these two libraries [34] , we randomly sub-selected sequence reads ( n = 7 , 527 , 263 ) from each library and aligned ‘adult’ and ‘juvenile’ read pools to the final transcriptome using the program SOAP2 [32] , requiring that each read mapped exclusively to one location in the transcriptome with a minimum alignment length of 40 nt and a maximum of three nucleotide mismatches per read . Relative levels of transcription in adult and juvenile stages were inferred based on the calculation of reads per kilobase per million mapped reads ( RPKM ) [35]; statistical differences in quantitative transcription between the juvenile and adult stages of O . viverrini was determined using a modified Audic-Claire equation [36] relating to a Bonferroni transformed p-value ( i . e . , False Discovery Rate ) of ≤0 . 01 and ≥2-fold absolute difference in RPKM levels , as described previously [28] . To assess the expansion of the cysteine peptide domain proteins in O . viverrini relative to the other parasitic trematodes for which extensive transcriptomic/genomic data are available , we compiled all representative sequences in the present optimized transcriptome or available from transcriptomic data for C . sinensis [12] and Fasciola spp . [19] , [20] as well as genomic data for Schistosoma spp . [17] , [21] , [22] . We extracted the nucleotide region encoding the C13 legumain-like ( PF01650 ) or C1 cathepsin-like ( PF01650 ) cysteine peptidase domains from each of these sequences , aligned these data using the program MUSCLE [37] through 50 iterations and manually verified this alignment by visual inspection in BioEdit ( http://www . mbio . ncsu . edu/bioedit/ ) . To eliminate redundancy , a complete or nearly complete transcript for each unique C13 legumain-like or C1 cathepsin-like domain detected in each alignment was retained as a representative of all sequences ( = contigs or singletons ) assembled in these datasets . To assess evolutionary relationships between and among the C13 legumain-like or C1 cathepsin-like peptidases represented in this consensus alignment , phylograms were constructed by Bayesian inference ( BI ) using the program Mr Bayes v . 3 . 1 . 2 [38] employing the Monte Carlo Markov chain method ( nchains = 4 ) over 1 , 000 , 000 tree-building generations , with every 100th tree being saved; 10% of the saved trees were discarded ( burnin = 1 , 000 trees ) to ensure stabilisation of the nodal split frequencies , and consensus trees for each peptidase family were constructed from all remaining trees , with the nodal support for each clade expressed as a posterior probability ( pp ) . The consensus trees were generated and labelled in Figtree ( http://tree . bio . ed . ac . uk/software/figtree/ ) .
In addition to refining the current assembly of the O . viverrini transcriptome , the RNA-Seq data generated here allowed , for the first time , a detailed assessment of quantitative differences in transcription between juvenile and adult stages of this species . Of the 24 , 896 contigs in the current assembly , 19 , 283 did not differ significantly in their observed levels of transcription between these two stages ( Bonferroni transformed p-value: >0 . 01; RPKM difference ≤2-fold ) , whereas 3 , 020 and 2 , 593 transcripts were differentially transcribed ( p-value≤0 . 01; RPKM difference ≥2 fold ) in juveniles and adults , respectively ( Table S3 ) . A significant percentage ( 11% ) of molecules with increased transcription in juvenile O . viverrini encoded peptides associated with energy metabolism , including oxidative phosphorylation ( e . g . , cytochrome c oxidase and NADH dehydrogenase ( ubiquinone ) 1 alpha sub-complex 1 ) , fatty acid metabolism and amino acid ( e . g . , valine , leucine , isoleucine and lysine ) degradation ( see Table S4 , S5 , S6 ) . In the juvenile stage , increased transcription was observed for both secreted and non-secreted cysteine peptidases ( e . g . , families C13 and C1A; see Table S3 and S7 ) , including molecules known to be involved in protein catabolism , proteolysis and lysosome-specific pathways in other species of fluke [43] . Of the expanded C13 legumain-like sequences predicted , six had increased transcription in the juvenile stage , as did four unique C1 cathepsin-like ( PF01650 ) cysteine peptidase domains , each of which has close homology to cathepsin L genes in Clonorchis and/or Fasciola . To support the annotation of these transcripts , as for the C13 legumains , each unique C1 cathepsin-like domain was aligned with cathepsin L-like sequences from C . sinensis [44] , Fasciola spp . [45] and Schistosoma spp . [46] , and then clustered by Bayesian inference ( see Figure 1 ) . In total , we identified six unique C1 domains using this approach . Although three additional , unique domains were identified based on homology with data in the MEROPs database , these transcripts were not sufficiently complete to allow an assessment of the entire C1 domain and thus were not considered further . Based on these analyses , three of the six putative cathepsin L-like domains clustered with the other representatives of the Opistorchiidae ( one juvenile-enriched , one constitutively transcribed and one adult-enriched ) and three with the recognized cathepsin Ls from Fasciola ( all juvenile enriched ) . Nodal support for these clusters was high in all instances ( pp values ranging from 0 . 8 to 1 . 0 ) . Significantly increased transcription in the adult stage of O . viverrini related to nucleotide processing and oocyte meiosis . For example , GO terms that were highly represented in the adult included nucleoside , nucleobase and nucleotide kinase activity as well as ribonucleotide binding ( for adenylate , nucleoside diphosphate and casein kinases ) ( see Table S4 ) . Transcription linked to nucleotide replication and processing was significantly higher in the adult compared with the juvenile stage ( Table S5 ) . In addition , chromosome–specific proteins ( e . g . , histone-lysine N-methyl transferase and chromosome-transmission fidelity protein ) , DNA repair and recombination proteins ( e . g . , meiotic recombination protein and DNA polymerases ) as well as proteins involved in the processing of mRNA ( = spliceosome ) , such as integrator complex subunit ( INTS ) and splicing factor ( e . g . , SFRS2 ) , were enriched in adult O . viverrini ( Table S3 and S6 ) . Pathways associated with the nucleotide synthesis , including purine and pyrimidine metabolism , were also enriched in this developmental stage ( Table S4 ) . Furthermore , transcription associated with oocyte meiosis was higher in the adult stage ( Table S3 and S7 ) , and included transcripts encoding adenylate cyclases , ribosomal kinases and egg-specific antigens .
Opisthorchis viverrini ( Trematoda; Platyhelminthes ) is a socioeconomically important liver fluke that affects ∼9 million people in southeast Asia and , due to poorly understood mechanisms , is one of a small number of parasitic helminths known to directly cause malignant cancer [5] . In an effort to better understand the biology of this neglected parasite , transcripts from the adult stage were sequenced by 454 technology in a previous study [12] , yielding ∼55 , 000 sequences ( ≥100 nt ) predicted to encode ∼49 , 000 peptides of ≥50 aa . Although this published dataset [12] provided significant , new insights into the transcriptome of O . viverrini , the authors acknowledged that substantial gaps remained to be addressed . Specifically , the published assembly was described as being fragmented and , due to limitations in the sequencing chemistry used , potentially contained homopolymer-associated sequencing errors [12] . In addition , because only the adult stage of O . viverrini was represented and the initial study focused on a qualitative exploration of the transcriptome , quantitative assessment of transcription during the life-cycle of this parasite was not possible at the time . In an effort to overcome these limitations and enhance the transcriptomic data available for this important parasite , we generated ∼25 million 50 bp ( single-end ) sequence reads using the Illumina platform , allowing transcription levels to be quantitated in the juvenile or adult stages of O . viverrini . An enhanced ( ‘new’ ) assembly was achieved from the combined output of the Illumina platform ( including data for each stage ) and the raw 454 data generated in the previous study [12] . Comparative analyses conducted here demonstrate that this approach improved considerably the assembly of the transcriptome . Notably , the number of contigs ( ≥100 bp ) in the ‘new’ combined assembly ( n = 24 , 869 ) was ∼31 , 000 fewer than in the previous assembly [12] , and both the N50 and N90 assembly metrics increased substantially ( by 360 and 225% respectively ) , with reductions particularly in the number of short sequences ( i . e , <200 bp: n = 11 , 177 and 5 , 175 in the ‘old’ and ‘new’ assemblies , respectively ) ( Figure 1 ) . This reduction in the number of sequences appears to relate specifically to an enhanced assembly and a reduction of redundancy overall , with 55 . 9% of the sequences and 80 . 4% of all contigs from the 454-only assembly having a close match in the new assembly . Indeed , of the remaining 24 , 330 ‘454-unique’ sequences ( contigs or singletons ) , only 1 , 480 encoded peptides of ≥50 aa with homology to proteins in other eukaryotes ( excluding likely contaminants , such as fungal or vertebrate sequences ) in a functional database ( Table S8 ) . Importantly , the mapping of the raw Illumina reads to the ‘454-unique’ sequences indicated little coverage in the present dataset ( Table S9 ) , suggesting that the transcripts represented by these sequences were of low abundance in the adult and juvenile stages from which mRNA was purified . Notably , the 454 data were generated from a normalized cDNA library , in order to specifically enrich for such lowly transcribed sequences [12] . A comprehensive analysis of the enhanced dataset indicated that the transition from the juvenile to the adult stage of O . viverrini relates primarily to a down-regulation of metabolic pathways ( likely in response to the reduced growth demands of the organism ) and , predictably , an increase in pathways associated with DNA replication and reproduction . It is likely that the increased transcription of these molecules relates to the production of eggs , sperm and embryos , a hypothesis supported by the increased transcription in adults of key meiosis-related genes such as the meisosi specific serine/threonine kinase mek1 . These findings are consistent with those reported previously for this developmental transition in other flukes , such as Schistosoma haematobium [17] . Many of the genes typically associated with highly specific reproductive functions ( i . e . , spermatogenesis ) appear to be constitutively transcribed in the juvenile and adult stages investigated here . Exceptions to this relate primarily to the transcription of genes expressed in the terminal stages of the reproductive process , such as vitelline B ( involved in egg-yolk production ) [47] and homologues of the tyrosinase genes tyr1 and tyr2 , known to be associated with late-phase egg shell synthesis in Schistosoma spp . [48] . Indeed , these genes were among the most highly transcribed genes in adult O . viverrini . All hamsters used in the present study to produce O . viverrini were helminth-free and infected at the same time point , with the juvenile worms being harvested two weeks following inoculation . The pre-patent period for O . viverrini is at least four weeks [4] , and all specimens yielded at two weeks were confirmed to be immature worms . Our data suggest that the transcription of many of the genes associated with the early phases of reproductive process ( e . g . , spermatogenesis and oogenesis ) begins long before the worm reaches adulthood . However , an alternative hypothesis is that these genes have different functional roles at different stages during the life-cycle of O . viverrini . The enhanced assembly of the transcriptome allowed greater insights into a variety of important and/or druggable groups of molecules , including receptors and transporters , kinases and peptidases . Most notably , we observed a substantial enhancement in the assembly of the complex cysteine peptidase families , including those representing C1 and C13 , which are noted for their important functional roles in many helminths [40] . In particular , we assembled 367 transcripts encoding a conserved cysteine peptidase domain based on BLASTx homology , and further characterized particular families of ES molecules based on subsequent phylogenetic analysis of known homologues from other helminths . Based on our analysis , a striking expansion of the C13 legumains in Opisthorchis relative to all other parasitic helminths for which extensive genomic and/or transcriptomic data are available was detected . This expansion appears to relate primarily to an independent lineage positioned close to , but , clearly , distinct from the nearest homologous sequence in another helminth species , being the single C13 legumain-like peptidase identified in C . sinensis , although we did detect also a close orthologue of this sequence . Interestingly , transcription data relating to these sequences suggest that all but two of the C13 legumain-like sequences detected here for O . viverrini were significantly up-regulated in the juvenile relative to the adult stage . The legumains , or asparaginyl endopeptidases ( AEPs ) , are known to cleave peptide bonds on the carboxyl-terminus of asparagine residues [49] and , in S . mansoni , have been implicated in the trans-processing and activation of cathepsin B haemoglobinase , thus enabling the degradation/digestion of host haemoglobin [50]–[52] . The asparaginyl endopeptidase sequences characterized here include typical histidine and cysteine residues , essential for enzymatic activity , suggesting that they are indeed functional . The specific divergence and radiation of these peptides in O . viverrini , and their transcription in the juvenile stage suggests an essential role during a critical phase of development . A likely hypothesis is that these molecules have radiated in O . viverrini to facilitate the exploitation of particular proteins in bile as a novel food source and/or cell detritus resulting from epithelial cell ( cholangiocyte ) turnover and/or immune cells undergoing diapedesis through the epithelium , particularly in chronically infected animals . The up-regulation of transcripts linked to these proteins is co-incidental with those associated with a variety of proteins involved specifically in metabolic pathways in the juvenile stage , which , at least circumstantially , supports this hypothesis . Also notable among the cysteine peptidases were the cathepsin L-like proteins . Close homologues of individual sequences defined previously in C . sinensis [44] were identified and clustered using phylogenetic inference . However , intriguingly , we also detected specific , close homologues in O . viverrini of genuine cathepsin Ls reported for Fasciola spp . [43] . These sequences do not segregate into clades by species , rather are positioned on the tree as monophyletic pairs ( one O . viverrini sequence grouping with one Fasciola sequence ) based on the nearest homologue in Fasciola , suggesting a common evolutionary ( i . e . , orthologous ) origin for each gene , rather than an independent radiation of this group of enzymes , as has been proposed previously for cathepsins of trematodes [43] . Despite this apparent orthologous relationship , we can only speculate as to the specific functional roles of these enzymes in O . viverrini . In Fasciola , cathepsin Ls appear to be directly involved in tissue penetration ( cathepsin L3 ) and the digestion of host proteins during migration and development of the immature stage ( cathepsins L1 and L2 ) and within the bile duct as adults ( cathepsin L5 ) [43] . To our knowledge , O . viverrini does not penetrate host tissues , and , consistent with the understanding of the cathepsin Ls in Fasciola , we find no evidence of a homologue of cathepsin L3 in Fasciola . We did detect , however , putative orthologues of cathepsins L1 and L5 . It is likely that these enzymes are critical also for feeding in Opisthorchis . For the O . viverrini cathepsin L1 orthologue , this inference is supported by differential transcription data , with the peptide being significantly up-regulated in the juvenile stage ( relative to the adult ) , consistent with the transcriptional profile known for this molecule in Fasciola spp . [43] . In contrast , in Fasciola , cathepsin L5 has been reported to be up-regulated in adults , with little evidence of transcription in the juvenile stage [43] . However , in O . viverrini , the cathepsin L5 orthologue is clearly transcribed at a higher level in the juvenile stage , suggesting a possible difference in the role of this enzyme in the latter species , despite the apparent orthologous origin . Intriguingly , although the Fasciola-like cathepsin L sequences appear to be enriched in the juvenile stage , the three Clonorchis-like cathepsin Ls , are , generally , represented by a higher level of transcription , and two of them are specifically enriched in the adult stage . Unlike Fasciola , for which juvenile stages migrate through the liver parenchyma , both the juvenile and adult stages of O . viverrini live in the bile ducts of the host . Therefore , it is possible that O . viverrini has developed ‘stage-enriched’ suites of cathepsins that allow both of these stages to fill complementary , but non-overlapping niches within the host ( e . g . , exploiting similar but distinct food-sources ) , thus reducing inter-generational competition for resources/nutrients . Clearly , an in-depth exploration of the cathepsin Ls in O . viverrini would enable better insights into the feeding mechanisms of these parasites . Our proposals could be explored at a functional level , either through gene knock-down and knockout approaches , which are already established for some parasitic flukes , including O . viverrini [53] , [54] or through the use of free-living ‘model’ flukes , such as Schmidtea mediterranea or Macrostomum ligano , in the same way that C . elegans has been employed as a ( surrogate ) functional tool for parasitic nematodes [11] . Given the utility of these model flukes as tools to study tissue and nerve regeneration [55]–[57] , ongoing efforts to sequence their genomes and transcriptomes ( of different developmental stages ) should assist in the establishment of effective functional genomic tools for trematodes [58] . The development of such tools would provide major support toward the development of novel interventions against socioeconomically important trematodiases . Through the use of Illumina-based sequencing , the present study provides a deep insight into the battery of cysteine peptidases that O . viverrini can deploy [43] , specifically in relation to the radiation of the C13 legumains and the presence of cathepsin Ls orthologous to those of Fasciola spp . These findings contrast the existing hypothesis for the evolution of the peptidases in flukes ( i . e . , family-specific radiation/evolution ) [43] . The finding that the cathepsin Ls of O . viverrini do not appear to have an homologue in C . sinensis , despite the close biological relationship shared by these species ( both being members of the Opistorchiidae ) , suggests that radiation prior to differentiation of the major trematode families , followed by subsequent gene loss , may also have shaped the cathepsins in trematodes [43] . This interpretation is further supported by the presence of cathepsin F-like proteins encoded in the adult transcriptome of F . hepatica [19] . Notably , previous knowledge and understanding of these molecules in trematodes has largely been based largely on proteomic observations , which tend to be biased toward abundantly expressed proteins [59] . Indeed , taking into account only the highly transcribed sequences in the current dataset , the results of the present study are consistent with the existing hypotheses for opistorchiids ( i . e . , a reliance on cathepsin Fs , proliferation of cathespin Bs and a single C13 legumain-like homologue ) and for the family specific radiation of these enzymes [43] . However , it is well established that one of the strengths of RNA-Seq technology is its ability to resolve both lowly transcribed sequences and alternative splicing events which are often not detectable using traditional technologies ( e . g . , microarray ) or even sensitive proteomic tools [60] . This point is clearly worthy of note , given that RNA-Seq has not be widely deployed for the characterization of other socioeconomically important flukes ( e . g . , species of Schistosoma , Fasciola and Clonorchis ) . It may well be the case that similar expansions of the cysteine peptidases and specialization relating to the exploitation of specific food sources has occurred in a range of fluke species , but has , as yet , not been resolved due to the limitations of previous technologies . Clearly , given the key functional roles that many of these molecules play in fluke biology and pathogenesis [40] , including in O . viverrini [61] , [62] , deeper exploration of fluke cathepsins using RNA-seq technology is needed . Coupling such investigations with expanded genomic sequencing of key parasites would provide much greater insight into the role that alternative splicing may play in the transcriptional biology of flukes; an area which to date has been explored only to a limited extent . | Opistorchis viverrini is an important and neglected parasite affecting ∼9 million people in South-east Asia . The parasite has a complex life-cycle which involves an intermediate phase in cyprinoid fishes . Consumption of raw or under-cooked fish infected with the metacercarial ( larval ) stage of O . viverreni results in infection , with adult worms living primarily in the intra-hepatic bile duct . In addition to the affects of the infection itself , O . viverrini is directly carcinogenic , with up to 70% of infected individuals in endemic regions developing malignant cholangiocarcinomas . Control of the parasite relies exclusively on the use of praziquantel and little is known about the mechanisms through which O . viverrini stimulates carcinogenesis . An improved understanding of the molecular biology of O . viverrini is urgently needed . In our study , we employed RNAseq technology to assess changes in gene transcription during the development of O . viverrini within the definitive host , and significantly improved the characterization of the transcriptome of this parasite . In so doing , we shed new light on the evolution of a major group proteins ( i . e . , the cysteine peptidases ) which , given their important function roles as excreted/secreted molecules , have been proposed as attractive drug/vaccine targets for a wide-range of neglected flukes , including species of Opistorchis , Clonorchis , Schistosoma and Fasciola . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"genome",
"expression",
"analysis",
"parasitic",
"diseases",
"parasitology",
"dna",
"transcription",
"neglected",
"tropical",
"diseases",
"zoology",
"infectious",
"diseases",
"gene",
"expression",
"biology",
"helminthology",
"opisthorchiasis",
"genetics",
"genomics",
"computational",
"biology",
"genetics",
"and",
"genomics"
] | 2012 | Molecular Changes in Opisthorchis viverrini (Southeast Asian Liver Fluke) during the Transition from the Juvenile to the Adult Stage |
G protein coupled receptors ( GPCRs ) allow for the transmission of signals across biological membranes . For a number of GPCRs , this signaling was shown to be coupled to prior dimerization of the receptor . The chemokine receptor type 4 ( CXCR4 ) was reported before to form dimers and their functionality was shown to depend on membrane cholesterol . Here , we address the dimerization pattern of CXCR4 in pure phospholipid bilayers and in cholesterol-rich membranes . Using ensembles of molecular dynamics simulations , we show that CXCR4 dimerizes promiscuously in phospholipid membranes . Addition of cholesterol dramatically affects the dimerization pattern: cholesterol binding largely abolishes the preferred dimer motif observed for pure phospholipid bilayers formed mainly by transmembrane helices 1 and 7 ( TM1/TM5-7 ) at the dimer interface . In turn , the symmetric TM3 , 4/TM3 , 4 interface is enabled first by intercalating cholesterol molecules . These data provide a molecular basis for the modulation of GPCR activity by its lipid environment .
Signaling across cell barriers by G protein coupled receptors ( GPCRs ) is increasingly associated with GPCR dimerization or more general oligomerization [1] . While G protein coupling was shown for many GPCRs to not necessarily depend on dimerization , a direct coupling of function with homo- or hetero-dimerization was reported for a number of GPCRs from both class A and class C [2–4] . The chemokine receptor type 4 ( CXCR4 ) is supposed to form dimers for proper biological activity [5] . This class A GPCR is responsible for directed migration of cells ( chemotaxis ) in nerve and immune cells of the human body [6 , 7] . The overexpression of CXCR4 has been shown to lead to metastasis [8] and increased levels of CXCR4 expression were in particular found in breast and lung cancer cells [9] . Moreover , CXCR4 is known to be a major receptor for the human immunodeficiency virus-1 ( HIV-1 ) upon infecting human T-cells leading to immunodeficiency and AIDS [10] . CXCR4 dimers have been suggested to be either preformed , induced , or stabilized by the interaction with its agonist CXCL12 [5 , 11–13] . Function of CXCR4 requires the presence of cholesterol [14] . Cholesterol may influence CXCR4 function by either inducing conformational changes upon binding or by cholesterol-altered CXCR4 dimerization . So far , no direct proofs for either of those mechanisms exist . For cancer cells , the chemotaxis response following CXCL12 binding to CXCR4 was abolished , inhibiting CXCR4-mediated cell migration , if the receptor homodimerization was modulated by removing cholesterol or blocked by a TM4 peptide ( an analog of the transmembrane ( TM ) helix 4 of CXCR4 ) [14 , 15] . Another study revealed that the CXCL12-induced increase of the bioluminescence resonance energy transfer ( BRET ) -signal intensity , was greatly reduced in presence of TM4 peptides in living cells , whereas TM6 and TM7 peptides showed significantly weaker influence [16] . In addition , CXCR4-mediated inhibition of cAMP production was almost abolished due to the binding of TM4 peptides . These observations hint to a mechanism where dimeric interfaces involving TM4 enable ligand-induced conformational changes which are required for CXCR4 signaling . Moreover , the dimerization of wildtype CXCR4 with a truncated form was shown to cause the WHIM ( Warts , Hypogammaglobulinemia , Infections and Myelokathexis ) syndrome [17] . Structural information on GPCR dimerization or oligomerization interfaces mainly comes from available crystal structures of class A GPCRs [18 , 19] displaying preferred interfaces under crystal conditions . Different dimer interfaces were found for various receptors: the transmembrane helix 1 ( TM1 ) and helix 8 ( H8 ) form the symmetric dimer interface of opsin [20] and of the human β2-adrenergic GPCR [21] . TM1 and TM2 as well as H8 were found at the symmetric interface of the human κ opioid receptor [22] , and TM1 , 7 and H8 for rhodopsin [23] . For the latter , atomic force microscopy data further hinted to TM4 , 5 as the physiological interface [24] . For the C-C chemokine receptor type 5 ( CCR5 ) a crystal packing with an asymmetric interface including TM1 , 7 and H8 on one monomer and TM4 , 5 for the second was obtained [25] . Some receptors even displayed different dimer configurations in different crystals . E . g . , for the β1-adrenergic GPCR both symmetric TM1 , 2 , H8 as well as TM4 , 5 interfaces were observed [26] . Also the high-resolution crystal structures of CXCR4 in complex with either a small antagonist ( IT1t ) , a cyclic antagoinst ( CVX15 ) , or a viral chemokine antagonist ( vMIP-II ) showed symmetric TM5 , 6 interfaces as the dominant dimerization interfaces [17 , 27] . However , CXCR4 dimers were likewise formed by TM1/TM5-7 interactions , or displayed TM1/TM1 as a dimerization interface ( Fig 1 ) . Another important source of information about GPCR dimerization are in silico molecular dynamics ( MD ) simulations . Coarse-grained ( CG ) simulations allow to study GPCR self-assembly on the microsecond timescale . The two commonly used techniques to obtain a statistical view on the oligomerization process include simulations of multiple receptors [28 , 29] and multiple independent parallel dimerization or oligomerization simulations [29–31] . These in silico methods have shown that individual GPCRs may interact by different interfaces and suggested for the β2-adrenergic receptor a possible modulation of the dimer interface by cholesterol [31] . The strength of GPCR dimer interfaces may be approximated by analyzing the potential of mean force [32 , 33] . This methodology has shown that both the β1- and β2-adrenergic receptors prefer the symmetric TM1 , H8 dimerization interface over the also symmetric TM3 , 4 interface [32] . Rhodopsin also mostly favors the TM1 , H8 interface , followed closely by symmetric TM5 and TM4 , 5 interfaces [33] . A coupling between GPCR function and dimer interface configuration was first shown by cross linking experiments . These revealed that the homodimer interface of the dopamine receptor ( class A GPCR ) differs considerably between active ( TM4/TM4 ) and inactive ( TM4 , 5/TM4 , 5 ) states of the receptor [34] . Similarly , the metabotropic glutamate receptor ( mGluR , class C GPCR ) adopts different dimer configurations depending on whether the receptor is in its active ( TM6/TM6 ) or inactive ( TM4 , 5/TM4 , 5 ) state [35] . Differently , the serotonine receptor ( 5HT2c , class A GPCR ) was found to form both TM1/TM1 and TM4 , 5/TM4 , 5 dimers , however , only the latter dimer was activation sensitive [36] . While evidence for a coupling of GPCR function and dimerization is increasing , it is not known whether activation selects between different pre-existing GPCR dimer configurations or whether activation drives a reorientation of the receptors in a dimer into an activation-compatible state . A couple of recent crystal structures revealed intercalating cholesterol molecules within assembled GPCR oligomers . In case of the A2A adenosine receptor , three cholesterol molecules from the upper membrane leaflet were reported to intercalate at the TM1-TM3/TM5 , 6 dimer interface . The cholesterol molecules were bound between TM2 and TM3 and closely packed around TM6 of the protomers [37] . The symmetric TM5 , 6/TM5 , 6 dimer interface of the P2Y12 receptor was stabilized by two cholesterol molecules from the lower leaflet which intercalate between TM3 and TM5 of the protomers [38] . Interestingly , the antagonist-bound , ( and thus considered inactive ) μ-opioid receptors shows a symmetric dimer interface involving TM5 , TM6 and cholesterol bound to TM6 [39] . Also for the agonist-bound receptor , in orchestra with a G protein mimetic fragment coupled to the intracellular part , a cholesterol molecule bound to TM6 . However , the dimerization via symmetric TM5 , 6 interactions was disabled due to the outwards movement of the lower part of TM6 [40] . On the other hand , the symmetric TM1 , 2 , H8/TM1 , 2 , H8 dimer ( also formed in the inactive state ) could still be observed . These observations from different GPCR dimer structures further support the hypothesis that GPCR dimer interfaces may differ between active and inactive configurations . Furthermore , cholesterol binding appears to be receptor specific and thus hard to classify in a general way . Here , we report a combined coarse-grained and atomistic simulation study of the dimerization of the chemokine receptor type 4 , showing a promiscuous , however , specific binding pattern of CXCR4 in a phospholipid bilayer . Our study further pinpoints how cholesterol binding to a specific binding site on TM1 , 7 largely suppresses the formation of a dimer interface involving TM6 , which was reported to be essential in receptor activation [41] . In contrast , cholesterol-rich domains are shown to favor a symmetrical TM3 , 4 dimer interface of CXCR4 . These findings connect the experimental observations that CXCR4 signaling is cholesterol-dependent and that the TM4 helix is a key player in building up signaling-competent homodimer interfaces in a molecular manner .
Dimerization of CXCR4 in pure POPC was observed in ≈ 50% of all simulations within 3 μs ( For a sample dimerization process see S1 Video ) . Addition of cholesterol led to a markedly decreased dimerization: Association was obtained in 29% of all simulations at 10% cholesterol concentration , and only 10% of the systems dimerized at 30% cholesterol content within 3 μs ( 145 and 53 simulations resulting in dimer formation , respectively , ( see S1 Fig ) ) . After 6 μs of simulation time , 55% and 21% of the simulation systems at 10% and 30% cholesterol content , respectively , showed dimerization . It is interesting to note that the decrease in dimerization can not be attributed to the decrease in protein mobility only: cholesterol reduced the diffusion coefficient of a monomeric receptor only by 25% and 50% ( for 10% and 30% cholesterol , respectively ) , compared to a corresponding decrease in dimerization by 42% and 79% ( see Materials and Methods ) . However , assuming a simple ‘hit-and-dimerize’ model , a reduced diffusion coefficient would lead to a proportionally decreased receptor dimerization . This finding suggests that cholesterol affects dimerization either by changing the membrane properties or by direct interaction with the receptors . For all dimerization simulations , the receptor did not substantially affect the membrane thickness in its surrounding , i . e . the lipid bilayer frustration [42] is expected to be of minor importance for the dimerization in the POPC systems ( compare S1 Table ) . A few dissociation events of CXCR4 dimers could be observed , allowing to estimate a lower bound for the binding free energy between -20 kJ/mol ( 30% cholesterol ) and -24 kJ/mol ( pure POPC , see Materials and Methods ) . Overall , an orientation analysis of the monomers in the sampled dimer structures revealed seven different dimer configurations ( Fig 2 ) . The most populated dimers in pure POPC were a compact asymmetric TM1/TM5-7 dimer ( configuration B , population 26% ) , a TM1/TM5 dimer ( C , 24% ) , a loosely bound configuration via terminal TM5/TM5 contacts ( A , 20% ) , and a symmetric TM1/TM1 dimer ( D , 13% ) . Interconversions between initially formed dimer interfaces were hardly observed with the exception of transitions between interfaces B and C ( compare S2 Table ) . The most populated dimer configuration B ( TM1/TM5-7 ) is in excellent agreement with one of the available CXCR4 dimer crystal structures ( marked as blue spots in Fig 2a , root mean square deviation , RMSD , 3 . 9 Å , for a molecular view see Fig 3 ) : The crystal structure ( pdb entry 3OE8 [17] ) has a symmetric TM5 , 6/TM5 , 6 interface in a trimeric complex , but also contains an interface formed by TM1 of one monomer and residues of TM5 , TM6 , and TM7 of the second receptor ( Figs 1b and 3a ) . Specifically , at the N-terminal side of the receptor , Lys38 ( TM1 ) and the aromatic Phe276 form a π-H pair and Ile261 ( TM6 ) interacts with the backbone of Lys38 . The central part of the interaction interface is built up by hydrophobic interactions between leucines at positions 50 ( TM1 ) and 253 ( TM6 ) . Contacts between Val54 , Leu58 , and Val62 of TM1 and Leu216 , Cys220 , Ile223 , and Ser224 of TM5 complete this dimeric CXCR4 interface at the C-terminal side . The second most populated dimer configuration ( C ) contains an asymmetric TM1/TM5 interface . A similar interface was reported for the CCR5 chemokine receptor [25] , although there TM4 contributed as well to the interface . The symmetric TM1/TM1 interface ( observed for 13% of all spontaneously formed dimers in pure POPC , D ) is also seen in the contact region between two TM5 , 6/TM5 , 6 dimers in the crystal structure ( compare Fig 1b , pdb entry 3OE9 ) . This dimer interface is mainly built by Thr43 and Ile47 of each receptor ( RMSD between crystal and simulated TM1/TM1 amounts to 4 . 1 Å , Fig 3b ) . In contrast to the well reproduced TM1/TM5-7 and TM1/TM1 crystal dimers , the rather compact TM5 , 6/TM5 , 6 crystal interface was not sampled in the simulations regardless of the cholesterol concentration . The formation of this dimer interface was found to be hindered by significant differences in the tilt of TM5 between the monomeric CXCR4 form ( tilt of upper part of TM5 , Asp193-Met205 , 28°; lower part of TM5 , Val206-Ser227 , 35° ) and the TM5 , 6/TM5 , 6 crystal dimer ( tilts of 12° and 22° , respectively ) , a result that was obtained independent of the membrane thickness ( S2 Fig , S1 Table ) . I . e . , under crystallization conditions the TM5 helix in the symmetric TM5 , 6 dimer is oriented more parallel to the ( hypothetical ) membrane normal as compared to CXCR4 monomers in a phospholipid membrane . This finding suggests that the formation of the crystal compact TM5 , 6/TM5 , 6 dimer in biological membranes requires either prior or induced receptor tilting upon dimerization or conformational receptor rearrangements resulting in an altered TM5 tilt . Additionally , atomistic simulations of the crystal TM5 , 6/TM5 , 6 dimer showed an asymmetric orientation of the TM5 helix at the interface , suggesting a reduced stability of the crystal TM5 , 6/TM5 , 6 CXCR4 dimer symmetry in a membrane environment ( S3 Fig ) . A different picture emerges for the TM1/TM5-7 and TM1/TM1 dimer configurations . Here , the TM5 helices of the simulated crystal dimers adopt similar orientations to those observed in CXCR4 monomer simulations . The presence of cholesterol in the lipid bilayer significantly affected the preference for the formation of different dimers . While the population of dimer A ( loose TM5/TM5 dimer ) was only slightly affected , the population of dimers including helices TM1 and TM5-7 ( dimers B , C , and D ) decreased significantly with increasing cholesterol concentration . At high cholesterol concentration , the formation of dimers B and D ( found in crystal structures ) is impeded resulting in relative populations of only ≈ 5% . On the other hand , in the presence of cholesterol dimers interacting via the TM3 , 4 interface ( dimers F , G ) are substantially enhanced . The abundance of the asymmetric dimer G ( TM1 , 7/TM3-5 interface ) was already enhanced at 10% cholesterol content while the occupancy of the symmetric dimer F ( TM3 , 4/TM3 , 4 ) , hardly sampled in a pure POPC membrane ( ≈ 1% ) , increased to ≈ 20% at high cholesterol concentration . These changes in the dimerization profile were largely induced by specific cholesterol-receptor interactions ( see below ) .
The function of several G protein coupled receptors was shown before to be coupled to dimerization . Recent pieces of evidence further prompted to the existence of multiple dimer interfaces , and the selective importance of individual interfaces for signaling . In addition , crystallographic studies suggest that GPCRs may adopt type-specific dimer configurations , despite their structural similarity . Here , based on ensembles of dimerization simulations , a high plasticity of the dimer interface of the chemokine receptor CXCR4 is reported . Three compact dimer structures emerged as main configurations in a cholesterol-free membrane , involving a TM1/TM5-7 interface , a TM1/TM5 interface , as well as a symmetric TM1/TM1 configuration . Interestingly , the symmetric and compact dimer including helices TM5 and partially TM6 , predominant in crystals of the receptor , could not be reproduced in the membrane environment . The reason for this difference could be ascribed to the different orientations of the TM5 helix in the membrane-embedded CXCR4 monomer and the crystal TM5 , 6/TM5 , 6 dimer . In cholesterol-depleted membranes CXCR4 is known to be inactive [15] . GPCR activation and desensitization are connected with movements of TM6 and TM7 , respectively [41] . Therefore , in the absence of cholesterol , the involvement of these helices in the dimerization interface , as reported here , is likely to hinder their activation-related movements resulting in chemotaxis-inactive dimers . This interpretation is further supported by a FRET study [15] that reported only a small reduction of FRET signals ( 15–20% ) between CXCR4 proteins upon depletion of cholesterol , while the CXCR4 chemotaxis response to CXCL12 was completely abolished . Either the binding of the CXCL12 ligand ( or its dimer ) is disabled or the G protein signaling is disrupted . Addition of cholesterol drastically affected the dimerization pattern: The cholesterol binding between TM1 and TM7 mostly disabled the formation of the TM1/TM5-7 dimer . In turn , a new cholesterol-induced symmetric TM3 , 4 dimer was enabled . This novel TM3 , 4 CXCR4 binding mode in cholesterol-rich membranes is corroborated by the experimental finding that addition of TM4 peptides weakened CXCR4 interactions and signaling in malignant cells [15]; the binding of TM4 peptides to CXCR4 likely competes with dimerization at this interface . Additional experimental support for the functional role of dimerization for signaling is provided by earlier work of Percheranicer et al . on HEK cells: Using BRET the authors demonstrated a CXCL12-induced rearrangement of CXCR4 homodimers coupled to the activation of the receptor [16] . TM6 and TM7 peptides did not affect these CXCL12-induced configurational changes , however , addition of these transmembrane peptides significantly lowered the receptor signaling . Together with our results , this experiment may be interpreted as follows: We assume a pool of different dimers , both active or activatable and inactive , formed by CXCR4 in cholesterol containing cells . Upon addition of CXCL12 the preformed TM3 , 4 homodimers are stabilized by the interaction with the ligand , shifting the pool towards this dimer configuration and resulting in the observed increased BRET intensity and receptor signaling . Addition of TM4 peptides impedes the formation of this complex resulting in drastically decreased BRET intensity and receptor signaling [15 , 16] . In turn , addition of TM6 and TM7 peptides hardly affect the formation of the activation-competent TM3 , 4 dimer but probably blocked receptor activation by binding to the activation site . Therefore , not the BRET intensity but the activity is influenced by the presence of TM6 or TM7 peptides in the experiment [16] . This in turn supports the conclusion derived above , that dimers including TM6 at the interaction interface likely represent inactive dimers . Activation-competent CXCR4 dimers are therefore suggested to necessitate binding via the TM3 , 4 interface which can only be achieved by cholesterol intercalation as shown here . To further validate these findings , a number of experiments could be performed to analyse different aspects of the discussed results . E . g . , cross-linking experiments are highly effective in identifying TM helices that contribute to dimeric interfaces and further to test dimer function . In a similar way as performed by Guo et al . [34] , cross-linking different sets of substituted cysteines on distinct TM helices will provide experimental insight in how different dimer interfaces can affect agonist-induced activation . Adding agonists to cross-linked TM1/TM5-7 dimers , should result in reduced CXCR4 signaling as compared to adding agonists to cross-linked TM3 , 4/TM3 , 4 dimers . In order to address the functionality of cholesterol-induced TM3 , 4/TM3 , 4 dimers , mutations of residues on TM3 and TM4 can be introduced that stabilize this interface in the absence of cholesterol , e . g . mutations that insert bulky and hydrophobic residues in order to pad the otherwise cholesterol-filled volume at the interface ( e . g . Leu166Phe , Thr168Trp or Ile169Phe ) . According to our observations , CXCR4 signaling is suggested to be less cholesterol-dependent for these mutants . In summary , our results suggest a highly dynamic dimerization pattern for the chemokine receptor CXCR4 , and provide evidence for a modulation of the dimerization pattern by the membrane composition . In agreement with experiments , the total dimerization is only slightly affected by cholesterol while the specificity of the interaction interface and thus probably the signaling capability are drastically altered .
The dimer crystal structure 3OE0 [17] was obtained from the Protein Data Bank and adjusted in the following way: First , all non-protein atoms , one monomer , and lysozyme ( cocrystallizing agent ) were removed . Then , the mutations of residues L125W and T240P , required for crystallization , were mutated back to represent the wild type protein . In the end , two missing intracellular loops were modelled by MODELLER [49] ( The loop comprising K67 , K68 , L69 , and R70 was not resolved in the crystal structure and in place of residues S229 and K230 lysozyme was present in the crystal structure . ) . The MODELLER output structure was energy-minimized using the GROMOS54a7 force field [50] . Because the termini were not resolved in the crystal structure , the final model contained 279 ( 25–303 ) of the 352 residues of the CXCR4 protein . The DAFT procedure involved the following steps: First , atomistic structures of CXCR4 were converted to the Martini2 . 2 CG force field [47] using martinize [47] . The secondary and tertiary structure of the protein was assured to be stable by applying a RubberBand network for all backbone bead pairs within the distance of 0 . 9 nm and excluding the i − i + 1 and i − i + 2 pairs ( which are connected by bonds and angles within the Martini force field ) . The RMSD of the TM helices of CXCR4 was as low as 1 . 5 Å ( 500 ns , CHARMM36 force field ) or even 1 . 0 Å ( Amber14sb/Lipid14 ) in atomistic simulations of CXCR4 ( see below ) . Therefore , the application of structural restraints in the coarse-grained model is not expected to influence the results on receptor dimerization . CG proteins were randomly rotated around the z-axis and placed in pairs with the minimum distance of circumscribed spheres of 3 . 5 nm into a rhombic box . In the next step a CG lipid bilayer [51] , either pure POPC or mixed POPC/cholesterol with 9:1 or 7:3 ratio , and CG water [52] were added by insane [53 , 54] . Those systems were then passed to martinate [55] where they underwent 500 steps of steepest-descent energy minimization and 10 ps of position restrained NVT simulation with a 2 fs time step . Furthermore , 100 ps of MD simulation in an NpT ensemble with a 20 fs time step were performed to heat the system to 310 K and adjust the pressure to 1 bar . Production runs succeeded in an NpT ensemble , where the temperature was kept at 310 K by applying the Berendsen thermostat [56] with a 1 ps time constant . The pressure was controlled in a semi-isotropical manner ( xy and z were independent ) to 1 bar using the Berendsen barostat [56] with a 3 ps time constant . The relative permittivity was set to be 15 and the electrostatic interactions were shifted to 0 between 0 and 1 . 2 nm . The van der Waals forces were described by the 12-6 Lennard-Jones potential that was shifted to zero between 0 . 9 and 1 . 2 nm . The integration time step was 20 fs and the movement of the center of mass of the system was removed linearly every 10 steps . Atomistic simulation systems were set up in Gromacs 5 . 0 . 4 [57] by backmapping relaxed coarse-grained systems to an atomistic resolution ( either CHARMM36 [58 , 59] or Amber14sb/ Lipid14 [60–62] force fields ) using backward [48] . In case of crystal structure simulations the corresponding monomer or dimer crystal structures were fitted on the backmapped structures and energy minimized . Before production run simulations , 2 ns simulations were performed with position restraints applied on the protein . The final simulation runs were performed in the NpT ensemble , where the temperature was kept constant at 310 K using the v-rescale thermostate [63] with a coupling time constant of 0 . 5 ps . The pressure was controlled semiisotropically to stay at 1 bar . The center of mass motion was removed for the whole simulation system . The simulation of the TM3 , 4/TM3 , 4 dimer in POPC with 30% cholesterol content was prepared in GROMACS 4 . 6 . 5 using the GROMOS54a7 [50] force field . The particle mesh Ewald summation [64] was applied to compute the long-range electrostatics . For further simulation details see Pluhackova et al [65] . A summary of all atomistic simulations is given in Table 2 . | The G protein coupled C-X-C chemokine receptor type 4 ( CXCR4 ) is a transmembrane receptor that plays an essential role in the human immune system . However , overexpression of CXCR4 has been shown to lead to metastasis of breast and lung cancer cells . Furthermore , the HI-Virus infects T-cells through binding to CXCR4 resulting in immunodeficiency and AIDS . The function of CXCR4 as well as its interaction with HIV depends on the presence of cholesterol in the cell membrane . Additionally , dimerization of CXCR4 is considered to be functionally relevant . Here , using more than a thousand of molecular dynamics simulations , we aim to connect the cholesterol-modulated function of CXCR4 with a potentially cholesterol-conditioned dimerization . Our results show that the dimerization pattern of CXCR4 is drastically altered by cholesterol . Further , we suggest that a functionally competent CXCR4 dimer is exclusively formed in the presence of cholesterol , while the preferred CXCR4 dimer interface in the absence of cholesterol is shifted to probably inactive forms . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"dimers",
"(chemical",
"physics)",
"crystal",
"structure",
"condensed",
"matter",
"physics",
"crystals",
"membrane",
"receptor",
"signaling",
"materials",
"science",
"crystallography",
"g",
"protein",
"coupled",
"receptors",
"materials",
"by",
"structure",
"physical",
"chemistry",
"chemical",
"properties",
"lipids",
"solid",
"state",
"physics",
"dimerization",
"proteins",
"chemistry",
"transmembrane",
"receptors",
"cholesterol",
"physics",
"biochemistry",
"biochemical",
"simulations",
"signal",
"transduction",
"cell",
"biology",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"computational",
"biology",
"chemical",
"physics",
"cell",
"signaling"
] | 2016 | Dynamic Cholesterol-Conditioned Dimerization of the G Protein Coupled Chemokine Receptor Type 4 |
Our understanding of the wiring map of the brain , known as the connectome , has increased greatly in the last decade , mostly due to technological advancements in neuroimaging techniques and improvements in computational tools to interpret the vast amount of available data . Despite this , with the exception of the C . elegans roundworm , no definitive connectome has been established for any species . In order to obtain this , tracer studies are particularly appealing , as these have proven highly reliable . The downside of tract tracing is that it is costly to perform , and can only be applied ex vivo . In this paper , we suggest that instead of probing all possible connections , hitherto unknown connections may be predicted from the data that is already available . Our approach uses a ‘latent space model’ that embeds the connectivity in an abstract physical space . Regions that are close in the latent space have a high chance of being connected , while regions far apart are most likely disconnected in the connectome . After learning the latent embedding from the connections that we did observe , the latent space allows us to predict connections that have not been probed previously . We apply the methodology to two connectivity data sets of the macaque , where we demonstrate that the latent space model is successful in predicting unobserved connectivity , outperforming two baselines and an alternative model in nearly all cases . Furthermore , we show how the latent spatial embedding may be used to integrate multimodal observations ( i . e . anterograde and retrograde tracers ) for the mouse neocortex . Finally , our probabilistic approach enables us to make explicit which connections are easy to predict and which prove difficult , allowing for informed follow-up studies .
Recent years have seen a surge in research effort devoted to obtaining the human connectome , a map of all the connections in the human brain at the level of macroscopic brain regions [1 , 2] . Technological advances , in particular diffusion-weighted MRI ( dMRI ) , have enabled bundles of white-matter fibers to be identified in vivo in unprecedented detail . However , dMRI suffers from a number of drawbacks [3–5] . For instance , it is an indirect measuring technique [6]: Rather than directly observing axons or large fiber bundles , these must be inferred from the diffuse movement of water molecules , using a process called tractography . In practice , the problem tractography tries to solve may be underspecified , as a single voxel may contain fibers that cross , ‘kiss’ , merge or split [7] . As a result , it may be unclear which path the estimated fibers follow . Further problems arise when interpreting the output of ( probabilistic ) tractography . The number of streamlines ( i . e . candidate fiber trajectories ) that connect two regions of interest is often used synonymously with fiber count , yet the actual number of streamlines between two regions is an intricate function of the actual fiber count and several parameters of the dMRI acquisition and tractography procedure [3 , 8] . In all , despite dMRI having greatly advanced the field of connectomics by being applicable in living human subjects , it is far from the be-all end-all solution to finding gross anatomical connectivity . Earlier approaches for studying brain connectivity [9] involve invasive techniques such as post-mortem dissection [10 , 11] as well as tract tracing in animal subjects [12] . In the latter approach a tracer ( such as a fluorescent dye or a virus ) is injected into neuronal tissue of a living animal . After appropriate waiting time , the animal is sacrificed to allow the tracer material to spread through the tissue , either in the direction from cell soma to axon terminal ( known as anterograde tracing ) , or vice versa ( retrograde tracing ) . Inspection of the virus expression or the fluorescence of the dye is subsequently used to determine to which other neuronal populations the injection site was connected [13–15] . Tract tracing has a number of advantages over dMRI-based connectivity estimation . First of all , tract tracing provides unequivocal proof that two regions are connected . In dMRI , there is always a possibility that fiber tracts follow the same highway , but do not mix . Furthermore , tract tracing can recover the direction of the tracts it recovers , something which is impossible to do with dMRI . Furthermore , the probed connections are measured directly , without the need for an additional processing step such as tractography . This results in very accurate connectivity estimates , in particular regarding long-range connections [6 , 16] and has prompted researchers to use tract tracing methods as a means to evaluate the performance of dMRI-based structural connectivity estimation [17–19] . Compared to dMRI , tract tracing is a very expensive procedure for probing connectivity [20 , 21] . It requires sacrificing animal subjects , as well as substantial manual labor in administering the tracers and processing the treated brain tissue . Through a process known as ‘link prediction’ [22–24] the number of experimental studies needed to evaluate all possible connections in a connectome may be reduced . The general idea behind this technique is that the connections that have been observed carry enough information for the missing connections to be predicted . One class of models used for making link predictions assumes that connections are the result of hidden properties of the nodes in the network ( i . e . regions of interest or neuronal populations ) [25] . For instance , stochastic block models assume the nodes of a network have latent class labels , and that the probability of a connection between two nodes depends on whether they share the same label [26] . By learning this latent structure from the data , i . e . which node has which label , new connections ( or the absence thereof ) may be predicted [27–31] . The concept of latent node classes also forms the basis of community detection [32] , for which the goal is to identify sets of nodes that have more connections among themselves than with nodes outside the set . Another latent structure approach assumes that the nodes of a network are actually embedded in an unknown physical space ( a ‘latent space’ ) [33 , 34] . When a latent space model ( LSM ) is used for link prediction , the ( Euclidean ) distance between the positions of nodes in the latent space is used to determine the likelihood of a connection . This approach is clearly applicable when networks represent geographically restricted phenomena , like traffic , power grids and the internet , but may also be used in more abstract settings , such as a social network with ties dependent on political ideology , rather than spatial location [33] . While stochastic block models have been used for modeling and prediction of links in structural connectivity [35–37] , LSMs have so far mostly been applied to social network analysis [38–40] instead of to structural connectivity . However , clearly the connectome is spatially embedded [41–43] , suggesting that the use of LSM can improve the quality of link prediction . In the current study , we describe an extended probabilistic LSM with which we embed tract-tracing connectomes into a latent space . This allows us to predict unknown connections in macaque visual cortex [44] and macaque cerebral cortex [45] . Additionally , the procedure is applied to combine anterograde and retrograde tract tracing data for the mouse neocortex [46] . While in this data set all connections have been observed , the different tracer directions disagree about connection strengths . We show that by embedding the network into a latent space , both sources of data can be explained by a single connectome . The predictive performance of the LSM is compared with two baseline models as well as with the more general latent eigenmodel , as described in [25] . Our analyses demonstrate that the LSM clearly outperforms the baseline model and slightly improves on the latent eigenmodel . The probabilistic nature of our approach provides an intuitive representation of the uncertainty in the parameters we estimate , in the form of their posterior distribution . This uncertainty may be used to determine which predicted connections are reliable and which require more data to be estimated with confidence . Finally , the spatial embedding obtained by the LSM may be interpreted to gain additional insight in the structural organization of a connectome .
The data sets used in this paper are publicly available . Surface data was available for the macaque data sets , but not for the mouse data as the node definitions for this data set are layer-specific . The properties of each of the data sets are summarized in Table 1 and discussed in detail below . The goal of our method is to predict connectivity for potential connections for which no tracer data is available , informed by the connections that do have observed data . To accomplish this , we assume that the p nodes of the network are in fact embedded in a latent space with dimensionality D , so that each node i has a latent position z i ∈ R D [33 , 34 , 38 , 39 , 41] . Furthermore , we assume that the propensity for two nodes to be connected , or the strength of such a connection , is inversely proportional to the distance lij = ‖zi − zj‖2 between the two nodes in the latent space . If no tracer data was available , the nodes are considered to be distributed uniformly within this latent space . As soon as connections between pairs of nodes become observed , this latent arrangement becomes constrained—for example , nodes that are strongly connected should be close to each other and conversely , disconnected nodes should be far apart . The higher the dimensionality of the latent space , the more complex configurations of the connectome the model can represent . For example , in a 1D model the latent positions are ordered on a line , which can host only a limited number of different connectivity structures . On the other hand , in a high-dimensional space the degrees of freedom of the model will be sufficiently high to capture a more complex network topology ( although for our purposes , a high-dimensional latent space will be prone to overfitting ) . As tracer data is typically available in a thresholded form , e . g . binary connections or ordinal connection weights , the latent space is accompanied by a set of boundaries that determine which range of distances corresponds to a particular connection weight . This idea is implemented using an ordinal regression model [25 , 47] . It defines the probability of an ordinal connection class k between nodes i and j as fijk = Φ ( i , j , k ) − Φ ( i , j , k − 1 ) , in which Φ ( i , j , k ) = ∫ - ∞ h ( i , j , k ) N ( x ∣ 0 , 1 ) d x ( 1 ) gives the cumulative density of the standard normal distribution on the interval [−∞ , h ( i , j , k ) ] . Here , h ( i , j , k ) = ( bk − lij ) /σ serves to scale and translate the Euclidean distance lij in the latent space to the intervals of the standard normal density function . Note that bk and σ are the same across all connections . The observed connection weights A = {aij} are assumed to follow a categorical distribution with probability vector fij , subject to 0 ≤ fijk ≤ 1 and ∑k fijk = 1 . If anterograde and retrograde tracer data are available separately , as in the mouse data collected by [46] , both A = {aij} and R = {rij} follow such a distribution . Importantly , once the latent positions zi and the class boundaries hijk have been learned using the available observations , the same parameters can be used to predict the class weight probabilities fij for unobserved connections , and subsequently predict the values for the missing observations in A . Thus , the latent space model as described here serves as a mechanism to ‘complete’ a partially observed connectome . The latent space model describes only symmetric connectivity behavior , as the Euclidean distance is a symmetric function . However , some part of the connectome topology may not be explained by Euclidean distance alone . For example , some nodes may be hubs—nodes which have a large number of connections compared to the rest . To allow this phenomenon in our model , we add two random vectors that model the additional likelihood of nodes having incoming and outgoing connections , using the vectors δ ∈ R p and ε ∈ R p , respectively [48] . When using these additional factors , we write instead h ( i , j , k ) = ( bk − lij + δi + εj ) /σ . The generative model complete with hyperpriors is described in detail in S1 Appendix , which also discusses constraints to make the model identifiable . Importantly , connections for which no tracer data is observed , provide no information to the model . Instead , by finding the spatial embedding of nodes within the latent space , the ordinal probabilities fij for these unknown connections may be inferred . The result is a probabilistic description of a connectome that assigns for each connection a probability for each ordinal category . To compute the posterior distribution of the latent locations and connection class probabilities , a Hamiltonian Monte Carlo sampling scheme is used [49] , which was implemented in Stan [50] . For more detail we refer the reader to S1 Appendix . The result of this procedure is a collection of samples that collectively represent the distribution over the parameters of interest , i . e . the latent positions and the unobserved connection weights . To determine the optimal dimensionality of the latent space , 10-fold cross-validation was used for each of the different data sets . For each fold , the model was trained using nine-tenth of the observed connections and evaluated using the likelihood of the remaining one-tenth . To evaluate the performance of different numbers of dimensions , the parameter D was varied in the range [1 , … , 5] . The dimensionality D ^ that resulted in the best generalizability ( i . e . highest likelihood on the withheld data ) was considered the optimal dimensionality . The model was then trained using D ^ and all available data . The performance of the predicted connectivity is evaluated using the cross-validation results . Per fold , we first compute for the tth collected sample A ( t ) = { a i j ( t ) } the absolute difference between the predicted connection weights a i j ( t ) and the observed connection weight aij , which is subsequently averaged over all connections , i . e . e abs = 1 | F | ∑ ( i , j ) ∈ F | a i j ( t ) - a i j | , ( 2 ) in which F is the set of edges ( i , j ) in that particular cross-validation fold . This error measure is in the range [0 , K − 1] . Note that eabs is a conservative measure of the performance , as Bayesian averaging would typically reduce the error . However , to be consistent with the error measures described next , we evaluate eabs per sample . Depending on the intended application of the predictions , it may be more relevant to consider only the presence or absence of a connection instead of the difference in connection weight . In other words , predicting a weak connection that should have been absent may be a more severe error than predicting a strong connection that is in fact only moderate . We therefore also compute the false positive rate efpr and false negative rate efnr , as e fpr = F P F P + T N e fnr = F N F N + T P , ( 3 ) with the terms given by F P = ∑ ( i , j ) ∈ F 1 [ a i j ( t ) > 0 ∧ a i j = 0 ] , T N = ∑ ( i , j ) ∈ F 1 [ a i j = 0 ] , F N = ∑ ( i , j ) ∈ F 1 [ a i j ( t ) = 0 ∧ a i j > 0 ] and T P = ∑ ( i , j ) ∈ F 1 [ a i j > 0 ] . For each error measure , the presented results are subsequently averaged over all samples and over the cross-validation folds . In addition to the prediction error , the probabilistic approach to latent space models allows us to compute the uncertainty that is associated with the predictions . To do so , we consider the posterior distributions over the parameters of interest . The width of these distributions provides a measure of uncertainty . If the model is certain about a prediction , the posterior will be peaked around the mode of the distribution , but if there is a lot of uncertainty the distribution will be more dispersed . To analyze how certain the predictions are , we quantify for each of the estimated parameters fijk the associated uncertainty as the 95% credible interval , i . e . the width of the range in which 95% of the posterior probability density lies . For each connection between node pairs ( i , j ) , the largest uncertainty of the K possible connection strength classes is reported as the final measure of uncertainty . Two baselines were constructed in order to interpret the performance of the latent space model . In the first , connection probabilities for a particular connection strength class k are determined by the fraction of connections in the training data having connection weight k . In this baseline , the probability vector fij is the same for all pairs ( i , j ) . In the second baseline , a zero-dimensional latent space is used . In other words , the only flexibility the model has , is in the random effects δ and ε . This baseline serves to evaluate the additional effect of the latent space , compared to the predictive performance of the degree distribution of the training data . To compare with a more flexible model , we use the latent eigenmodel that was introduced by [25] . Here , the distance lij = ‖zi − zj‖2 is replaced by l i j = - z i T Λ z j , with Λ ∈ R D × D a diagonal matrix . The elements λii may be interpreted as the relative weights of each of the latent node attributes , which no longer ( necessarily ) represent spatial position . As described in [25] , the latent eigenmodel may serve as a weak generalization of a latent distance model by setting Λ to identity . The baselines and latent eigenmodel are described in detail in S1 Appendix . For the two macaque data sets their respective cortical surfaces were available . This allowed us to determine the position of each region of interest as the center-of-gravity of the vertices in that region . This was used for what we will refer to as a fixed-position model , in which all parameters of the latent space model are learned as before , but the latent positions z are determined by the surface . For the mouse data , we were unable to find a cortical reconstruction for the parcellation used in [46] , and hence the corresponding analyses have been omitted for this data set . Visualizing the latent space embedding is not straightforward , as the different samples of the posterior may be rotated , translated and scaled relative to each other . Furthermore , the latent space is potentially high-dimensional . An approximate solution to these issues is to use ( classical ) multidimensional scaling , which visualizes the node positions in an arbitrary number of dimensions , while minimizing the error with respect to the relative distances between the nodes [51] . Throughout the remainder of this paper , we show both the anatomical positioning of nodes on the cortex , as well as the latent space embedding , on a two-dimensional plane , using this procedure . For the latent embedding , we use the posterior expectation of the distances L ^ = { l ^ i j } , with l ^ i j = 1 T ∑ t = 1 T l i j ( t ) , with T the number of samples . For the connections , we use the posterior expectation of the connectomes A ^ = { a ^ i j } , with a ^ i j = 1 T ∑ t = 1 T a i j ( t ) . Subsequent to multidimensional scaling , the latent embeddings are transformed using the Procrustean transform [38] , using the empirical positions as reference . This approach keeps the relative distances between the nodes intact , but rotates the system so that it is maximally similar to the reference . Note that this step is only applied when a reference is available , i . e . for the two macaque data sets .
Instead of predicting unobserved connections , the latent space model may also be used to integrate different modalities . Here , we combine both anterograde and retrograde tracing data collected for the mouse neocortex [46] into a unifying estimate of the underlying connectome . The model captures the remaining asymmetry using the random effects parameters δ and ε . As shown in Fig 2 , the best generalization performance is obtained using a two-dimensional latent space when using either retrograde or anterograde tracers on their own . However , once the data are combined , a four-dimensional space is optimal instead . Fig 10A shows the predicted connectome for the mouse neocortex using either a single data source with a latent space dimensionality of two , or the combined data sources with a latent space dimensionality of four . The overall structure of the matrices appears to be the same in either setting . In Fig 10 , the uncertainty that comes with the predictions is shown . The average uncertainty when using only anterograde data is 0 . 35 ( SD = 0 . 20 ) , for only retrograde data 0 . 32 ( SD = 0 . 20 ) and 0 . 29 ( SD = 0 . 19 ) when the modalities are combined , demonstrating that the resulting connectome is predicted with slightly more confidence when both modalities are used simultaneously . The histograms in Fig 10C show that the distribution of connection weights remains approximately the same for the three conditions , but are most peaked for the data fusion case . In Fig 11 , the latent embedding is shown for the two single-modality connectomes , as well as for the data fusion approach . The most clear example of a region that is connected differently according to the single-modality connectomes is the medial entorhinal cortex ( ENTm ) . It has outgoing connectivity to retrosplenial areas ( RSP ) and visual cortex ( VIS ) , but receives only a few connections itself , mainly from olfactory areas such as PIR and insular cortex ( GU ) . This causes it to be placed nearly disconnected in the retrograde tracer case . When both modalities are used , ENTm is placed closer to retrosplenial areas as it is in the anterograde tracer connectome .
We have demonstrated that connectivity can be predicted between regions for which no data has been observed directly . We accomplished this by estimating a latent space in which the nodes of the connectome are embedded . In this latent space , the distances between nodes determine the probability or strength of the corresponding connections: nodes that are close together are more likely to be connected ( or have a higher connection strength ) than nodes that are far apart . The latent space model ( LSM ) was applied to predict connectivity for two connectomes of the macaque , as well as to integrate anterograde and retrograde tracer data for the mouse neocortex . We evaluated the prediction performance by comparing the LSM with two baselines and a more general ( hence less constrained ) model known as the latent eigenmodel ( LEM ) [25] . Our results indicate that the LSM predictions were accompanied by low average errors and typically outperformed the baselines and the LEM . Although the LEM is a more general latent variable , the Euclidean distance function of the LSM appears beneficial for capturing connectivity structure . By visualizing the latent embedding of nodes in comparison to their positions on the actual cortex ( available for the two macaque data sets ) , it becomes apparent which regions have ‘surprising’ locations if we consider solely their connectivity . For the macaque visual system connectome for example , the frontal eye field and prefrontal area 46 are strongly connected to other visual areas , which causes the LSM to place them more closely to these areas than they are on the cortex . For the cerebral cortex connectome we observe the opposite; regions V6 and V6a are placed further away from early visual cortex than in actual anatomy . The fact that the number of regions that are located surprisingly in this sense is small , indicates further that connectivity and physical distance often go hand in hand . When many of the connections are actually not observed ( as for the macaque cerebral cortex connectome ) , the model may be uncertain about its predictions . As the described LSM is probabilistic , this prediction uncertainty can be made explicit . This revealed that even though on average the model performs well , depending on the data a substantial number of ( potential ) connections could not be predicted with confidence . This is due to too few data points being available for the involved regions , so that the LSM cannot fixate their positions in the latent space . The representation of uncertainty provides target areas for additional experiments with novel tracers . Those connections with maximal uncertainty should be probed first in future work , so that these connections become known , and the latent space parameters with the most degrees of freedom become anchored , which will propagate to making other predictions more certain as well . Our approach is therefore applicable both for prediction of unseen connections , as well as for guidance of optimal experimental design [52 , 53] . The prediction performance and certainty may be increased by using alternative statistical models for prediction , such as a non-parametric Gaussian process based approach [54] , or different latent variable models . However , in the current paper and the context of connectomics , the choice for a latent space model was motivated by findings that show that the probability of two regions being connected ( and the strength of this connection ) is inversely correlated with the Euclidean distance between them [42 , 43 , 55 , 56] . This made it likely that a LSM could fit well to the data and has the additional benefit of an interpretable representation , which more advanced statistical procedures may lack . We observed that this was indeed the case , and that only a small number of latent dimensions sufficed in modelling each of the considered connectomes . The results of the fixed-positions baseline ( in which the latent positions were replaced by the anatomical centers-of-gravity of the regions involved ) further corroborate that indeed much of the connectome can be explained by the physical layout of the brain alone . However , the fact that the LSM can improve upon the predictive performance of this baseline indicates that there are additional principles at play , confirming recent work on generative principles of the connectome [57 , 58] . There are a number of ways in which the link prediction approach may be extended in future work . First , following [59 , 60] , a nonparametric variant of the model may be constructed that learns the dimensionality D ^ of the latent space from the data itself . This avoids the need for a cross-validation procedure , as all the data can be used to train the model and learn D ^ simultaneously . Another extension could describe the FLNe weights in the macaque cerebral cortex data [45] as a continuous variable rather than the currently used thresholded categorization . We have refrained from this approach in order to present one model that was applicable to all three data sets , but it is to be expected that continuous weights better inform the latent distances than the ad-hoc ordinal representation of connection strengths , and subsequently increase prediction performance . A further model extension consists of adding other covariates of the regions to the model , such as cortical thickness . In [25] , a general framework for such an approach is described . A few studies are related to the presented work and have analyzed the spatial embeddedness of structural connectivity . For example , [43] use the 29 × 29 submatrix of fully observed connectivity in the macaque cerebral cortex to predict global graph-theoretical properties of the full 91 × 91 connectome . Here , it is assumed that the fully observed submatrix is representative of the entire connectome . This relates to our observation about the ( relative ) degree of regions in the data and in the predicted connectome , which have been found to be quite similar in the macaque cerebral cortex data . [61] consider a more abstract notion of topological dimensionality . Here , different nervous systems as well as computer circuits are studied and found to have a higher topological complexity than the physical embedding of the networks would suggest . As [61] argue , this implies that connectomes optimize for a trade-off between minimal wiring length and maximum topological complexity . Finally , [62] demonstrate that many large-scale features of connectivity in mouse , macaque and human can be explained by simple generative mechanisms based on spatial embedding . We have demonstrated the usage of the latent space model for prediction of connectivity on animal tracer data . The advantage of tracer data over other modalities is their reliability ( as , for example , dMRI-based structural connectivity estimates are often accompanied by uncertain estimates [63 , 64] ) , which allowed us to evaluate the performance of link prediction using the latent space model . In terms of application however , other modalities may benefit more from our approach . For example , dMRI in combination with tractography is often used to estimate structural connectivity in vivo , making it applicable to human subjects . This approach has a number of well-known shortcomings that affect the resulting connectomes [3–5] . By using the LSM approach , connections that are difficult to estimate in living human subjects may be predicted from the connections that were more easily obtained . Furthermore , recent advancements in electron microscopy imaging have enabled connectivity analysis at the single-cell resolution [65 , 66] . Here , link prediction may be used to complete the connectivity that has not yet been probed and at the same time to obtain insight in the possible spatial embedding of connectivity at this scale . For completely observed connectomes where link prediction is not directly relevant , the LSM may still be used to contrast the latent embedding with the anatomical organization of the connectome , e . g . for the recently published mouse meso-scale connectome [67] . With the increase in available connectivity data , proper statistical analyses of connectomes is paramount for furthering our understanding of the brain [66 , 68] . There are two sides to the coin of these analyses . On the one hand , latent variable models such as the one employed here and elsewhere [35 , 60 , 69] allow for a succinct description of otherwise dauntingly large amounts of data . At the same time , models that fit sufficiently well to the data can be used to predict the status of hitherto unobserved connectivity . As successful prediction is testament to a useful underlying model , the results we have shown here corroborate that spatial embedding of a connectome provides a sensible representation of the data . | Tract tracing is a highly accurate procedure for identifying animal brain connectivity . However , the technique is labor intensive and requires the sacrifice of animal subjects . In our work , we describe a computational method that is able to predict the presence or absence of unobserved connections , without having to probe these connections physically . The model works by learning for each of the nodes in the connectome its position in a latent space . Nodes that are connected according to the available data are placed close to one another , while disconnected nodes are positioned far apart . Unobserved connections may now be inferred by looking at the corresponding distance in the latent space . We apply the procedure to two data sets of the macaque brain and show that the latent space model is able to predict the strength of unknown connections . Furthermore , we use the model to integrate anterograde and retrograde data for the mouse connectome . Because the model is probabilistic , it allows us to quantify how certain we are about our predictions . This enables future research to determine which connections can confidently predicted , and which connections require further data acquisition . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] | [
"medicine",
"and",
"health",
"sciences",
"diagnostic",
"radiology",
"tractography",
"nervous",
"system",
"brain",
"vertebrates",
"neuroscience",
"animals",
"mammals",
"magnetic",
"resonance",
"imaging",
"primates",
"brain",
"morphometry",
"brain",
"mapping",
"neuroimaging",
"old",
"world",
"monkeys",
"research",
"and",
"analysis",
"methods",
"sensory",
"physiology",
"diffusion",
"magnetic",
"resonance",
"imaging",
"imaging",
"techniques",
"monkeys",
"neocortex",
"visual",
"cortex",
"macaque",
"connectomics",
"visual",
"system",
"radiology",
"and",
"imaging",
"diagnostic",
"medicine",
"neuroanatomy",
"anatomy",
"diffusion",
"tensor",
"imaging",
"physiology",
"biology",
"and",
"life",
"sciences",
"sensory",
"systems",
"amniotes",
"cerebral",
"cortex",
"organisms"
] | 2017 | The missing link: Predicting connectomes from noisy and partially observed tract tracing data |
Dengue , chikungunya , and Zika virus epidemics transmitted by Aedes aegypti mosquitoes have recently ( re ) emerged and spread throughout the Americas , Southeast Asia , the Pacific Islands , and elsewhere . Understanding how environmental conditions affect epidemic dynamics is critical for predicting and responding to the geographic and seasonal spread of disease . Specifically , we lack a mechanistic understanding of how seasonal variation in temperature affects epidemic magnitude and duration . Here , we develop a dynamic disease transmission model for dengue virus and Aedes aegypti mosquitoes that integrates mechanistic , empirically parameterized , and independently validated mosquito and virus trait thermal responses under seasonally varying temperatures . We examine the influence of seasonal temperature mean , variation , and temperature at the start of the epidemic on disease dynamics . We find that at both constant and seasonally varying temperatures , warmer temperatures at the start of epidemics promote more rapid epidemics due to faster burnout of the susceptible population . By contrast , intermediate temperatures ( 24–25°C ) at epidemic onset produced the largest epidemics in both constant and seasonally varying temperature regimes . When seasonal temperature variation was low , 25–35°C annual average temperatures produced the largest epidemics , but this range shifted to cooler temperatures as seasonal temperature variation increased ( analogous to previous results for diurnal temperature variation ) . Tropical and sub-tropical cities such as Rio de Janeiro , Fortaleza , and Salvador , Brazil; Cali , Cartagena , and Barranquilla , Colombia; Delhi , India; Guangzhou , China; and Manila , Philippines have mean annual temperatures and seasonal temperature ranges that produced the largest epidemics . However , more temperate cities like Shanghai , China had high epidemic suitability because large seasonal variation offset moderate annual average temperatures . By accounting for seasonal variation in temperature , the model provides a baseline for mechanistically understanding environmental suitability for virus transmission by Aedes aegypti . Overlaying the impact of human activities and socioeconomic factors onto this mechanistic temperature-dependent framework is critical for understanding likelihood and magnitude of outbreaks .
Over the last 30–40 years , arboviral outbreaks have dominated the public health landscape globally [1] . These viruses , most notably dengue ( DENV ) , chikungunya ( CHIKV ) , and Zika ( ZIKV ) , can cause symptoms ranging from rash , arthralgia , and fever to hemorrhagic fever ( DENV ) , long-term arthritis ( CHIKV ) , Guillain-Barré syndrome and microcephaly ( ZIKV ) [2–4] . DENV , which historically spread worldwide along shipping routes [5] , places 3 . 97 billion individuals at risk worldwide [6] and causes an estimated 390 million cases annually , including 96 million symptomatic cases [7] . CHIKV was introduced into the Americas in December 2013 after an outbreak in St . Martin Island [8] . Since then , autochthonous transmission has been reported in 45 countries [9] , and 1 . 3 billion people worldwide are at risk of contracting CHIKV [10] . More recently , the ZIKV epidemic in the Americas captured global attention after the World Health Organization ( WHO ) designated it a Public Health Emergency of International Concern in February 2016 in response to its association with neurological disorders . Following the first reported case in Brazil in May 2015 , ZIKV has spread to 48 countries and territories where it is transmitted autochthonously [11] . Because DENV , CHIKV , and ZIKV are mostly transmitted by Aedes aegypti mosquitoes , they may have similar geographic distributions and risk factors . Informed public health decisions to limit the spread and magnitude of these arboviral epidemics depend on a robust understanding of transmission dynamics . One mechanistic modeling framework , the Susceptible–Infected–Recovered ( SIR ) model , has been implemented successfully to model the dynamics of outbreaks of influenza , measles , and vector-borne diseases such as CHIKV and ZIKV [12–14] . This approach tracks virus population dynamics by compartmentalizing individuals by their state in an epidemic ( i . e . , Susceptible ( S ) , Infected ( I ) , Recovered ( R ) ) . This framework can be extended to include additional compartments , such as a latency stage , or to incorporate the dynamics of the mosquito population for vector transmission . Arbovirus dynamics are strikingly seasonal and geographically restricted to relatively warm climates [6 , 7] . This arises because several life history traits of the mosquitoes that transmit DENV , CHIKV , and ZIKV are strongly influenced by temperature and seasonality [15–22] . For simplicity , many existing models assume static life history traits [14] , and those that address seasonal forcing tend to incorporate sinusoidal variation as a single transmission parameter , β [23] . Treating seasonal temperature variation as a sinusoidal forcing function on the transmission parameter implies a monotonic relationship between temperature and transmission , such that transmission is maximized at high temperatures and decreases at low temperatures . However , decades of experimental work have demonstrated strongly nonlinear ( often unimodal ) relationships between mosquito and pathogen traits and temperature that are not well captured in a single sinusoidal forcing function [24] . Efforts by Yang et al . [25 , 26] addressed the need to include seasonal variation by adopting an SEI-SEIR compartmental framework with time-varying entomological parameters and fitting the model to DENV incidence data in Campinas , Brazil . Other previous work has integrated the effects of temperature on mosquito and parasite traits into temperature-dependent transmission models for DENV , CHIKV , and/or ZIKV , and revealing a strong , nonlinear influence of temperature with peak transmission between 29–35°C [27–34] . However , we do not yet have a mechanistic estimate for the relationship between seasonal temperature regimes and transmission potential , incorporating the full suite of transmission-relevant , nonlinear thermal responses of mosquito and parasite traits . Here , we expand on previous work with three main advances: ( 1 ) we incorporate the full suite of empirically-derived , unimodal thermal responses for all known transmission-relevant mosquito and parasite traits; ( 2 ) we examine the influence of seasonal temperature mean and variation ( in contrast to constant temperatures or daily temperature variation ) ; and ( 3 ) we use a dynamic transmission framework to explore the impact of different seasonal temperature regimes on the epidemiologically-relevant outcomes of epidemic size , duration , and peak incidence ( in contrast to R0 , or vectorial capacity , which are difficult to measure directly ) . To do so , we incorporate previously estimated and independently validated thermal response functions for all vector and parasite traits [24] into a dynamic SEI-SEIR model [25 , 26] . We explore field-relevant temperature regimes by simulating epidemics across temperature means ( 10–38°C ) and seasonal ranges ( 0–17°C ) from across the predicted suitable range for transmission . Specifically , we use the model to ask: ( 1 ) How does final epidemic size vary across constant temperatures ? ( 2 ) Under seasonally varying temperatures , how does the temperature at the start of the epidemic affect the final epidemic size and duration ? ( 3 ) How do temperature mean and seasonal range interact to determine epidemic size ? ( 4 ) Which geographic locations have high epidemic suitability based on climate ?
We first examined how epidemic dynamics varied across different constant temperatures . Here , we did not introduce seasonal forcing into the model but rather assumed static life history traits for Aedes aegypti for the simulation period . We simulated the model under default starting conditions ( see Appendix ) at four different constant temperatures: 20°C , 25°C , 30°C , and 35°C . These temperatures were chosen to span the range of temperatures at which arbovirus transmission is likely to be possible [24] . Using the model that included seasonal variation in temperature , we examined how the dynamics of an epidemic varied due to the temperature at which the epidemic began , under two temperature regimes . First , we set Tmax = 40 . 0°C , Tmean = 25 . 0°C , and Tmin = 10 . 0°C in the time-varying seasonal temperature model under default parameters ( see Appendix ) and varied the temperature at the start of the epidemic from 10 . 0°C to 40 . 0°C in increments of 0 . 1°C . We examined the response of final epidemic size , epidemic length , and maximum instantaneous number of infected individuals . We then repeated this process for a regime with a lower magnitude of seasonal temperature variation: Tmax = 30 . 0°C , Tmean = 25 . 0°C , and Tmin = 20 . 0°C . By comparing these temperature regimes , we can examine how epidemics respond to starting temperatures that are outside the range of plausible temperatures of arbovirus transmission ( regime 1 ) versus restricted to the plausible temperatures for transmission ( regime 2 ) [24] . Using the compartmental modeling framework with the default starting conditions , we examined the variation in final epidemic size as a result of seasonal forcing . To do so , we simulated over a wide range of temperature mean and seasonal variance regimes . The mean annual temperature varied from 10 . 0°C to 38 . 0°C in increments of 0 . 1°C , while the seasonal variation about the mean ( i . e . , Tmax−Tmin2 ) ranged from 0 . 0°C to 17 . 0°C in increments of 0 . 1°C . Many of these temperature regimes are unlikely to be observed empirically . However , the simulated temperature regimes spanned the full range of feasible temperature conditions . We recorded the final epidemic size , measured as the number of individuals in the recovered compartment at the end of the simulation , for each unique combination of mean annual temperature and seasonal variation . In addition , we examined the effect of epidemic starting temperature on final epidemic size across the same seasonal temperature regimes . We ran the model under default starting conditions , but allowed the starting temperature to equal Tmin , Tmean , or Tmax . To observe the interaction of population immunity with the seasonal temperature regime , we simulated the model assuming that 0 , 20 , 40 , 60 , or 80% of the population was initially immune . Each simulation began with the introduction of the infected individual occurring at the mean seasonal temperature . We then compared simulated climate regimes with actual climates in major cities , to measure relative epidemic suitability of the following cities: São Paulo , Brazil; Rio de Janeiro , Brazil; Salvador , Brazil; Fortaleza , Brazil; Belo Horizonte , Brazil; Recife , Brazil; Bogotá , Colombia; Medellín , Colombia; Cali , Colombia; Barranquilla , Colombia; Cartagena , Colombia; Tokyo , Japan; Delhi , India; Manila , Philippines; Shanghai , China; Beijing , China; New York City , USA; Guangzhou , China; Kobe , Japan; and Buenos Aires , Argentina , given 0 , 20 , 40 , 60 , and 80% population immunity . These cities were chosen because they represent some of the most populous urban areas across South America and throughout the world . To characterize uncertainty in the model , we sampled 50 joint posterior estimates for c , Tmin , and Tmax for each life history trait provided by Mordecai et al . [24] . We examined the variability in epidemic dynamics with starting temperatures under each parameterization and report the 95% credible interval for the epidemiological indices . We similarly characterize uncertainty in our estimates of the final epidemic size as a function of the seasonal temperature regime by simulating under each parameterization and reporting the 95% credible interval .
Holding temperature constant , we examined variability in epidemic dynamics across four temperatures: 20°C , 25°C , 30°C , and 35°C . As temperature increased from 20°C to 30°C , the number of susceptible individuals depleted more rapidly ( Fig 2 , SH ) . At 20°C and 35°C , the epidemics were small ( 1 . 33% and 5 . 92% of the population infected , respectively ) and burned out rapidly . Although simulations run at 25°C and 30°C produced final epidemic sizes of 94 . 73% and 99 . 98% of the population infected , respectively ( Fig 2 , RH ) , the epidemic peaked much faster at 30°C . Next , we examined variability in epidemic dynamics due to the temperature at which the epidemic began , given two seasonal temperature regimes ( 25°C mean and a seasonal range of 10°C to 40°C or 20°C to 30°C , respectively ) . Given that an epidemic occurred , epidemic length monotonically decreased as a function of starting temperature for the first temperature regime ( Fig 3A ) : warmer temperatures at the start of the epidemic produced shorter epidemics , and vice versa . In the second temperature regime , epidemic length monotonically decreased as a function of starting temperature until ~29°C . When temperature varied from 10°C to 40°C , the longest epidemic simulated was 137 . 8 days and occurred at starting temperatures of 11 . 2°C , and the shortest epidemic lasted 16 . 82 days and occurred when the temperature at the epidemic start was 35 . 7°C . When the temperature was 35 . 8°C or higher or 10 . 2°C or lower , no epidemic occurred . When temperature was constrained between 20°C and 30°C , the longest epidemic simulated was 253 . 64 days at a starting temperature of 20°C , and the shortest epidemic lasted 136 . 1 days at a starting temperature of 28 . 9°C . In contrast to epidemic length , the response of final epidemic size and maximum number of infected individuals to the temperature at epidemic onset depended on the amount of seasonal temperature variation . When temperature varied widely , from 10°C to 40°C , both final epidemic size and the maximum number of infected individuals responded unimodally to starting temperature , with peaks at 23 . 9°C and 24 . 1°C , respectively ( Fig 3A ) . By contrast , when temperature varied more narrowly from 20°C to 30°C , the final epidemic size and the maximum number of infected individuals were insensitive to starting temperature ( Fig 3B ) . Taken together , these results show that epidemics introduced at different times within identical seasonal temperature regimes can produce very similar final epidemic sizes and maximum infection rates , provided that the temperature range is sufficiently constrained . If temperature variation is large , dramatically different final epidemic sizes and maximum infection rates may result . To address how mean temperature and seasonal variance combined to influence the final epidemic size , we simulated over a wide range of temperature regimes that accounted for variation in the mean and temperature range over a calendar year . We calculated relative epidemic suitability , defined as the final epidemic size as a proportion of the human population , for twenty major cities worldwide ( Table 4 ) . In a low-variation thermal environment , a band of mean temperatures between approximately 25°C and 35°C supports the highest epidemic suitability ( Fig 4 ) . As the seasonal temperature range increases , lower mean temperatures are capable of supporting large epidemics . However , outside this narrow band of temperature regimes , epidemic suitability rapidly diminishes , and most temperature regimes did not produce epidemics . Of the focal 20 major cities , those with high mean temperature and small average temperature variation exhibited the highest epidemic suitability . For instance , Manila , Philippines , which has a monthly mean temperature of 29°C and average seasonal amplitude in mean temperature of 1 . 50°C , had an epidemic suitability of 0 . 9998 . Cartagena and Barranquilla , Colombia had epidemic suitability of 0 . 9997 . On the other hand , areas with low average temperature and greater temperature variation , such as Beijing and New York , exhibited lower—but still non-zero—epidemic suitabilities of 0 . 5268 and 0 . 04088 respectively . Notably , Guangzhou and Shanghai , China have high epidemic suitability ( 0 . 9996 and 0 . 9966 , respectively ) despite moderate mean temperatures ( 22 . 9 and 17 . 6°C , respectively ) due to high seasonal variation in temperature . By contrast , high seasonal variation reduced suitability to 0 . 9537 in Delhi , India , which has a high mean temperature of 26 . 3°C ( Fig 4 ) . The relationship between epidemic suitability and seasonal temperature regime was consistent across varying levels of population immunity . Locations with high mean temperatures and small average temperature variation had higher epidemic suitability , regardless of the level of population immunity ( S8–S10 Figs ) . However , as the level of immunity increased from 20% to 80% , the epidemic suitability at given seasonal temperature regime decreased ( Table 4 ) . Epidemic suitability also varied by starting temperature , depending on the seasonal temperature regime . The epidemic suitability of cities with high mean temperature and small average temperature variation—such as Manila , Philippines and Cartagena and Barranquilla , Colombia—did not depend on starting temperature ( Table 5 ) . However , areas with low to moderate mean temperature and large average temperature variation ( e . g . , Kobe , Japan and Shanghai , China ) exhibited low epidemic suitability ( both 0 . 0001000 ) at the minimum starting temperature and moderate-to-high epidemic suitability at the maximum starting temperature ( 0 . 6890 and 0 . 8905 , respectively ) ( Fig 5 ) . The opposite occurred in regimes with high mean temperature and large temperature variation , though these temperature regimes are rarer . Estimated epidemic suitability is close to one in the most suitable temperature regimes because we assumed that: ( i ) the population was fully susceptible at the start of the epidemic; ( ii ) mixing was homogeneous among humans and mosquitoes; ( iii ) all cases of infection are included regardless of whether or not they are symptomatic; and ( iv ) no other environmental or social drivers are limiting transmission . As a result , the epidemic suitability metric should be considered an upper bound on the proportion of the population that could become infected based on temperature alone . Final epidemic size was not sensitive to life history trait parameterization ( S8–S10 Figs ) , using samples from the posterior distribution of thermal response fits for each temperature-dependent trait . There was uncertainty in the specific numerical values of the epidemiological indices across starting temperatures ( S1 Fig ) . However , the overall functional response of the final epidemic size , maximum number of infected individuals , and the epidemic length to starting temperature was consistent across the samples from the joint posterior distribution .
Recent outbreaks of DENV , CHIKV , and ZIKV in Latin America and across the globe have captured the attention of the public health community and underscore the importance of preparation for future outbreaks . As temperatures rise , the global landscape suitable for such outbreaks will expand and shift geographically , potentially placing a larger proportion of the world’s population at risk [24 , 29 , 31] . Understanding how local temperature regimes govern epidemic dynamics is increasingly important for determining resource allocation and control interventions [41] . While previous work has investigated the effects of temperature on DENV , CHIKV , and/or ZIKV transmission , until now we have lacked comprehensive , mechanistic , and dynamic understanding of the effects of seasonally varying temperature on transmission via its ( nonlinear ) effects on mosquito and parasite traits [27–34] . With our model , which expands on [24] and [25] , we show that seasonal temperature mean and amplitude interact with the temperature at epidemic onset to shape the speed and magnitude of epidemics . At constant temperature , epidemics varied substantially in the rate at which susceptible individuals were depleted . Epidemics simulated at 25°C and 30°C reached similar sizes but the epidemic at 25°C proceeded at a much slower rate ( Fig 2 ) . This “slow burn” phenomenon occurs because slower depletion of susceptible individuals can produce epidemics of similar size to epidemics that infect people very rapidly . This phenomenon also occurs in more realistic , seasonally varying temperature regimes . The temperature at which an epidemic started affected dynamics only under large ranges of temperature variation . When temperature ranged from 10°C to 40°C , the final epidemic size peaked at intermediate starting temperatures ( 24°C; Fig 3A ) . However , in highly suitable seasonal environments , final epidemic size was large regardless of the starting temperature ( Fig 3B ) . At mean starting temperatures , epidemic suitability was sensitive to the interaction between annual temperature mean and seasonal variation . Under low seasonal temperature variation , a narrow band of annual mean temperatures ( approximately 25–35°C ) had the highest epidemic suitability ( Figs 4 & S2–S5 ) . Outside this band of temperature regimes , suitability diminishes rapidly . Larger seasonal variation in temperature lowers the range of optimal annual mean temperatures ( i . e . , suitability is high in cooler places with larger seasonal variation in temperature; Fig 4 ) . The relationship between epidemic suitability and the seasonal temperature regime also depended on the temperature at the epidemic onset . Three distinct relationships emerged ( Figs 5 & S6 and S7 ) . At intermediate annual mean temperatures of ~25–35°C and low seasonal temperature variation ( ~0–10°C ) , epidemic suitability is insensitive to starting temperature because temperature is suitable for transmission year-round . At lower annual mean temperatures ( ~10–25°C ) and higher seasonal temperature variation ( ~10–15°C ) , epidemic suitability is highest when epidemics start in moderate to warm seasons , and lower when epidemics start during cooler seasons . Finally , at high annual mean temperatures ( > 35°C ) and low seasonal temperature variation ( ~0–10°C ) , epidemic suitability is high only when epidemics start at the coldest period of the year , because otherwise the temperature is too warm for efficient transmission . The interaction between temperature mean , annual variation , and starting point sharply illustrates the unimodal effect of temperature on transmission . Models that do not include unimodal effects of temperature ( e . g . , those with sinusoidal forcing on a transmission parameter ) may fail to capture the limits on transmission in warm environments . With rising mean annual temperatures and increasing seasonal temperature variation due to climate change , the landscape of epidemic suitability is likely to shift . Importantly , areas with previously low epidemic suitability may have increasing potential for transmission year-round . By contrast , warming temperatures may drive epidemics in cities with high current suitability ( e . g . , Manila , Philippines , Barranquilla , Colombia , and Fortaleza , Brazil ) to shift toward cooler months . Thus , climate change may alter not only epidemic size and duration but also seasonal timing globally , as it interacts with other important drivers like rainfall and human behavior . It is important to note that model-estimated epidemic suitability should be treated as an upper bound on the potential for large epidemics because within highly suitable climate regimes , epidemics can vary in magnitude due to human population size and movement dynamics [28] , effective vector control , and other mitigating factors . Likewise , our estimates are conditioned on Aedes aegypti presence and virus introduction to support an outbreak . Although seasonal temperature dynamics provide insight into vector-borne transmission dynamics , other factors like mosquito abundance , vector control , and rainfall also determine transmission dynamics . Thus , temperature must be considered jointly with these factors . Moreover , accurately describing epidemic dynamics of emerging and established vector-borne pathogens will ultimately require integrating realistic models of environmental suitability , as presented here , with demographic , social , and economic factors that promote or limit disease transmission [42 , 43] . Conversely , we show that the interaction between temperature and the availability of susceptible hosts alone can drive epidemic burnout even in the absence of other limiting factors like vector control and seasonal precipitation . This suggests that correctly representing the nonlinear relationship between temperature and epidemic dynamics is critical for accurately inferring mechanistic drivers of epidemics and , in turn , predicting the efficacy of control interventions . | Mosquito-borne viruses like dengue , Zika , and chikungunya have recently caused large epidemics that are partly driven by temperature . Using a mathematical model built from laboratory experimental data for Aedes aegypti mosquitoes and dengue virus , we examine the impact of variation in seasonal temperature regimes on epidemic size and duration . At constant temperatures , both low and high temperatures ( 20°C and 35°C ) produce small epidemics , while intermediate temperatures like 25°C and 30°C produce much larger epidemics . In seasonally varying temperature environments , epidemics peak more rapidly at higher starting temperatures , while intermediate starting temperatures produce the largest epidemics . Seasonal mean temperatures of 25–35°C are most suitable for large epidemics when seasonality is low , but in more variable seasonal environments epidemic suitability peaks at lower annual average temperatures . Tropical and sub-tropical cities have the highest temperature suitability for epidemics , but more temperate cities with high seasonal variation also have the potential for very large epidemics . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"china",
"immunology",
"geographical",
"locations",
"animals",
"seasons",
"insect",
"vectors",
"infectious",
"diseases",
"south",
"america",
"aedes",
"aegypti",
"disease",
"vectors",
"brazil",
"insects",
"arthropoda",
"people",
"and",
"places",
"seasonal",
"variations",
"mosquitoes",
"eukaryota",
"colombia",
"asia",
"immunity",
"earth",
"sciences",
"biology",
"and",
"life",
"sciences",
"species",
"interactions",
"organisms"
] | 2018 | Seasonal temperature variation influences climate suitability for dengue, chikungunya, and Zika transmission |
Human genome-wide association studies have linked single nucleotide polymorphisms ( SNPs ) on chromosome 9p21 . 3 near the INK4/ARF ( CDKN2a/b ) locus with susceptibility to atherosclerotic vascular disease ( ASVD ) . Although this locus encodes three well-characterized tumor suppressors , p16INK4a , p15INK4b , and ARF , the SNPs most strongly associated with ASVD are ∼120 kb from the nearest coding gene within a long non-coding RNA ( ncRNA ) known as ANRIL ( CDKN2BAS ) . While individuals homozygous for the atherosclerotic risk allele show decreased expression of ANRIL and the coding INK4/ARF transcripts , the mechanism by which such distant genetic variants influence INK4/ARF expression is unknown . Here , using rapid amplification of cDNA ends ( RACE ) and analysis of next-generation RNA sequencing datasets , we determined the structure and abundance of multiple ANRIL species . Each of these species was present at very low copy numbers in primary and cultured cells; however , only the expression of ANRIL isoforms containing exons proximal to the INK4/ARF locus correlated with the ASVD risk alleles . Surprisingly , RACE also identified transcripts containing non-colinear ANRIL exonic sequences , whose expression also correlated with genotype and INK4/ARF expression . These non-polyadenylated RNAs resisted RNAse R digestion and could be PCR amplified using outward-facing primers , suggesting they represent circular RNA structures that could arise from by-products of mRNA splicing . Next-generation DNA sequencing and splice prediction algorithms identified polymorphisms within the ASVD risk interval that may regulate ANRIL splicing and circular ANRIL ( cANRIL ) production . These results identify novel circular RNA products emanating from the ANRIL locus and suggest causal variants at 9p21 . 3 regulate INK4/ARF expression and ASVD risk by modulating ANRIL expression and/or structure .
Atherosclerotic vascular disease ( ASVD ) is a leading cause of human mortality worldwide [1] . While there are well-recognized risk factors for ASVD such as tobacco use , obesity and hyperlipidemia , the identification of common genetic variants associated with the disease has proven difficult despite strong evidence that susceptibility is heritable . Recently , multiple unbiased genome-wide wide association studies ( GWAS ) have linked single nucleotide polymorphisms ( SNPs ) on chromosome 9p21 to ASVD and other related conditions ( i . e . coronary artery disease , stroke , myocardial infarction and aortic aneurysm ) [2]–[12] . These associations have been replicated in multiple independent studies and are not associated with “classical” ASVD risk factors such as hypertension , obesity , tobacco use or lipid levels . While causal variants within 9p21 . 3 have yet to be identified , the risk associated SNPs cluster together within a 53 kb region roughly 100 kb centromeric to the INK4/ARF ( CDKN2a/b ) tumor suppressor locus ( [13] and Figure S1 ) . Congenic mapping using mice differentially susceptible to ASVD has identified syntenic susceptibility alleles near the murine Ink4/Arf locus [14] , suggesting that this risk interval is conserved in mammals . The INK4/ARF locus encodes three archetypal tumor suppressor genes: p16INK4a , p15INK4b and ARF ( p14ARF in humans , p19ARF in mice ) , as well as a long non-coding RNA called Antisense Non-coding RNA in the INK4 Locus ( ANRIL or CDKN2BAS ) . Other than a role for ARF in development of the optic vasculature , all of the INK4/ARF proteins are thought to be largely dispensable for normal mammalian development , but play important roles in restraining aberrant proliferation associated with cancer and other disease states ( reviewed in: [15] ) . In purified T-cells from healthy individuals , we have shown a pronounced effect of ASVD-associated SNPs at 9p21 on INK4/ARF expression , with those harboring the risk alleles demonstrating reduced levels of p15INK4b , p16INK4a , ARF and ANRIL [16] . Decreased expression of such anti-proliferative molecules could promote pathologic monocytic or vascular proliferation , thus accelerating ASVD development ( reviewed in [17] , [18] ) . For example , mice lacking p16INK4a exhibit increased vascular hyperplasia following intra-arterial injury [19] and ARF deficiency has been implicated in atherosclerotic plaque formation [20] . Additionally , TGF-β signaling , which induces the expression of p16INK4a and p15INK4b , is anti-atherogenic in some settings [21]–[23] . Most recently , excess proliferation of hematopoietic progenitor cells , which is in part controlled by p16INK4a expression during aging [24] , has been associated with atherosclerosis in a murine model [25] . Moreover , targeted deletion of a region syntenic to the ASVD risk interval in mice resulted in severely attenuated expression of p15INK4b and p16INK4a [26] . Although these results suggest that the ASVD-associated 9p21 SNPs control INK4/ARF expression , and that decreased expression of the INK4/ARF tumor suppressors may promote ASVD , it is not known how polymorphisms located ∼120 kb away from the locus might influence INK4/ARF expression . ANRIL was first uncovered in a genetic analysis of familial melanoma patients with neural system tumors [27] . Based upon EST assembly , ANRIL has 19 exons with no identified open reading frame [27] ( Figure S2 ) . Although cloning a full-length version of the predicted transcript has proven difficult , a growing number of alternatively spliced ANRIL transcripts have recently been reported in the literature [28] , [29] . Many of these reports suggest that multiple ANRIL isoforms can be expressed in a single cell type . For example , two ANRIL variants have been reported in testes , five in HUVECs and three in lung [27] , [28] . Further confounding the study of ANRIL , the majority of predicted exons are <100 nucleotides in length and many consist entirely of repetitive LINE , SINE and Alu elements ( i . e . exons 7 , 8 , 12 , 14 and 16 ) [30] . Without a firm understanding of ANRIL structure , deciphering the biological function of this non-coding RNA has become increasingly complicated . More than a decade ago , elegant genetic work by Van Lohuizen , DePinho and colleagues demonstrated that the Ink4/Arf locus is potently repressed by Polycolmb group ( PcG ) complexes [31] . Such repression appears critical for the persistence and proliferation of somatic stem cells and other self-renewing tissues such as pancreatic beta-cells [32] , [33] . Until recently little progress was made in understanding the biochemical basis for PcG-mediated INK4/ARF silencing , however two independent groups have since shown that the PcG complexes , PRC-1 and PRC-2 , localize to the INK4/ARF locus and repress its activity through the establishment of repressive chromatin modifications such as H3K27 trimethylation [34]–[36] . Prior work has also shown that long , non-coding RNAs such as Xist , Kcnq1ot1 and HOTAIR can repress genes in cis- or trans- through interaction with PcG complexes [37]–[40] . Moreover , dysregulation of HOTAIR has been recently implicated in breast cancer progression , suggesting that long , non-coding RNAs play an important role in human disease [41] . Based on this evidence , we and others postulated that ANRIL could play a similar role in PcG-mediated repression of the INK4/ARF locus [16] . Against this background of prior work , we sought to better determine ANRIL structure and expression in relation to ASVD-SNP genotype and INK4/ARF expression . Toward that end , we performed comprehensive DNA and RNA analyses of ANRIL using RACE and next-generation sequencing in primary and transformed cell lines . We showed that the expression of ANRIL isoforms containing exons 1–2 or 4–6 correlated with ASVD-SNP genotype , however those containing exons 18–19 did not . Surprisingly , we also uncovered circular ANRIL ( cANRIL ) RNA species whose expression correlated with ASVD-SNP genotype . These new data suggest that the causal variant ( s ) within 9p21 influence ASVD susceptibility through regulation of ANRIL expression and splicing , leading to differential PcG recruitment and INK4/ARF repression .
Multiple ANRIL isoforms have been proposed based upon the assembly of ESTs and the sequencing of cDNA libraries ( See Figure S2 ) . To determine which of these isoforms predominates in vivo , we performed RNA ligase mediated ( RLM ) -RACE in cell lines and primary human peripheral blood T-lymphocytes ( PBTL ) . In addition to using the RLM procedure to maximize the detection of mRNA transcripts , we employed a high-fidelity Taq polymerase capable of amplifying complex DNAs such as those containing SINE , LINE and Alu elements . Primers for 3′ and 5′ RACE were designed within exons 1 , 2 , 4 , 6 , 9 , 13 , 16 and 18 of the originally reported transcript , NR_003529 [27] , but only primers in exons 4 and 6 selectively amplified ANRIL sequences ( data not shown ) . These amplicons were cloned and sequenced to verify the resulting DNA sequences . Using exon 4 and 6 primers , we identified multiple ANRIL variants , including novel splice isoforms that were not previously reported ( Figure 1A , novel exons shown in blue ) . We also detected several transcripts with a peculiar , non-colinear exon sequence , most notably in the HeLa and primary PBTL populations ( shown in red in Figure 1A ) . Given that no particular ANRIL isoform appeared to predominate from our RACE data , we exploited the observation that ANRIL RNAs containing exon 15 frequently maintained the canonical exonic structure ( e . g . 15–16–17–18–19 ) . We termed these exons as “distal” because they are located at the 3′ end of ANRIL , and those prior to exon 15 as “proximal” . To examine the absolute abundance of transcripts containing the “proximal” or “distal” exons in a variety of cell types , we developed and validated quantitative Taqman primer-probe sets spanning ANRIL exons 1–2 , 4–6 and 18–19 ( See Materials and Methods and Figure 1B ) . We also quantified the expression of other INK4/ARF transcripts ( e . g . p15INK4b , p16INK4a and ARF ) since we previously observed a correlation between the expression of ANRIL4-6 and p16INK4a in PBTLs ( r2 = 0 . 32 , p<0 . 0001; [42] ) . Importantly , none of these Taqman strategies amplify ANRIL exons containing repetitive sequences . Moreover , the products from these assays were gel-purified , cloned and sequenced to verify their identities . Equal amounts of RNA from a panel of 27 primary and transformed cell lines were subjected to reverse transcription using a combination of random hexamer and oligo dT primers . Deletion or methylation of the 9p21 tumor suppressors is a common event in cancer and was observed in 59% of our cell line panel ( Figure S4 ) . Even after exclusion of these lines from our analyses , all three ANRIL species ( exon 1–2 , 4–6 and 18–19 ) were rare: based on an estimate of 15 pg of total RNA per cell , expression of ANRIL species ranged between 0 . 01 to 1 copy per cell ( Figure 1B ) . For comparison , INK4/ARF expression is considered quite low in primary cells [43] , [44] , yet p16INK4a and p15INK4b were more abundant than any ANRIL species in non-transformed Hs68 , HUVEC and IMR90 cell lines ( on average 155 . 9±18 . 9- and 14 . 5±5 . 5-fold , respectively; See Figure S4 ) . These data indicate that ANRIL species are readily detectable but non-abundant in a wide variety of cell types . To further characterize 9p21 transcription in a manner unbiased by PCR primer choice , we analyzed publically available , next-generation RNA sequencing ( RNA-seq ) datasets of oligo dT-primed and reverse transcribed RNA from primary human brain or HeLa cells . ( [45]; Genbank Short Read Archive Study ID: SRP002274 and ENCODE RNA-seq replicates , respectively ) . These datasets were solely chosen based upon the high-depth of sequencing provided ( >368 million and >110 million reads , respectively ) , allowing for the study of non-abundant ncRNAs . In accord with the TaqMan-based analyses , reads mapping to ANRIL were rare in both the human brain and HeLa datasets ( Figure 2 ) . In fact , the peak height of reads mapping to p16INK4a was >40 times higher than that of any ANRIL exon , even though brain expresses low levels of p16INK4a [44] , [46] . The level of ANRIL expression detected in both datasets ( maximal peak height of 12 ) was comparable to that of other long , non-coding RNAs including HOTAIR and Kcnq1ot1 ( maximal peak heights of 10 and 18 , respectively ) . Intriguingly , by both TaqMan ( Figure 1B ) and RNA-seq ( Figure 2B ) , we observed a disparity in the number of molecules of each ANRIL exon detected , with those containing the ‘central’ ANRIL ( 4–12 ) exons being the least abundant when averaged across the entire locus . The most prevalent ANRIL peaks localized to the 5′ and 3′ ends of the transcript ( i . e . exons 1–3 and 13–19 ) . The relative excess of the more distal exons might be a consequence of 9p21 deletion , as we observed splicing between MTAP , a gene 100 kb telomeric to the INK4/ARF locus , and the 3′ end of ANRIL in RNA-seq datasets from cell lines with INK4/ARF loss ( i . e . SUM102 and MCF7 , data not shown ) . Although such MTAP-ANRIL fusions have been previously described [47] , using a highly sensitive Taqman strategy we only detected these fusions in cell lines with 9p21 deletion , and not in any primary cells or lines with two copies of an intact INK4/ARF locus ( data not shown ) . Therefore , while exon deletions and/or MTAP splicing explain the relative excess transcription of distal ANRIL exons in some cancer cell lines harboring 9p21 deletion , these mechanisms do not explain the decreased transcription of exons 4–12 versus exons 1–3 , in any cell line . Likewise , somatic deletions and MTAP splicing do not explain the uniform decrease of the central exons compared to the proximal and distal exons in cultures of primary cells ( Figure S4 ) . Together , these RNA-seq and TaqMan analyses indicate that ANRIL is a rare , multi-variant RNA species in which transcripts containing exons 1–3 or 13–19 predominate over those containing exons 4–12 . Based on these findings as well as the observation of non-colinear RNA species by RACE ( Figure 1A ) , we hypothesized that the reduced level of the central exons ( 4–12 ) in mature , polyadenylated RNA may be the result of alternative splicing events in which the central exons of ANRIL are frequently skipped . We observed non-colinear RNA species in RACE analysis of several cell lines and primary cells ( Figure 1A ) . We considered the possibility that this represented alterations of the germline DNA sequence ( e . g . duplications or transversions ) but found no evidence for this using BreakDancer [48] to analyze next-generation DNA sequencing from 10 individuals . Given that cryptic germline DNA alterations did not appear to explain these non-colinear forms , we considered that non-canonical RNA splicing events such as trans-splicing or re-splicing of RNA lariats might occur as part of ANRIL processing . In the latter case , the splice sites of certain consecutive exons are not recognized by the splicing machinery , resulting in the inclusion of exonic sequences within the RNA lariat [49] . Internal splicing of this exon-containing lariat structure may then occur , leading to the production of exon-only RNA circles in which non-consecutive exon junctions are generated by splicing across the branch site ( e . g . exon 13–14–4–5 ) . To determine the structure of such non-colinear ANRIL RNAs , we designed and validated a Taqman strategy using outward facing primers to detect transcripts containing the non-colinear exon 14-5 junction ( Figure 3A ) . This detection scheme was chosen based upon preliminary PCR studies wherein amplification of the 14-5 splice junction predominated over that of the exon 14-4 junction . As lariat or circular RNAs would not be polyadenylated , we assayed the amount of each ANRIL species detected in RNA samples reverse transcribed with random hexamers ( HEX ) , oligo dT ( dT ) or both ( Figure 3B ) . As would be expected for polyadenylated RNAs , conversion of ANRIL1-2 , ANRIL18-19 , p16INK4a and p15INK4b into cDNA was efficiently accomplished with either oligo dT or HEX primers alone . Conversely , ANRIL4-6 and 14-5 were effectively primed with HEX but not oligo dT , confirming that these transcripts were not polyadenylated . To further examine the structure of these non-canonical transcripts , we determined transcript sensitivity to the RNA exonuclease , RNAse R . RNAse R specifically digests both structured and non-structured linear RNAs , but spares RNA circles and lariats [50] . Using equal amounts of RNA from normal and immortalized human fibroblasts ( IMR90 and Hs68 , respectively ) treated with or without RNAse R , we generated cDNA and conducted TaqMan analysis for ANRIL expression . As expected for linear species , RNAse R treatment caused a marked reduction in the number of p16INK4a and p15INK4b transcripts detected ( 4 . 5- and 8 . 4-fold decrease , respectively ) , demonstrating that these coding transcripts are predominantly linear ( Figure 3C and 3D ) . In contrast , we observed a 6-fold enrichment of ANRIL14-5 molecules with RNAse R treatment , confirming their circular nature . We were surprised to find that ANRIL4-6 expression also exhibited RNAse R-dependent enrichment in both cell lines . ANRIL1-2 levels decreased consistent with a predominantly linear species , and ANRIL18-19 demonstrated an intermediate behavior consistent with a mix of linear and circular forms ( Figure 3C and 3D ) . Together , these data provide evidence that ANRIL4-6 and 14-5 are predominantly contained within non-polyadenylated , circular ( or lariat ) ANRIL ( cANRIL ) transcripts . Based upon analysis of non-sequential cDNA and EST sequences , stable RNA circles have been hypothesized to represent <1% of the total transciptome [51] . To determine the ubiquity of cANRIL RNAs , we employed the Taqman-based strategy shown in Figure 3A . Using this method , we observed amplification of ANRIL14-5 in 16 of 20 ANRIL expressing primary cultures and cell lines ( Figure S5 ) . The finding of cANRIL in diverse cell types is perhaps surprising given the low abundance of ANRIL transcripts and that RNA lariats are usually unstable intermediates which undergo rapid degradation . These data suggest that cANRIL is a stable , naturally occurring , circular RNA species produced in most INK4/ARF expressing cells . To better define the structure of cANRIL RNAs , we used outward-facing primers located in five ANRIL exons to amplify cDNA produced using a mix of random hexamers and oligo dT ( Figure 4A ) . In Hs68 and IM90 cells , PCR products specific to 9p21 were detected using primers sets targeting exons 4 and 6 . These products were more efficiently amplified using cDNAs from RNase R-treated samples , suggesting that they arose from circular RNAs . In contrast , PCR products specific to chromosome 9 were not detected using primers internal to exons 1 and 18 , and only detected in one instance with primers internal to exon 16 , suggesting these exons were not commonly included in stable , circular ANRIL species . To determine the contents of the circular transcripts containing ANRIL exons 4 and 6 , we cloned and sequenced the observed PCR products . The resulting sequences predominantly included ANRIL exons 4 to14 . Several exons were never observed in the looped structures ( i . e . exons 1 , 2 , 3 , 8 , 9 , 11 , and 12 ) ( Figure 4B and 4C ) . Novel sequences were also discovered within the products , most notably , regions intronic to the previously described ANRIL transcript ( i . e . parts of intron 3 ) . These data explain the paucity of expression of ANRIL exons 4 to 14 as determined by Taqman and RNA-seq ( Figure 1B and Figure 2B ) , suggesting that ANRIL processing may involve exon skipping events which preferentially incorporate the central exons into lariats which may then be internally spliced to form cANRIL . We have previously shown [42] a correlation between expression of the coding INK4/ARF transcripts and ANRIL4-6 . We now recognize that the latter is largely contained in circular RNA species ( Figure 3B and 3C ) . Next , we investigated correlations between other , newly identified ANRIL isoforms and the coding INK4/ARF transcripts ( Figure 5A and Figure S6 ) . We performed Taqman-based gene expression analyses in 106 primary human peripheral blood T-lymphocyte ( PBTL ) samples from two previously published cohorts [16] , [42] . Remarkably , cANRIL expression ( ANRIL14-5 and ANRIL4-6 ) was observed in all PBTL samples . Expression of ANRIL18-19 did not correlate with any other ANRIL or INK4/ARF transcript ( Figure 5A ) . In contrast , a strong correlation was observed between the expression of ANRIL4-6 and all of the INK4/ARF tumor suppressors ( p<1*10−15 for all pair-wise comparisons ) . Significant associations ( p<0 . 05 ) were also observed between ANRIL1-2 and ARF , and between ANRIL14-5 and p16INK4a . We have also demonstrated an effect of a replicated SNP associated with atherosclerosis ( rs10757278 ) on the expression of p15INK4b , p16INK4a , ARF and ANRIL4-6 in PBTLs [16] . We therefore examined whether expression of other ANRIL isoforms correlated with ASVD SNP genotype ( Figure 5B ) . While no significant correlation between ANRIL18-19 with SNP genotype was noted , expression of ANRIL1-2 and ANRIL14-5 showed significantly decreased expression in individuals harboring the risk ( G ) allele ( p<0 . 0001 and p<0 . 04 , respectively ) . In aggregate , these data demonstrate that expression of both circular and linear transcripts containing the proximal but not distal ANRIL exons correlates with ASVD-SNP genotype and expression of the coding INK4/ARF transcripts . We performed next-generation DNA sequencing of the ASVD risk interval to identify polymorphisms that might influence ANRIL expression or splicing . To enhance detection of rare variants , we pooled DNA from 5 individuals of Asian and European descent who were homozygous for either the A or G allele of rs10757278 . To increase the chance that the analyzed DNAs would harbor causal variants that influence INK4/ARF expression , we selected individuals based on their age-adjusted expression of p16INK4a ( i . e . AA donors with higher than average expression , GG donors with lower than average expression [16] ) . We performed sequence capture using an Illumina tiling array on the AA vs . GG pooled samples . The tiling array was designed to bind all non-repetitive human chromosome 9 regions from 22 , 054 , 888 to 22 , 134 , 171 bp , a region chosen as it contains the previously identified ASVD and type 2 diabetes mellitus risk intervals ( Figure 6A ) . The 9p21 enriched DNA was sequenced using both Roche 454 ( 400 bp reads ) and Illumina GAII ( 35 bp reads ) technologies . The resulting sequences were aligned to the UCSC reference genome ( hg18 ) using three separate alignment algorithms: gsMapper , BWA and SOAP [52] , [53] . As shown in Figure 6B , SOAP identified the highest number of polymorphisms within each pooled sample; however , many of these were not recapitulated using other mapping techniques . Thus , for further analysis we focused our efforts on polymorphisms which were identified by at least two mapping algorithms . Employing these criteria , we discovered 101 SNPs that differed from the reference genome within our samples: 11 in the AA pool , 64 in the GG pool , and 26 in both ( Figure 6C ) . As expected , more differences from the reference genome were identified in the GG sample , as the reference genome harbors the A allele , and most SNPs in the captured region are in moderate to strong linkage disequilibrium with rs10757278 ( Figure S1 ) . Therefore , using next generation sequencing of captured DNA from 10 informative individuals biased to harbor causal variants , the chosen sequencing approach found 75 ( 11+64 ) SNPs that differed between the two pooled samples . Given the finding that cANRIL expression correlates with rs10757278 genotype , we sought to determine if any of the identified SNPs might influence ANRIL splicing . For this analysis , we restricted the SNP list to include only those within 200 bp of an ANRIL intron-exon boundary , as genetic alterations in this region have the highest potential to influence RNA splicing [54] . Of the 75 SNPs present in only one of the two pooled samples , four were within 200 bp of an ANRIL intron-exon boundary ( Figure 6C , Table S1 ) . Using previously described prediction algorithms [54] , [55] , we determined the likelihood of these variants near intron-exon boundaries to change cis-regulatory splice elements including exon splicing enhancers ( ESEs ) and silencers ( ESSs ) as well as intronic splicing enhancers ( ISEs ) and silencers ( ISSs ) . A score of -1 indicates that the minor allele destroys one cis-element and +1 indicates that the minor allele creates one cis-element ( Table S1 ) . ESEs and ISEs favor exon recognition by the spliceosome when they occur in exons or introns , respectively . However , ISSs and ESSs promote exon skipping irrespective of position [56] . A pair of SNPs ( rs34660702 and NOVEL ) 10 bp apart near the start of exon 12 were predicted to alter ANRIL splicing , but the effects of these SNPs worked in opposite directions , and both exhibited low frequencies in the CEU population ( Figure 6C and Table S1 ) . More provocatively , we also observed two SNPs ( rs7341786 and rs7341791 ) in strong linkage disequalibrium with rs10757278 and which bracketed exon 15 . These SNPs demonstrated high minor allele ( CC and GG , respectively ) frequencies and with the major allele of both SNPs enhancing the “exon-ness” of exon 15 . These results indentify common , linked SNPs whose major allele is likely to inhibit skipping of exon 15 and thereby promote the production of cANRIL species ending in exon 14 in individuals homozygous for the A ( protective ) allele of rs10757278 . Due to the small number of pooled individuals for sequencing ( n = 5/group ) , we also sought to identify additional , rare polymorphisms with the potential to modulate ANRIL splicing . We analyzed all SNPs reported in the HapMap database in the chromosome 9 region 22 , 054 , 888 to 22 , 134 , 171 within 200 bp of ANRIL intron-exon boundaries for their potential to alter RNA splicing in cis- . Using this method , we identified 31 additional SNPs in HapMap that modified ESE , ESS , ISE or ISS sequences ( Table S1 ) . Notably , the majority of these variants were rare and almost never reported in the CEU population [57] . Therefore , among all of the SNPs examined from the ASVD risk interval , rs7341786 and rs7341791 appear most likely to influence ANRIL splicing . These SNPs are also in very strong linkage disequilibrium with the ASVD-associated SNP , rs10757278 ( r2>0 . 96 , based upon the nearest database SNPs in CEU ) , and therefore would be expected to correlate with ANRIL and INK4/ARF expression . However , it is important to note that other classes of germline variants ( e . g . complex indels or alterations in repetitive elements ) would not have been identified by our sequencing strategy , and therefore we are unable to exclude a role for other such variants in ANRIL expression and/or splicing .
Multiple ANRIL isoforms have been reported in the literature and EST databases , with some exhibiting differential expression patterns and SNP associations ( Figure S2 and [28] , [29] ) . Using RACE , RNA-seq and sensitive quantitative real-time PCR techniques , we now show that no single ANRIL species predominates in vivo , and that splicing to MTAP does not occur in cells with an intact 9p21 locus . Moreover , our analyses identified new ANRIL exons and variants , uncovering a novel group of circular ANRIL ( cANRIL ) species ( Figure 4 ) . Using independent approaches , we found that all ANRIL exons were expressed at very low levels , orders of magnitude lower than even the relatively rare INK4/ARF tumor suppressors , p15INK4b and p16INK4a ( Figure 1B , Figure 2B , and Figure S4 ) . This low level of expression is comparable to what we observed for other regulatory non-coding RNAs ( i . e . HOTAIR and Kcnq1ot1 ) associated with PcG-mediated repression . The discovery of non-colinear ANRIL species whose expression correlated with INK4/ARF transcription suggested that alternative splicing events might modify ANRIL structure leading to changes in PcG-mediated INK4/ARF repression . There are two major mechanisms by which non-colinear RNAs are thought to arise: trans-splicing and exon skipping [58] . During trans-splicing , a Y-branched RNA structure is formed that is sensitive to RNase R digestion [50] . In contrast , exon skipping events generate large lariat structures , which can then undergo cis-splicing to create RNase R-resistant circular RNAs . To confirm the circular nature of the ANRIL species , we showed that transcripts containing ANRIL4-6 and 14-5 were resistant to RNase R degradation , were not polyadenylated and could be PCR amplified using sets of outward facing primers ( Figure 3 and Figure 4 ) . Therefore , cANRIL species appear to result from exon skipping events occurring during RNA splicing . Based on these observations , we propose a model suggesting that common polymorphisms in 9p21 . 3 modifiy INK4/ARF gene expression through cis-regulation of ANRIL transcription and/or splicing ( Figure 7 ) . In turn , such changes in ANRIL levels or structure influence the ability of these ncRNAs to repress the INK4/ARF locus . Specifically , changes in PcG-mediated repression of p15INK4b , p16INK4a and/or ARF occur , altering potentially atherogenic cellular proliferation and ASVD risk as previously suggested [16]–[20] , [26] . Prior studies provide evidence for this model . For example , we and others have described an effect of ASVD-genotype on ANRIL transcription [16] , [30] , [59] , [60] . Moreover , meticulous allele-specific expression analyses by Cunnington et al . suggested that the cis-effects of 9p21 SNPs were stronger for ANRIL than for the coding INK4/ARF transcripts ( 20% vs . <8% ) [60] . In contrast , the effect of these SNPs on expression of the coding INK4/ARF transcripts predominantly occurred in trans- . Such findings are consistent with our model whereby causal variants directly influence ANRIL structure in cis- , thereby controlling repression of the INK4/ARF locus , possibly in trans- ( Figure 7 ) . Likewise , recent work has supported a role for ncRNAs in INK4/ARF repression . Knockdown of the RNA helicase , Mov10 , led to INK4/ARF deregulation , suggesting a role for RNA metabolism in regulation of the locus [61] . Supporting these data , Yap et al . recently demonstrated that the 5′ end of ANRIL contains stem-loop structures capable of binding CBX7 , a member of the PRC-1 complex [62] . Disruption of this interaction led to premature senescence marked by increased INK4/ARF expression . Therefore , strong evidence supports the notion that ANRIL expression correlates with ASVD-SNP genotype and that PcG-mediated repression of the INK4/ARF locus is modulated by this expression . The outstanding question of this model is the biochemical mechanism whereby causal variants at 9p21 . 3 located more than 100 kb distant from the proximal exons of ANRIL ( Figure S1 ) modulate ANRIL expression and/or structure . Our data are consistent with a “transcriptional model” wherein distal cis-regulatory elements within the ASVD risk interval directly influence ANRIL transcription . Jarinova et al . have identified an enhancer sequence within ANRIL intron 17 , which in heterologous reporter assays , enhanced transcription in a genotype-specific manner [30] . Consistent with this view , we also observed a significant effect of ASVD-SNP genotype on the expression of the linear ANRIL1-2 transcript ( Figure 5B ) . As such , the correlation of cANRIL expression with ASVD-SNP genotype would presumably reflect passive production of these circular species in the setting of increased or decreased in total ANRIL transcription . However , this model would not easily explain the positive correlation we observed between the expression of coding INK4/ARF transcripts and proximal ANRIL exons ( Figure 5A and Figure S6 ) , as the 5′end of ANRIL is reported to foster INK4/ARF repression [62] . Alternatively , we believe our present data are more consistent with a “splicing model” in which causal variants in the ASVD risk interval influence ANRIL splicing by regulating exon skipping . Provocatively , the ASVD risk interval includes exon 15 where the termination of most exon skipping events that produce cANRIL occur ( Figure 4B and 4C ) . Using next-generation sequencing of this region in individuals biased to harbor causal variants , we identified exon 15 SNPs ( rs7341786 and rs7341791 ) which are in strong linkage disequilibrium with the ASVD-associated SNP rs10757278 ( r2>0 . 96; Table S1 ) . These polymorphisms have high minor allele frequencies in our sample and the CEU population and are predicted to alter ANRIL splicing ( Figure 6 , Table S1 ) . In particular , the major allele ( ‘A’ ) of rs7341786 , detected only in individuals homozygous for the A-allele at rs10757278 , is predicted to increase the strength of exon 15 as a splice acceptor , which would promote the expression of spliced RNA circles ending in exon 14 , as observed ( Figure 4B and 4C ) . For the splicing model to explain the observed positive correlation between the expression of ANRIL and the coding INK4/ARF transcripts , distinct species of ANRIL would need to differ in their ability to repress the INK4/ARF locus . Supporting this possibility , PcG-mediated repression of the murine Kcnq1 locus is dependent upon the length of the long , ncRNA , Kcnq1ot1 [63] , [64] . Specifically , longer Kcnq1ot1 transcripts are associated with increased PcG recruitment and Kcnq1 silencing potential . Therefore , shorter ANRIL variants ( i . e . those lacking exons 4–16 ) generated by exon skipping events may also be less efficient at repressing the INK4/ARF locus . Important predictions of the splicing model are that cANRIL expression ( reflecting ANRIL splicing ) should correlate with INK4/ARF expression ( which it does , Figure 5A and Figure S6 ) , that individuals homozygous for the ‘A’ allele of rs7341786 ( and rs10757278 ) should exhibit increased production of cANRIL species containing exon 14 but not exon 15 ( which they do , see Figure 4B and 4C ) and that these individuals with increased propensity for splicing should exhibit de-repressed INK4/ARF expression ( which they do , [16] ) . A caveat to the splicing model is that the identified exon 15 SNPs need not be the true causal variant ( s ) influencing ANRIL splicing . It is possible that other polymorphisms not detected by our sequencing strategy could regulate ANRIL splicing . In particular , the sequencing strategy employed would not find differences between the pooled samples in repetitive regions , such as the large LINE element in the 12th intron of ANRIL ( Figure 6A ) . ANRIL harbors several LINE and SINE elements , and such repetitive motifs have been reported to modulate RNA splicing in other systems [65] , [66] . Therefore , while the exon 15 SNPs appear to be prime candidates to regulate ANRIL splicing , a variety of other classes of polymorphisms could also influence splicing and would not have been observed by the chosen sequencing approach . Importantly , the splicing and transcriptional models are not mutually exclusive . A single causal variant may influence both processes or there may be multiple causal variants that influence either process within the ASVD risk interval . Finally , while we believe the correlation of cANRIL expression and ASVD-SNP genotype is most likely explained by an effect of common 9p21 polymorphisms on transcription and/or splicing , a third possibility also exists . While circular RNA byproducts of exon skipping have generally been regarded as inconsequential , circular RNAs with catalytic activities ( e . g . group I and some group II introns ) are well described in bacteria , lower eukaryotes , plants [67] . In addition , some viroids and the hepatitis delta satellite virus have circular RNA genomes [68] , [69] . Although we are not aware of any endogenously produced circular RNA with discrete function in mammals , clearly circular RNAs species can possess independent functions in non-mammalian species , and we remain open to the possibility that cANRIL itself can directly participate in INK4/ARF regulation . In summary , this work links ASVD-genotype to ANRIL structure and INK4/ARF regulation , providing evidence for what we believe is a first association between endogenous circular RNA expression and a mammalian phenotype ( ASVD ) . Even were cANRIL expression merely a marker of exon skipping with no specific biologic function , we believe its expression may be a useful marker of ANRIL isoform selection , which our findings suggest is of pathogenic relevance to ASVD susceptibility . We believe this work has implications beyond ASVD , as altered ANRIL splicing could influence INK4/ARF expression , explaining the association of other nearby 9p21 SNPs with a variety of non-ASVD phenotypes in humans including longevity , type II diabetes , endometriosis and several tumors types [70]–[77] .
Research involving human subjects was approved by the University of North Carolina Institutional Review Board and all participants provided informed , written consent . WM266-4 , UACC 257 , A2058 , A375 , SUM-149 , RPMI-8322 and telomerized Hs68 cells were obtained and grown as previously described [78]–[80] . MDA-MB-468 , MDA-MB-436 , MDA-MB-231 , MCF7 , BT-474 , BT-549 , T-47D , COLO 205 , T84 , LoVo , LS 174T , SW480 , LS1034 , HeLa , HUVEC , IMR90 , Ramos , Raji , Jurkat and U-87 cells were originally obtained from ATCC and cultured as suggested . CD3 positive T-cells were isolated from human peripheral blood samples as previously described [42] . RNA was generated from proliferating cell lines and isolated human T-cells using the RNAeasy system ( Qiagen Inc . , Valencia , CA ) . 3′ and 5′ RACE was performed as described in the Firstchoice RLM-RACE manual ( Ambion Inc . , Austin , TX ) . This procedure is optimized for the detection of rare transcripts and provides additional steps to improve the specificity of mRNA amplification . Gene-specific primers were designed within ANRIL exons 4 and 6 as shown in Figure 1A and Table S2; RACE primers for other ANRIL exons tested did not amplify chromosome 9 specific products . All PCR reactions were conducted using SuperTaq-Plus ( Ambion ) in a Bio-Rad DNA Engine thermocycler . SuperTaq-Plus is a high fidelity , long range polymerase with the capability to amplify complex DNAs such as repetitive SINE , LINE and Alu elements . Cycling conditions for 5′RACE were: 94°C 3 min , 34×[94°C 30 s , 60°C 30 s , 68°C 3 min] , 68°C 5 min ( inner reaction ) and 94°C 3 min , 34×[94°C 30 s , 62°C 30 s , 68°C 3 min] , 68°C 5 min ( outer reaction ) . Cycling conditions for 3′RACE were: 94°C 3 min , 34× [94°C 30 s , 57 or 60°C 30 s , 68°C 3 min] ( outer reaction ) and 94°C 3 min , 34×[94°C 30 s , 60°C 30 s , 68°C 3 min] ( inner reaction ) . Cloning of the resulting PCR products was conducted using the TOPO-Blunt cloning kit ( Invitrogen ) . Sequencing of the resulting clones was conducted using both M13F and M13R primers . RNA was isolated using the Qiagen RNAeasy Kit and subjected to reverse transcription using the ImPromII reverse transcription system ( Promega , Madison , WI ) . RNA samples from breast and colorectal cell lines were kindly provided by Drs . N . Mitin and J . Yeh ( UNC ) . Taqman primer and probe sets for the detection of ANRIL exons 1–2 , 4–6 , 14–5 , and 18–19 as well as those for p15INK4b , p16INK4a , ARF and 18S rRNA are described in Table S2 . The ANRIL primer sets were designed to span at least one intron and were shown to have high specificity with linear amplification efficiencies between 88 and 94% ( Figure S3 ) . Final primer and probe concentrations were 900 and 250 nM , respectively . Products from the ANRIL 1–2 , 4–6 , 14–5 and 18–19 qRT-PCR reactions were cloned separately into the pBluntII-TOPO vector ( Invitrogen ) and verified . Real-time PCR was carried out in triplicate on an ABI 7900HT thermocyler . For relative expression studies , differentials were first calculated between each sample and the average 18S cycle threshold ( Ct ) for the entire experiment . Transcript Ct values were normalized using the equation: Expression = Lower limit of detection ( 37 or 40 ) - ( Ct target – Ct 18S differential ) and plotted on a log2 scale . The relative expression of p16INK4a , p15INK4b , ARF and ANRIL4-6 in PBTLs has been previously reported [42]; in this work , these same samples were reanalyzed for ANRIL1-2 , ANRIL14-5 and ANRIL18-19 expression as shown in Figure 5 and Figure S6 . In Figure 5 and Figure S6 , data was corrected to account for batch effects between the pilot and verification datasets previously described [42] . To determine the absolute number of transcripts present , a standard curve of five dilutions was generated for each experimental plate using known amounts of linearized plasmid containing the target sequence . The number of molecules detected in each sample was calculated using the equation: #Molecules = 10∧ ( ( Ct – y-intercept ) /slope ) . Primer efficiencies were calculated using the equation: Efficiency = ( 10∧ ( -1/slope ) ) -1 . Multiple comparison plots and statistical analysis appearing in Figure 5A and Figure S6 was performed using the gpairs function of the R YaleToolkit library . A total of 368 , 846 , 235 reads generated on the Illumina platform ( study SRP002274 ) were downloaded from the NCBI Short Read Archive . The reads were first screened for unique 20mers deriving from chromosome 9∶21 , 700 , 000–22 , 300 , 000 using the UCSC genome bowser Duke uniqueness mapability table . The resulting reads were mapped using the TopHat spliced aligner ( PMID: 19289445 ) to the reference human genome ( hg18 ) . The resulting coverage plot was imported into the UCSC genome browser for display . Also analyzed were two independent CalTech ENCODE mRNA-seq datasets ( http://genome . ucsc . edu/cgi-bin/hgTrackUi ? g=wgEncodeCaltechRnaSeq ) from HeLa cells . These datasets were chosen because they lacked deletion of 9p21 and showed the highest level of ANRIL expression of any publically available RNA-seq data we analyzed . For Figure 2B , the peak number of reads at a given base ( SRP002274 data ) or peak reads per kilobase mapping ( CalTech HeLa data ) were quantified and scaled relative to the average exonic coverage . Total RNA was isolated from proliferating Hs68 and IMR90 cells using a Qiagen RNAeasy kit . Equal amounts of RNA ( 43–50 µg , depending on the experiment ) were incubated with or without 40 U of Rnase R ( RNR07250 , Epicentre Biotechnologies ) for 2 . 5 hours at 37°C . The resulting RNA was purified using the RNAeasy column and quantified . Equal amounts of RNA were then subjected to reverse transcription using the ImPromII reverse transcription system and a mixture of random hexamer and oligo dT primers ( Promega , Madison , WI ) . Transcripts were quantified using Taqman-based real-time PCR . PCR primers pointing in opposite directions on ANRIL exons 1 , 4 , 6 , 13 , 16 and 18 were designed using Primer3 software and analyzed for hairpins using Netprimer ( Premier Biosoft and http://frodo . wi . mit . edu/primer3/ ) ( Table S2 ) . PCR reactions were conducted using cDNA representing 15 ng of mock or Rnase R-treated RNA . Reactions were performed using Apex Hot Start Taq DNA polymerase and Buffer 2 ( Genesee Scientific ) in a Bio-Rad DNA Engine thermocycler . The cycling conditions were as follows: 95°C 15 min , 40×[94°C 30 s , 59°C 30 s , 72°C 1 min] , 72°C 2 min . The resulting PCR products were cloned into the TOPO-Blunt cloning kit ( Invitrogen ) and sequenced using M13F and M13R primers . Genomic DNA was generated from T-cells of healthy human volunteers of known rs10757278 genotype , age and p16INK4a expression status using the Qiagen Peripheral Blood DNA isolation kit . For sequence capture , 21 ug of DNA was pooled from five individuals homozygous for the G-allele of rs10757278 and another five individuals homozygous for the A-allele . Samples were sent to NimbleGen for sequence capture using a tiled array spanning human chromosome 9 ( 22 , 054 , 888-22 , 134 , 171 ) . The resulting amplified DNA fragments were analyzed at the UNC Genome Analysis Facility using both Illumina GAII and Roche 454 technology . For sequencing on the Illumina platform , DNA was randomly sheered and appropriate adapters ligated . Resulting sequences were aligned to the entire human genome ( hg18 ) using MAQ , SOAP , and gsMapper software [52] , [53] . MAQ was run with default settings and output was translated into BAM format . SOAP alignment was performed allowing up to 10 gap bases and 2 mismatches and also translated into BAM format . Mapping with gsMapper was performed with default settings . SNP calls from MAQ and SOAP were generated using the pileup function of the SAMtools library [81] . Calls were culled to include only those SNPs appearing in ≥20 percent of reads . | Unbiased studies of the human genome have identified strong genetic determinants of atherosclerotic vascular disease ( ASVD ) on chromosome 9p21 . 3 . This region of the genome does not encode genes previously linked to ASVD , but does contain the INK4/ARF tumor suppressor locus . Products of the INK4/ARF locus regulate cell division , a process thought to be important in ASVD pathology . We and others have suggested that genetic variants in 9p21 . 3 influence INK4/ARF gene expression; however , the mechanisms by which these distant polymorphisms ( >100 , 000 bp away ) influence transcription of the locus is unknown . The ASVD–associated genetic variants lie within the predicted structure of a non-coding RNA ( ncRNA ) called ANRIL . Based upon recent work suggesting that other ncRNAs can repress nearby coding genes , we considered the possibility that ANRIL structure may regulate INK4/ARF gene expression . Coupling molecular analysis with state-of-the-art sequencing technologies in a wide variety of cell types from normal human donors and cancer cells , we found that ANRIL encodes a heterogeneous species of rare RNA transcripts . Moreover , we identified novel , circular ANRIL isoforms ( cANRIL ) whose expression correlated with INK4/ARF transcription and ASVD risk . These studies suggest a new model wherein ANRIL structure influences INK4/ARF expression and susceptibility to atherosclerosis . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"molecular",
"biology/rna",
"splicing",
"genetics",
"and",
"genomics/cancer",
"genetics",
"cell",
"biology/gene",
"expression",
"cardiovascular",
"disorders/peripheral",
"vascular",
"disease"
] | 2010 | Expression of Linear and Novel Circular Forms of an INK4/ARF-Associated Non-Coding RNA Correlates with Atherosclerosis Risk |
A family of hydrophilic acylated surface ( HASP ) proteins , containing extensive and variant amino acid repeats , is expressed at the plasma membrane in infective extracellular ( metacyclic ) and intracellular ( amastigote ) stages of Old World Leishmania species . While HASPs are antigenic in the host and can induce protective immune responses , the biological functions of these Leishmania-specific proteins remain unresolved . Previous genome analysis has suggested that parasites of the sub-genus Leishmania ( Viannia ) have lost HASP genes from their genomes . We have used molecular and cellular methods to analyse HASP expression in New World Leishmania mexicana complex species and show that , unlike in L . major , these proteins are expressed predominantly following differentiation into amastigotes within macrophages . Further genome analysis has revealed that the L . ( Viannia ) species , L . ( V . ) braziliensis , does express HASP-like proteins of low amino acid similarity but with similar biochemical characteristics , from genes present on a region of chromosome 23 that is syntenic with the HASP/SHERP locus in Old World Leishmania species and the L . ( L . ) mexicana complex . A related gene is also present in Leptomonas seymouri and this may represent the ancestral copy of these Leishmania-genus specific sequences . The L . braziliensis HASP-like proteins ( named the orthologous ( o ) HASPs ) are predominantly expressed on the plasma membrane in amastigotes and are recognised by immune sera taken from 4 out of 6 leishmaniasis patients tested in an endemic region of Brazil . Analysis of the repetitive domains of the oHASPs has shown considerable genetic variation in parasite isolates taken from the same patients , suggesting that antigenic change may play a role in immune recognition of this protein family . These findings confirm that antigenic hydrophilic acylated proteins are expressed from genes in the same chromosomal region in species across the genus Leishmania . These proteins are surface-exposed on amastigotes ( although L . ( L . ) major parasites also express HASPB on the metacyclic plasma membrane ) . The central repetitive domains of the HASPs are highly variant in their amino acid sequences , both within and between species , consistent with a role in immune recognition in the host .
Kinetoplastid parasites of the genus Leishmania cause a diverse spectrum of infectious diseases , the leishmaniases , in tropical and subtropical regions of the world ( reviewed in [1] ) . Mammalian-infective Leishmania species are divided into two subgenera , Leishmania ( Leishmania ) and Leishmania ( Viannia ) , that differ in their developmental cycles within the female sandfly vector . Transmission of species of both subgenera from vector to mammalian host requires parasite differentiation into non-replicative flagellated metacyclic promastigotes . These forms are inoculated when a female sandfly takes a blood meal; the parasites enter resident dermal macrophages and transform into replicative amastigotes that can be disseminated to other tissues , often inducing immuno-inflammatory responses and persistent infection . The fate of Leishmania amastigotes in the host determines disease type , which can range from cutaneous or mucocutaneous infection to diffuse cutaneous or the potentially fatal visceral leishmaniasis [1] . Comparative sequencing of three Leishmania genomes , L . ( L . ) major and L . ( L . ) infantum from the L . ( Leishmania ) sub-genus and L . ( V . ) braziliensis from the L . ( Viannia ) sub-genus , has revealed high conservation of gene content and synteny across the genus [2] , [3] , [4] . A number of loci show significant variation in size and gene complement between species , however . One example is the GP63 locus , containing tandemly arrayed genes coding for surface glycoproteins that are critical for macrophage invasion and virulence [5] , [6] . This locus is present in all three sequenced Leishmania species but varies considerably in size and number of genes present . Another example is the LmcDNA16 locus , originally identified on chromosome 23 of L . ( L . ) major [7] , [8] , [9] , [10] but since also found in L . ( L . ) donovani [11] , L . ( L . ) infantum [3] and other L . ( Leishmania ) species . This locus is characterised by the presence of two Leishmania-specific gene families encoding hydrophilic acylated surface proteins ( HASPs; [7] , [9] , [10] , [12] , [13] ) and small hydrophilic endoplasmic reticulum associated proteins ( SHERPs; [14] ) . The HASPs have conserved N- and C- termini but a sub-set , the HASPBs , possess divergent central domains containing hydrophilic amino acid repeats that exhibit both inter- and intra specific variation in their size and composition [10] , [11] , [15] . Acylation of the HASPBs involves N-terminal myristoylation and palmitoylation , modifications that are required for protein targeting to the parasite plasma membrane [13] . In L . major , HASPB expression is confined to mammalian-infective stages of the parasite life cycle , the metacyclics and amastigotes . While HASPBs are antigenic in the host [16] , [17] and can induce protective immune responses [18] , [19] , [20] , the biological functions of both the HASP and SHERP proteins remain unresolved [7] . To date , while the LmcDNA16 locus has been identified in all L . ( Leishmania ) species analysed , expression and localization of the encoded proteins has not been studied in New World L . ( Leishmania ) species . Here , we present analysis of HASPB expression in two representative sub-species , L . m . mexicana and L . m . amazonensis . Furthermore , the LmcDNA16 locus has been reported as absent from the published L . ( V . ) braziliensis genome , one of the few chromosomal regions showing strong divergence between sub-genera [3] . Instead , an apparently unrelated region containing several putative genes of unknown coding capacity is found in this position on chromosome 23 [3] . In this paper , we investigate this region further and identify at least two novel but closely-related L . ( V . ) braziliensis genes coding for putatively acylated repeat-containing proteins . These , like the HASPB proteins in L . ( L . ) mexicana but unlike those in L . major , are predominantly expressed on the plasma membrane of amastigotes . We name these proteins orthologous HASPs ( oHASPs ) and refer to the locus as the orthologous HASP locus ( OHL ) . Sequencing one of these new L . ( V . ) braziliensis genes in clinical isolates taken from Brazilian leishmaniasis patients has identified extensive sequence variation in the amino acid repeat regions , while some but not all sera samples taken from the same patients recognise recombinant protein expressed from the same open reading frame expressed in E . coli . These data identify a new molecular marker for L . ( V . ) braziliensis infection and suggest the potential for antigenic change within this class of amastigote proteins .
The L . ( L . ) major , L . ( L . ) infantum , L . ( V . ) braziliensis and Leptomonas seymouri genome sequences [21] , [22] were obtained from GeneDB ( www . genedb . org - [23] ) during the period June – September 2009 . Comparative alignments of the target loci ( and flanking regions ) were performed using the BLASTALL program [24] and visualised using the Artemis Comparison Tool [25] . N-terminal myristoylation and palmitoylation sites in target sequences were predicted using NMT – The MYR Predictor [26] , [27] and CSS-Palm 2 . 0 [28] with default settings . CLUSTAL alignments were generated for inter- and intra-species analysis of the oHASP protein repetitive regions using the CLUSTALW2 program ( default settings ) hosted by EBI . The Leishmania species and strains used in this study are described in Table 1 and include 11 L . ( V . ) braziliensis clinical isolates , provided as genomic DNA by the Leishmaniasis Immunobiology Laboratory , Institute of Tropical Pathology and Public Health , Goiás Federal University ( Leishbank - IPTSP/UFG/GO ) . The identities of species and strains were confirmed using restriction fragment length polymorphism ( RFLP ) analysis [29] . The clinical isolates were identified as L . ( V . ) braziliensis by PCR-typing with ribosomal DNA and glucose-6-phosphate dehydrogenase/META2 genes as described [30] , [31] , [32] . L . ( L . ) major , L . ( V . ) braziliensis and L . ( L . ) infantum parasites were maintained in culture as described [33] . L . ( L . ) mexicana and L . ( L . ) amazonensis parasites were maintained in culture and differentiated according to the method of Bates [34] . L . ( V . ) braziliensis promastigotes and intramacrophage amastigotes were generated and purified as described [33] . In brief , macrophages were incubated with stationary-phase L . ( V . ) braziliensis at a ratio of 1∶10 for 2 hr at 34°C , prior to washing twice with DMEM , replacement with fresh complete DMEM and further incubation for 48 hr at 34°C before amastigote harvesting , using 0 . 05% saponin and a single density isotonic Percoll gradient . Genomic DNA from each species and strain was extracted as follows: 5×108 – 5×109 parasites were pelleted by centrifugation ( 2000 g , 10 min , 4°C ) and washed twice with sterile PBS . Pellets were resuspended in 9 ml NET Buffer ( 0 . 01 M Tris pH 8 . 0 , 0 . 05 M EDTA , 0 . 1 M NaCl ) and 1 ml 10% SDS , ribonuclease A ( Sigma Aldrich ) added to a final concentration of 100 µg/ml and the mixture incubated at 37°C for 30 min . 200 µl proteinase K ( 20 mg/ml ) was added and the mixture incubated at 55°C overnight . Parasite genomic DNA was extracted with phenol-chloroform , washed twice in 70% ethanol , resuspended in TE buffer and stored at 4°C . PCR primers were designed using the Primer3 web utility [35] with default settings and synthesised by Eurogentec . All primer sequences used are shown in Table S1 . PCR amplifications for sequencing and cloning were carried out in either a Peltier PTC-200 Thermocycler ( MJ Research ) or a TechGene Thermocycler ( Techne ) using the Kod polymerase ( Novagen ) in 3-step reactions , according to the manufacturer's instructions . Briefly , the initial denaturing step required a 2 min incubation at 94°C and was followed by 35 reaction cycles ( 1 cycle = 95°C , 30 sec; 55°C , 10 sec; 72°C , 40 sec ) and a final extension step of 40 sec at 72°C . Southern blotting was carried out as described [7] with DIG-labeled probes and hybridization reagents ( Roche ) using the manufacturer's protocols . Primers for probe amplification were targeted against the intergenic region within the OHL locus ( Table S1 ) . The membrane was exposed to autoradiography film ( Amersham Hyperfilm HP ) and processed using a XoGraph Compact x4 ( XoGraph Imaging Systems ) . DNA sequencing of the repeat domains of oHASP genes utilised cloned PCR products amplified with suitable flanking primers ( Table S1 ) and cloned into pGEM-T easy vector . All sequencing was carried out on an Applied Biosystems 3130 sequencer , using T7 forward and Sp6 reverse primers , in the University of York Technology Facility; all data were analysed using Applied Biosystems Sequence Scanner v1 . 0 . Total RNAs ( 15 µg per track ) from procyclic , metacyclic and amastigotes of L . ( L . ) mexicana and L . ( L . ) amazonensis , generated by axenic culture [34] , were extracted and analysed by formaldehyde denaturing electrophoresis in the presence of commercial RNA markers , prior to blotting and hybridization as described [14] . The radioactive probe used for hybridization , NREP , was an oligo-labelled PCR product generated from the repetitive central domain of the HASPB gene ( GenBank: AJ251974 . 1 ) using primers NREP1 and NREP2 ( Table S1 ) . L . ( V . ) braziliensis amastigote pellets were resuspended using TRIzol Reagent ( Invitrogen ) and total RNA extracted according to the manufacturer's instructions . Further purification and quantitative real-time PCR ( RT-qPCR ) analysis was carried out as described [33] . The data generated were normalised using the constitutively expressed γ-glutamyl cysteine synthetase ( LbrM18_V2 . 1700 ) [36] . The L . mexicana HASPB open reading frame ( ORF ) was amplified using the primers LEXP5 and LEXP32 ( Table S1 ) prior to cloning into the Nde1site of pET15b and expression in E . coli BL21 ( DE3 ) pLysS [14] . His-tagged recombinant protein was purified by affinity chromatography , checked for purity by SDS-PAGE , and used to raise polyclonal antibodies ( anti-Lmex HASPB ) in rabbits , as described in [14] . The Lbr1110 ORF was PCR-amplified from L . ( V . ) braziliensis genomic DNA ( wild-type strain ) , using the Lb1110 primers ( Table S1 ) and subject to ligation-independent cloning within the University of York HiTel facility ( http://www . york . ac . uk/depts/biol/tf/hitel/index . htm ) . The resulting recombinant plasmid was introduced into E . coli Rosetta 2 and expression achieved in auto-induction medium [37] with overnight growth at 30°C . For protein purification , bacterial cells were resuspended in 70 ml buffer containing 300 mM NaCl , 20 mM sodium phosphate pH 7 . 4 , 20 mM imidazole , protease inhibitors and DNAse I . Lysis was performed by one pass through a continuous flow French Press at 20 kPSi and 4°C . The crude lysate was cleared by centrifugation at 50 , 000 g for 40 min at 4°C followed by filtration of the supernatant through a 0 . 8 µm membrane . All purification steps were carried out on an AKTA100 ( GE ) fitted with a direct loading pump . The lysate was loaded directly onto an equilibrated 1 ml HisTrap column ( GE ) at a flow rate of 1 ml/min . Following a 10 column volume ( CV ) wash with buffer A ( 300 mM NaCl , 20 mM sodium phosphate pH 7 . 4 , 20 mM imidazole ) , bound proteins were eluted with buffer B ( 300 mM NaCl , 20 mM sodium phosphate pH 7 . 4 , 0 . 5 M imidazole ) using a gradient of 0–100% B over 10 CV . Fractions of 1 ml were collected and analysed by SDS-PAGE; peak fractions were pooled and concentrated to ∼2 ml . Gel filtration was then performed using a Superdex 75 16/60 column ( GE ) and PBS buffer at a flow rate of 1 ml/min , collecting 1 ml fractions for SDS-PAGE analysis . Purified protein ( final yield , ∼4 mg/L cells ) was concentrated , stored at −20°C in PBS containing 25% glycerol and used for polyclonal antibody production in rabbits ( Eurogentech ) . Antibodies were purified using a 1 ml NHS-activated HP column ( GE ) coupled with 1 mg recombinant Lbr1110 protein . Following column equilibration with 10 ml binding buffer ( 20 mM sodium phosphate pH7 , 150 mM NaCl ) , 15 ml rabbit serum was loaded onto the column at 0 . 3 ml/min . Unbound sample was removed with 5 ml binding buffer and antibody eluted at low pH ( in 0 . 1 M glycine pH2 . 7 , 0 . 5 M NaCl ) in 0 . 5 ml fractions directly into tubes containing 50 µl 1 M Tris-HCl pH9 for neutralisation and storage . For immunoblotting , total protein lysates from 2×106 parasites were separated by SDS-PAGE prior to transfer on to PVDF Immobilon P membrane ( Millipore ) , as described [14] . The resulting blots were probed with rabbit anti-Lb1110 ( 1∶1000 ) , anti-Lmex HASPB ( 1∶500 ) and mouse anti-EF1-α ( 1∶1000; Millipore ) . Immune complexes were detected by ECL reagents ( Amersham Biosciences ) , with 30 sec exposure times . To detect immune recognition by clinical sera , similar blots were probed with sera from CL patients ( 1∶300 to 1∶500 ) and control healthy individuals ( also Brazilian ) , prior to detection with anti-human HRP ( 1∶5000; Sigma ) . For detection by confocal microscopy , antibody-labelling was performed on live parasites , to detect surface Lb1110 , and on permeabilised cells , to detect total Lb1110 localisation . 2×107 parasites were collected by centrifugation at 800 g for 10 min , washed and resuspended in 100 µl of 1% fatty acid-free BSA blocking solution ( BB International ) for 20 min . Live parasites were labelled with rabbit anti-Lb1110 ( 1∶100 ) for 30 min at 20°C , then fixed in 4% paraformaldehyde ( PFA ) before secondary detection with AlexaFluor-488-conjugated goat anti-rabbit IgG ( 1∶250 in blocking solution; Invitrogen ) . Labelling was also carried out on permeabilised cells which were first fixed in 4% PFA , washed , then incubated with 0 . 1% Triton-X100 ( Sigma ) for 10 min , washed and then incubated in 1% BSA blocking solution for 20 min at 20°C before labelling as above . Parasites were allowed to adhere to polylysine slides ( Sigma ) for 20 min and coverslips mounted with Vectashield containing DAPI ( Vector Laboratories ) , prior to imaging using a Ziess LSM 510 meta with a Plan-Apochromat 63X/1 . 4 oil DIC I objective lens . Images were acquired using LSM510 version 3 . 5 software . For detection by epifluorescence microscopy ( Figure 1C , lower panel ) , axenic amastigotes of L . mexicana were fixed and permeabilised as described above before labelling with anti-LmexHASPB ( 1∶100 ) and detection with goat-anti-rabbit-FITC secondary antibody ( Sigma ) . Fluorescent parasites were viewed using a Nikon Microphot FX epifuorescent microscope , images captured with a Photometrics CH350 CCD camera and data analysed via IPLab Spectrum software ( Scanalytics ) . Intramacrophage L . mexicana amastigote infections were carried out as described above for L . braziliensis , except that macrophages were grown on glass coverslips . Infected macrophages were fixed and permeabilised as described above . HASPB localisation ( Figure 1C , upper panel ) was determined using anti-Lmex HASPB , with detection by AlexaFluor-488-conjugated goat anti-rabbit IgG ( 1∶250; Invitrogen ) . For analysis by flow cytometry , parasites were labelled live as described above . Samples were analysed on a Dako CyAn ADP and data evaluated by Summit 4 . 3 Software . Sera samples taken from 6 patients , from whom L . ( V . ) braziliensis parasites were also isolated , were kindly provided by the Leishmaniasis Immunobiology Laboratory , Institute of Tropical Pathology and Public Health , Goiás Federal University ( Leishbank - IPTSP/UFG/GO; see Table 1 ) . The blood samples were collected as part of the initial diagnostic procedure , at the time of first clinical evaluation and prior to treatment .
The LmcDNA16 locus , encoding HASP and SHERP proteins , is conserved in all New and Old World L . ( Leishmania ) species analysed including L . ( L . ) major , L . ( L . ) donovani , L . ( L . ) infantum , L . ( L . ) mexicana and L . ( L . ) amazonensis [3] , [8] , [9] , [10] , [11] , [38] . In the two New World L . ( Leishmania ) species , L . ( L . ) mexicana and L . ( L . ) amazonensis , the LmcDNA16 locus on chromosome 23 contains several HASP genes [39] . However , unlike in L . ( L . ) major in which HASPB sequences are expressed highly in both metacyclics and amastigotes , HASPB expression occurs predominantly in amastigotes , both at the RNA and protein level , in species of the L . ( L . ) mexicana complex ( Figure 1 ) . RNA blotting with the NREP probe overlapping the central repetitive region of the predicted HASPB open reading frame ( ORF ) detects a single 2 Kb transcript in both L . ( L . ) mexicana and L . ( L . ) amazonensis that is ∼10-fold more abundant in axenic amastigotes than in metacyclic promastigotes and barely detectable in procyclic parasites ( Figure 1A ) . This expression pattern correlates with that observed at the protein level , using an antibody raised against recombinant protein expressed from the central repetitive region of the L . ( L . ) mexicana ORF to detect wild type proteins in lysates of the different parasite stages in both species ( Figure 1B ) . A single HASPB protein of ∼35 kDa ( L . ( L . ) mexicana ) and ∼29 kDa ( L . ( L . ) amazonensis ) is detected by immuno-blotting in axenic amastigotes only . As observed with L . ( L . ) major HASPB , these proteins run aberrantly when separated by SDS-PAGE [10]; the molecular masses deduced from the gene sequences are 18 . 6 kDa and 14 . 9 kDa respectively . Fluorescence microscopy using the same antibody shows localisation of the HASPB protein in a punctate pattern at the plasma membrane of both axenic and intra-macrophage parasites in L . ( L . ) mexicana ( Figure 1C ) and also , in L . ( L . ) amazonensis ( data not shown ) . These data confirm that the HASPBs of the L . ( L . ) mexicana complex are differentially regulated during the parasite life cycle , as in L . ( L . ) major , but unexpectedly , expressed predominantly in the macrophage-dwelling amastigotes . Although conserved in L . ( Leishmania ) species , the LmcDNA16 locus was reported as missing in L . ( V . ) braziliensis , one of the few chromosomal regions showing significant divergence between Leishmania species sequenced to date [3] . Instead , a non-syntenic region of ∼7Kb ( named here the OHL locus ) is positioned at the same chromosomal location in L . ( V . ) braziliensis , as determined by examination of the LmcDNA16 locus flanking regions that contain genes that are conserved in all sequenced L . ( Leishmania ) species ( Figure 2A ) . In addition , partial genome sequencing of Leptomonas seymouri , a related insect parasite , has identified a similar variable region between the same conserved flanking genes which is of reduced size and contains two ORFs ( Figure 2A ) . To confirm the content of the Leishmania loci , PCR amplification was used to probe L . ( Viannia ) and L . ( Leishmania ) species for HASPB and SHERP sequences , as well as for the two new ORFs identified in the OHL region of L . ( V . ) braziliensis ( LbrM23V2 . 1110 and LbrM23V2 . 1120 ) . This analysis confirmed the presence of conserved HASPB and SHERP genes in all analysed L . ( Leishmania ) species and their absence in L . ( Viannia ) species ( data not shown ) . Similarly , the newly identified ORFs were only detected in the L . ( Viannia ) species although notably , the sizes of the bands observed were variable in both number and size ( data not shown ) . Previous studies have shown that the genes within the LmcDNA16 locus exhibit both inter- and intra-species variation in size and content [11] , [15] . Similar variation in the size of the OHL region was demonstrated by hybridisation analysis of genomic DNAs from L . ( V . ) braziliensis , L . ( V . ) peruviana and L . ( V . ) guyanensis ( Figure S1 ) . Southern blots of HinDIII/XhoI-digested DNA ( utilising restriction sites flanking the L . ( V . ) braziliensis OHL region ) probed with a specific intergenic fragment ( located between LbrM23V2 . 1110 and LbrM23V2 . 1120; see Figure 2B ) identified single bands of different sizes larger than 12 Kb in L . ( V . ) braziliensis , L . ( V . ) guyanensis and L . ( V . ) peruviana DNA . These fragments were all considerably larger than the ∼7 Kb predicted to span the break in chromosomal synteny derived from L . ( V . ) braziliensis genome analysis ( Figure 2A ) . Additional bioinformatics analysis revealed a sequence mis-assembly derived from a ∼3 . 2 kb collapsed repeat sequence within the OHL locus . Collapsed repeats of this type frequently arise during automated genome assembly when sequence reads originating from distinct repeat copies are incorrectly joined to generate a single unit . They are identified as genomic regions with significantly increased read depth . The collapsed repeat identified here contains conserved ∼1 . 2 kb sequences ( A ) flanked by ∼0 . 8 kb sequences ( B ) forming an ABAB motif , as shown in Figure 2B . Each A sequence contains a putative ORF containing multiple iterations of conserved 30 nt sequences that code for a large amino acid repeat domain ( see Figure 3 ) . The number of repetitive 30 nt sequences varies between the two ORFs identified in GeneDB ( http://www . genedb . org ) as LbrM23V2 . 1110 and LbrM23V2 . 1120 , with 4 and 14 iterations respectively . It is important to note however that these two ORFs differ only in the number of repeat units present . Examination of the individual sequence reads that map to the collapsed repeat region reveal the presence of another ORF variant ( containing 12 iterations of the repeat motif ) . While only three variant ORFs were detected in this analysis , the increased read depth within the OHL region suggests that multiple copies of each motif could be present and that the structure of this repeat region consists of a tandemly repeated ABAB pattern , with sequence diversity within the iterated sequences , spanning more than 12 Kb of genomic DNA . Further analysis to more precisely define the size and composition of the OHL locus is in progress . From the analysis above , the two ORFs identified within the OHL region ( LbrM23V2 . 1110 and LbrM23V2 . 1120 ) represent only part of the coding capacity of this domain; there are several more related genes that are not mapped within the OHL locus representation shown in Figure 2 . Focusing on the sequence of the single LbrM23V2 . 1120 ORF , features characteristic of Leishmania genes were identified: a translation initiation site ( Figure 3Ab ) with an upstream AG splice acceptor site flanked by a conserved consensus sequence motif ( −12cCNcccNcNCAGNaN ( C/T ) N+5; Figure 3Aa ) preceded by a long polypyrimidine tract . CLUSTALW alignments of all putative ORFs identified in this locus , together with their flanking regions , revealed strong conservation of the 5′- and 3′-UTRs and putative conserved splice acceptor sites ∼230 nt upstream of the translation initiation site ` ( data not shown ) . While 3′ polyadenylation ( poly A ) sites show significant variation between characterised Leishmania genes and cannot usually be identified by simple sequence consensus motifs , use of the PREDATERM program here ( which predicts poly A sites based on local nucleotide composition ) facilitated identification of putative poly A sites within the flanking B sequences of the oHASP genes ( as positioned in Figure 3Ac ) . This information suggested that the 3′-UTRs of these genes are extensive , in common with other Leishmania genes . While these predicted RNA processing sites require experimental verification , their positions confirm that the OHL genes span the A and B sequences in Figure 2B , with the repeats arranged in an AB , AB reiterating pattern for RNA expression . Comparative analysis of the putative proteins encoded by the OHL ORFs revealed significant conservation although , as described above , variation was observed in the composition and number of iterations of the 30 nt repeats that code for hydrophilic 10 amino acid repeats ( Figures 3B ) . Of particular interest is the presence of conserved N-terminal residues , including a 2nd position glycine and a 5th position cysteine , confirmed as potential sites for N-myristoylation and palmitoylation using the NMT- The MYR Predictor and CSS-PALM predictive tools [27]–[28] . By contrast , screening for potential prenylation sites ( by PrePS ) , GPI-modification sites ( by big-PI Predictor ) or GPI-anchor signal sequences ( by GPI-SOM ) returned no positive predictions . Overall , these data indicate that the AB sequence repeats embedded within the OHL locus have the necessary sequence components for identification as functional genes coding for proteins that contain large internal hydrophilic repeat domains and may be modified both co- and post-translationally by N-myristoylation and palmitoylation . The OHL ORFs , therefore , have very similar characteristics to the L . ( Leishmania ) HASPB proteins , features evident in the comparisons and alignments presented in Figure 3 and Figure S2 . A similar analysis of the two L . seymouri ORFs reveals that both contain large hydrophilic amino acid repeat domains that are larger than those observed in the HASPs and oHASPs and also more degenerate ( Figure 3B ) . Predicted sites for N-myristoylation and palmitoylation are also found in these deduced protein sequences ( Figure 3B ) . To investigate RNA expression from the two characterised OHL genes ( LbrM23_V2 . 1110 and LbrM23_V2 . 1120 ) , RT-qPCR was used for quantitative analysis of transcript levels in macrophage-derived amastigotes and axenic procyclic and metacyclics of L . ( V . ) braziliensis . The results were normalised using the experimentally-characterised γ-glutamyl cysteine synthetase ( LbrM18_V2 . 1700 ) as a constitutive control and Meta1 ( LbrM17_V2 . 0980 ) as a marker for metacyclic expression [36] . Data were analysed using the Pfaffl method [33] and showed that the transcript abundances from both genes are increased 5 – 10 fold in the infective metacyclic and amastigote stages relative to the procyclic stages of the parasite ( Figure 4 ) . As expected , expression of the Meta1 gene was highly up-regulated in metacyclics versus amastigotes , as previously reported [36] . These observations show strong correlation with the stage-regulated transcript abundances of the HASPB genes in L . ( L . ) major [8] but less so with the L . ( L . ) mexicana species , that show predominant RNA expression in amastigotes , by RNA blotting and hybridisation analysis ( Figure 1 ) . To further probe gene expression at the protein level , the OHL gene encoding the smallest number of repeats ( LbrM23_V2 . 1110 or Lb1110 ) was chosen for detailed analysis . Using an affinity-purified antibody raised against recombinant Lb1110 expressed in E . coli ( see Materials and Methods ) , several approaches were used to investigate Lb1110 expression in the different L . ( V . ) braziliensis life cycle stages ( Figure 5 ) . Firstly , immunoblotting detected strong reactivity with the 15 kDa recombinant protein and recognised a protein of the same size only in amastigotes of L . ( V . ) braziliensis ( Figure 5A ) . In view of the RNA expression analysis above , this observation suggests that expression of the Lb1110 gene may be regulated translationally , unlike the L . ( L . ) mexicana homologue described in Figure 1 . A second protein , migrating at ∼2x the observed molecular mass of recombinant Lb1110 , was also detected by anti-Lb1110 in amastigote lysates although this molecule was also detectable at low levels in metacyclics and procyclics ( when compared to the differential loading indicated by detection of the constitutive marker , EF1α ) . It is likely that this protein is an additional oHASP containing a more extended repeat domain that is recognised by anti-Lb1110 . Other OHL genes may also be expressed as proteins that are of lower abundance and therefore not readily detectable by immunoblotting . Flow cytometry was used as a second method to quantify Lb1110 protein surface expression in procyclic , metacyclic and amastigote L . ( V . ) braziliensis . In these experiments , surface-exposed Lb1110 was detected by live primary antibody labelling prior to fixation and detection with AlexaFluor 488-conjugated goat anti-rabbit IgG ( see Materials and Methods ) . Prior to live cell staining , the amine-reactive fluorophore sulfo-succinimidyl-7-amino-4-methylcoumarin-3-acetic acid ( Sulfo-NHS-AMCA ) was used to confirm cell viability; dead cells staining with this reagent emit a strong blue fluorescence and could be omitted from further analyses . As shown in Figure 5B , live cell staining with anti-Lb1110 was not detectable in control ( no primary antibody used ) or procyclic parasite populations . Conversely , 13% of parasites in a metacyclic population stained with anti-Lb1110 while 90% of amastigotes were positive with this antibody . These results confirm the stage-specificity of Lb1110 expression and demonstrate its surface exposure on the majority of amastigotes . As a third approach to determining the localisation of the Lb1110 protein , expression was visualised in either live or permeabilised and fixed L . ( V . ) braziliensis by indirect immunofluorescence and confocal microscopy ( Figure 5C ) . Antibody labelling was carried out either pre- or post-fixation at 20°C , in order to compare antigen localisation at the surface membrane with that detected both externally and internally within the parasite . DAPI staining of the parasite nucleus and kinetoplast was used as a counter-stain in these experiments . As shown in the upper panel of Figure 5C , anti-Lb1110 staining is specific to L . ( V . ) braziliensis amastigotes and , in live antibody labelled cells , Lb1110 localises to a site close to the protrusion of the rudimentary flagellum , which could be indicative of antibody capping of the surface exposed protein . In permeabilised cells ( labelled Total Lb1110 ) , by comparison , staining is evident in a punctate pattern indicative of plasma membrane and flagellar localisation on both faces of the membrane bi-layer . In the lower panel of Figure 5C , a single L . ( V . ) braziliensis amastigote is shown at higher magnification , clearly demonstrating the plasma membrane localisation following permeabilisation but surface localisation to the rudimentary flagellum in the non-permeabilised L . ( V . ) braziliensis cell . In contrast , the live labelling pattern on the L . ( L . ) mexicana amastigote in the same figure ( ‘Surface HASPB’ ) is very similar to the total labelling pattern on the fixed L . ( V . ) braziliensis amastigote ( and to the fixed labelling seen in Figure 1C ) , suggesting that antibody capping is minimal on live L . ( L . ) mexicana under the labelling conditions used . Overall , these data suggest that the surface distribution of Lb1110 to the amastigote flagellum is not an artefact of antibody capping in these live cells . Given the amastigote-dominant expression of Lb1110 and its surface exposure on live parasites , in a pattern similar to that observed for HASPB expression in L . ( L . ) major metacyclic parasites [40] , we next investigated whether this protein is recognised by human immune serum collected from patients infected with L . ( V . ) braziliensis . Six serum samples derived from infections with the L . ( V . ) braziliensis clinical isolates listed in Table 1 were used to probe blots of separated parasite proteins from different stages , together with recombinant protein ( as used in Figure 5A ) . Two examples of these immunoblots , representative of the patterns observed , are shown in Figure 6 , probed with serum taken from HPV-06 and TMB-06 infections ( using the same serum dilution and length of chemical exposure for all blots ) . Recognition of a broad size range of proteins in total parasite extracts was evident in all stages with each antiserum , while normal human serum detected few proteins above background levels and did not recognise recombinant Lb1110 . Interestingly , the recombinant protein was strongly detected by HPV-06 but not by TMB-06 . Overall , these data confirm the antigenicity of Lb1110 and , as with the central repetitive domain of L . ( L . ) major HASPB , it can be predicted that the Lb1110 repeats may provide dominant epitopes for antibody recognition . It is also evident that not all antisera taken from infected patients recognise Lb1110 , suggesting that this antigen could be unstable or variant in vivo . To investigate this further , oHASP gene repeats were analysed in a number of L . ( V . ) braziliensis clinical isolates ( provided as genomic DNAs from Leishbank - IPTSP/UFG/GO and listed in Table 1 ) . Variation in the number of repeat iterations present in each of the 2 OHL ORFs described above ( LbrM23V2 . 1110 and LbrM23V2 . 1120 ) , coupled with the large size of the non-syntenic region , raised the possibility that further ORFs with distinct repeat regions might be present in this region , as discussed earlier . To verify this prediction , genomic DNA from the L . ( V . ) braziliensis genome strain ( MHOM/BR/75/M2904 ) was subjected to PCR with primers designed to amplify the repetitive domain in the oHASPs ( Figure 3Ad , Table S1 ) . The PCR products were sub-cloned into the pGEM-T-easy vector , 10 clones of each selected and their insertions sequenced . The repeat domain structure was then determined for each clone and each unique sequence translated and aligned using CLUSTALW . Three unique sequences were identified ( comprising 9 , 12 and 14 repeat units ) and aligned ( Figure S3 ) revealing variations in the number and sequences of the repeat domains of the oHASP ORFs in a single strain . The variations in size and composition of the repeat domains of the L . ( V . ) braziliensis genome strain oHASPs , as described above , are similar to those reported in the repetitive domains of L . ( L . ) major and L . ( L . ) donovani HASPBs [11] , [15] and typical of the observed inter- and intra- species variation in this protein family in the L . ( Leishmania ) subgenus . To determine the extent of similar variation in the L . ( Viannia ) oHASPs , a total of 11 isolates of L . ( V . ) braziliensis ( Table 1 ) and single strains of L . ( V . ) peruviana and L . ( V . ) guyanensis were analysed , using the same approach as above , cloning and sequencing multiple clones of each strain . These data are presented in Figure 7 , which includes the organisation of translated repeat domains in up to 4 independent clones from each PCR amplification ( A ) and the composition of each repeat unit analysed ( B ) . The number of distinct repeat domains identified in each strain of L . ( V . ) braziliensis varied from 2–6 per strain . The average repeat domain comprised 14 iterations with 6 being the lowest observed and 15 the highest ( Figure 7 ) . While the composition of the repeat domains varied both between and within strains , it is interesting to note that the most prevalent motif within the repeat region across all strains ( excepting LTB300 and RPL-05 ) is GGDHGHEHMD ( Figure 7B , sequence G ) . Also of note , the repeat domains in the LTB300 strain are very similar to those observed in L . ( V . ) peruviana strain LpLCA08-90 . Similarly , the RPL-05 strain shares greater conservation with the L . ( V . ) guyanensis strain LgM4147-75 repeat domains than with the L . ( V . ) braziliensis genome strain ( M2904-75 ) . In these cases , the most prevalent repeat unit appears to be GGDHVPEKAN ( Figure 7B , sequence X ) and GGDHGHGNMD ( Figure 7B , sequence F ) respectively . Intriguingly , the overall structure of the repeat domains is well conserved between the representative sequences from each strain with the individual motifs occupying very specific positions ( Figure 7 ) . The functional significance of this conservation has not yet been investigated further . In comparison to L . ( V . ) braziliensis , considerable variation in the repeat domains was observed for both L . ( V . ) peruviana and L . ( V . ) guyanensis ( although only a single isolate of each species was investigated ) . Analyses of these sequences show the level of conservation at the amino acid level between L . ( V . ) braziliensis and L . ( V . ) guyanensis to be ∼76% , L . ( V . ) braziliensis and L . ( V . ) peruviana ∼62% and L . ( V . ) guyanensis and L . ( V . ) peruviana ∼57% , suggesting that the L . ( V . ) peruviana repeat domain is the most divergent in content ( Figure 7 ) . Moreover the average size of the oHASP repeat domain is significantly less in L . ( V . ) peruviana ( Figure 7 ) . These species-specific variations in the size and content of the repeat domains are similar to those observed in the HASPB sequences in L . ( Leishmania ) species .
In addition to the complete genomes of L . ( L . ) major , L . ( L . ) infantum and L . braziliensis [2] , [3] , sequence data are also currently available for L . seymouri , a monogenetic protozoan that parasitizes insects , nematodes and ciliates and is the closest sequenced relative to Leishmania . The presence of a syntenically-positioned locus containing ORFs that code for putative N-acylated proteins containing large hydrophilic amino acid repeat domains suggests the presence of this hypermutable locus in the pre-Leishmania state . Whether this locus is present in Crithidia species remains unknown . Given the comparative simplicity of the locus in L . seymouri , the expansion seen in Leishmania spp . could be representative of the shift from the monogenetic life cycle of ancestral Leishmania to the digenetic life cycle of parasites from the Leishmania sensu strictu genus . A key step in this process is the evolution of the parasite-parasitized insect relationship allowing Leishmania to use sand flies as their vector . Our recent observation that the LmcDNA16 locus is essential for L . ( L . ) major differentiation in Phlebotomus papatasi [41] may be of relevance in this respect . Recent studies have demonstrated the importance of HASP proteins for L . major differentiation in the sand fly vector , while the antigenic properties of these molecules suggest their suitability as targets for vaccine development . Previous comparative genomic analyses of L . ( V . ) braziliensis , L . ( L . ) major and L . ( L . ) infantum , however , reported the absence of the HASP/SHERP ( or LmcDNA16 ) locus on chromosome 23 in L . ( V . ) braziliensis – with a smaller non-syntenic locus ( the OHL locus ) found at that location . In this paper , we show that the oHASP proteins coded within the OHL locus are orthologues of HASPB , possessing similar expression , localisation and antigenic properties . Of particular interest is the inter- and intra-species variation in the size and composition of the oHASP repeat domains ( also observed in HASPBs ) which could indicate that host ( and/or vector ) immune pressure is driving sequence diversification within this locus . Further study is now required to investigate the antigenic properties of the oHASPs , explore their interaction with the host immune system and investigate their utility as diagnostic agents for L . ( L . ) Viannia clinical infections . | Single-celled Leishmania parasites , transmitted by sand flies , infect humans and other mammals in many tropical and sub-tropical regions , giving rise to a spectrum of diseases called the leishmaniases . Species of parasite within the Leishmania genus can be divided into two groups ( referred to as sub-genera ) that are separated by up to 100 million years of evolution yet are highly related at the genome level . Our research is focused on identifying gene differences between these sub-genera that may identify proteins that impact on the transmission and pathogenicity of different Leishmania species . Here we report the presence of a highly-variant genomic locus ( OHL ) that was previously described as absent in parasites of the L . ( Viannia ) subgenus ( on the basis of lack of key genes ) but is present and well-characterised ( as the LmcDNA16 locus ) in all members of the alternative subgenus , L . ( Leishmania ) . We demonstrate that the proteins encoded within the LmcDNA16 and OHL loci are similar in their structure and surface localisation in mammalian-infective amastigotes , despite significant differences in their DNA sequences . Most importantly , we demonstrate that the OHL locus proteins , like the HASP proteins from the LmcDNA16 locus , contain highly variable amino acid repeats that are antigenic in man and may therefore contribute to future vaccine development . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"genetics",
"and",
"genomics/genomics",
"molecular",
"biology",
"cell",
"biology/gene",
"expression",
"infectious",
"diseases/protozoal",
"infections"
] | 2010 | Leishmania-Specific Surface Antigens Show Sub-Genus Sequence Variation and Immune Recognition |
To ascertain the clinical features and visual outcome of toxoplasma retinochoroiditis in a large series of cases . Two hundred and thirty subjects diagnosed with active toxoplasma retinochoroiditis were prospectively followed for periods ranging from 269 to 1976 days . All patients presented with active retinochoroiditis and positive IgG T . gondii serology at the beginning of the study and received a standardized drug treatment for toxoplasmosis , both in the first episode and in the subsequent recurrences . The group involved 118 ( 51 . 3% ) men and 112 ( 48 . 7% ) women , with ages ranging from 14 to 77 years , mean of 32 . 4 years ( SD = 11 . 38 ) . Primary retinochoroidal lesions were observed in 52 ( 22 . 6% ) cases and active retinochoroiditis combined with old scars in 178 ( 77 . 4% ) subjects at the beginning of the study . A hundred sixty-two recurrent episodes in 104 ( 45 . 2% ) patients were observed during follow-up . New subclinical retinochoroidal lesions were detected in 23 of 162 ( 14 . 2% ) recurrences episodes during the follow-up . Posterior segment complications were observed in 73 ( 31 . 7% ) subjects . Retinochoroidal lesions adjacent to the optic nerve and in the macular area were observed in 27 of 40 ( 67 . 5% ) cases of severe visual impairment ( VA = 20/200 or worse ) . Toxoplasma retinochoroiditis in this population had a high recurrence rate after an active episode . Severe visual impairment was associated with location of the retinochoroidal scar , recurrences and posterior segment complications . It is crucial to consider the location of the lesion in studies analyzing visual prognosis as a measure for treatment effectiveness and prevention strategies .
Ocular involvement by Toxoplasma gondii may cause severe visual impairment and is the main etiology of posterior uveitis in many parts of the world , however questions about the course of the disease remain unanswered . [1–3] Up to this day there is no consensus on treatment effectiveness and risk factors for the occurrence and recurrence of this eye disease . [2 , 4–6] Brazil hosts the highest prevalence ever described for ocular toxoplasmosis with a preponderance of different strains from those that predominate in Europe and North America , this alone would possibly justify clinical differences and long term outcome . [7–10] The study of ocular toxoplasmosis offers many challenges to researchers , mainly because of the plethora of factors that may affect the course of the disease and the need of long term follow-ups . To observe outcomes of toxoplasma retinochoroiditis ( TRC ) and especially the causes of vision loss in affected patients it is crucial to guide the analysis of prevention strategies and treatment . The study herein presented describes the clinical characteristics of a series of prospectively followed cases of toxoplasma retinochoroiditis , analyzing the course of the disease , recurrence frequency and factors that may influence the visual prognosis .
Follow-up visits were scheduled within 30 or 45 days and every year after the initial consultation . Yearly returns were programmed for every subject as well as individual consultations according to the presence of complications . Subjects were instructed to seek the Infectious Ophthalmology Laboratory of INI at any time in case of eye symptoms , especially blurred vision , red eye or floaters . Subjects who did not return on the scheduled dates were contacted by phone or mail . A minimal of three consultations was required to consider the patient followed , one for admission , the second within 45 days after the first one and the third between 8 months after the first visit and July 2015 . Retinography was performed for assessment of retinal status before and after treatment , whenever the transparency of the ocular media and the patient allowed and depending on the availability of the retinal camera . For the purpose of this study visual acuity ( VA ) of 20/200 or worse with the best correction in the affected eye was considered severe vision impairment . [12] All subjects with active retinochoroidal lesions , both in the initial active episode and in the subsequent reactivations , were treated for 30 to 45 days with standardized drug scheme for toxoplasmosis ( sulfadiazine 4 x 1g daily , pyrimethamine 25–50 mg daily , folinic acid 15 mg every other day and prednisone 0 . 5 to 1 mg / kg / day as the initial dose followed by gradual tapper off ) for 4 weeks or alternatively with 800mg-160mg 12/12h sulfametoxazole/ trimethoprim ( SMZ-TMP ) and prednisone or clindamycin ( 4 x 300 mg daily ) replacing sulfadiazine in case of allergy . Serologic tests for toxoplasmosis were performed at the Toxoplasmosis Laboratory of IOC-Fiocruz by immunoenzymatic test ( ELISA ) and indirect immunofluorescence assay ( RIFI ) . Follow-ups through July 2015 were used in the result analysis , which was performed as planned ( Fig 5 ) .
We prospectively followed 230 subjects from January 2010 until July 2015 , 118 ( 51 . 3% ) men and 112 ( 48 . 7 women for periods ranging from 269 to 1976 days , mean 1060 days . ( S1 Chart ) All subjects included in the study lived in the State of Rio de Janeiro and only one reported nationality other than Brazilian . Ages ranged from 14 to 77 years old ( mean = 32 . 4 , SD = 11 . 4 ) distributed as shown in Table 1 . At the beginning of the study 52 ( 22 . 6% ) patients with primary retinochoroidal lesions and 178 ( 77 . 4% ) with recurrent lesions were observed . In 91 of 178 ( 51 . 1% ) individuals it was associated with one scar and in 87 ( 48 . 9% ) individuals it was associated with more than one retinochoroiditis scar . There were 162 recurrence episodes in 104 ( 45 . 2% ) subjects during follow-up , 53 episodes ( 32 . 7% ) occurred within the first year , 53 ( 32 . 7% ) in the second year and 56 ( 34 . 6% ) in the subsequent years . Among 104 subjects who presented recurrence , 14 ( 13 . 4% ) individuals had two episodes in the same year and 43 ( 41 . 7% ) had more than one episode of recurrence throughout follow-up . Recurrences were observed away from previous scars in 20 ( 12 . 3% ) of the 162 episodes . Recurrences were not observed in 19 of the 52 ( 36 . 5% ) patients with a primary lesion during the studied period . At the beginning 177 of 230 ( 76 . 9% ) individuals had unilateral disease and 9 ( 3 . 9% ) developed retinochoroidal lesion in the otherwise healthy eye during follow-up . Most subjects ( 68 . 7% ) were between 20–39 years old and it was possible to determine the age of the first episode in only 83 ( 36 . 1% ) patients . Subclinical new retinochoroidal lesions were detected in 23 of 162 ( 14 . 2% ) recurrences episodes during the follow-up , by visualization of small active lesions in routine consultations in 3 cases ( 13 . 0% ) , by photographic documentation in 7 cases ( 30 . 5% ) or description of the number of lesions 13 cases ( 56 . 5% ) . A hundred fifty nine ( 69 . 1% ) individuals could be identified with evidences of a subclinical scenario of ocular toxoplasmosis up to the moment they were examined in this study . It was possible to retrospectively identify congenital toxoplasmosis in only one case ( 0 . 4% ) , and acute systemic symptomatic infection in three ( 1 . 3% ) of 230 individuals; the time of first exposure for T . gondii was undeterminable in the remaining patients . Positive IgM serology for T . gondii was found in 21 ( 9 . 1% ) individuals . At least one central lesion ( within the large vascular arcades ) was observed in 172 ( 74 . 8% ) subjects , and in 58 ( 25 . 2% ) only peripheral lesions ( outside the large vascular arcades ) were seen . Ninety ( 52 . 3% ) of these 172 patients had retinochoroidal lesions adjacent to the optic disc or in the macular area . Posterior segment complications attributable to TRC ( described in Table 2 ) were observed in 73 of 230 ( 31 . 7% ) patients and it was more frequent in subjects older than 40 years of age ( p = 0 . 008 ) . Five patients ( 2 . 2% ) had diagnosis of outer punctate toxoplasmosis . Vasculitis , neuro-retinitis and perilesional neovascularization were observed respectively in 14 ( 6 . 1% ) , 3 ( 1 . 3% ) and 4 ( 1 . 7% ) of the 230 subjects and were not considered complications but clinical manifestations of ocular toxoplasmosis Vasculitis and perilesional neovascularization were seen in association with active exudative lesions before treatment and remission was observed in all cases after the standardized drug scheme used . Neuro-retinitis cases progressed to optic atrophy in 2 subjects . The four patients with retinal detachment were referred to surgery , as well as 3 subjects with epi-retinal membranes and 3 with residual vitreous opacities . Criteria of surgery recommendation for macular hole , epi-retinal membranes and macular pucker was based on the presence of retinal traction or visual acuity worse than 20/100 . Residual vitreous opacities were defined as those seen until the last visit of the patient follow-up and were only referred to surgery if visual acuity remained worse than 20/100 , at least 6 months after the last active episode . When these complications were associated with macular scars surgery was discussed with the patient regarding visual prognosis . Peripheral retinal tear was referred to treatment with laser photocoagulation , retinal hemorrhages and vitreous hemorrhage were managed conservatively with spontaneous resolution . Subjects with clinically evident macular edema that persisted after standardized drug scheme for toxoplasmosis were monitored and resolution observed within 3 months after the active episode . Two hundred and eight ( 90 . 4% ) followed subjects initiated treatment of acute episodes with a standardized scheme for toxoplasmosis with sulfadiazine , pyrimethamine , prednisone and folinic acid , 15 subjects with SMZ-TMP ( 6 . 5% ) and 7 ( 3 . 0% ) had allergic reactions that justified the replacement of sulfadiazine by clindamycin and 11 ( 4 . 8% ) subjects discontinued treatment without medical advice due to side effects of medications . Discontinuation of the treatment was not associated with recurrences ( p = 0 . 76 ) or presence of complications ( p = 0 . 99 ) . At the end of the follow-up 122 ( 53 . 0% ) subjects had a decrease in visual acuity attributable to after-effects of the posterior uveitis ( VA <20/30 ) , 40 of 230 subjects ( 17 . 4% ) had severe vision impairment ( VA = 20/200 or worse ) in the affected eye but in only 1 of 62 ( 1 . 6% ) patients with bilateral lesions the visual acuity was 20/200 or worse in both eyes . There was an association between severe vision impairment and retinochoroidal location , complications and recurrences during follow up but not with age >40 years ( Table 3 ) .
Clinical manifestations of toxoplasmic retinochoroiditis have been described extensively in the literature , however the vast majority of the studies involve subjects treated at different centers , with different therapeutic regimens , with short follow-ups or retrospectively analyzed . [13–17] This study prospectively followed 230 subjects from a single center and evaluated in a standardized manner . Monitoring of asymptomatic subjects involved active search and persistent calls especially after the first year of follow-up which provided the detection of subclinical episodes of recurrence during follow-up . We observed 159 ( 69 . 1% ) subjects with evidence of asymptomatic episodes up to the moment they were examined in the study , this shows that the disease can be subclinical in many individuals . This data was obtained by observing subclinical episodes during follow up and by comparing the number of retinochoroidal scars in the first visit of the study with the number of episodes referred by the patients . We cannot estimate how many of them had congenital or early childhood lesions that could justify the absence of previous history . Other possible causes of this observed phenomenon is the existence of previous vitreous opacities or vision impairment , which could make new symptoms more difficult to be perceived by the patient . Finding high rates of asymptomatic lesions may denote an underestimation of the frequency of the disease in individuals infected with T . gondii , a benign aspect due to the low morbidity associated with many lesions , but a hidden danger , with indefinite prognosis in patients in whom a retinochoroiditis scar is randomly diagnosed in clinical practice . This demonstrates the importance of studying risk factors associated with the recurrence of ocular toxoplasmosis in order to identify which individuals are more susceptible to the disease and devise strategies for monitoring and prevention . Active seeking of asymptomatic subjects during follow-up was a strategy used to avoid the bias of monitoring just patients who had more severe and recurrent disease . Unfortunately , it was not possible to separate individuals in congenital and acquired ocular toxoplasmosis groups because in only 4 ( 1 . 7% ) of 230 cases the moment of first exposure was certain . Serology for toxoplasmosis with positive IgM antibodies was detected in 21 ( 9 . 1% ) followed subjects , assuming acquired disease in such cases . The low percentage of serologic confirmed acquired cases does not allow inferences about the mode of infection acquisition by the entire population . The predominance of quiescent retinochoroidal scars regarding the frequency of primary lesions ( 22 . 6% ) has been described previously . [1 , 13 , 18 , 19] Individuals with primary retinochoroiditis lesion and no recurrence during the follow-up were included after negative results for HIV and syphilis tests and if they responded to specific drug treatment for toxoplasmosis , meeting these criteria , they were considered highly probable of TRC . There are several sources of potential bias in this study . First of all , we cannot completely rule out mild tuberculous , histoplasmosis and herpes uveitis especially in those subjects where recurrence was not observed . Other published studies often have the same limitation as these diagnosis are frequently made on exclusion or presumed basis and ocular toxoplasmosis is only confirmed by intra-ocular fluid analyses eventually . [13 , 20 , 21] On the other hand , clinical evolution and the observed fundoscopic characteristics speaks favorably towards TRC diagnosis , cases with inaccessible retinal examination were considered ineligible and those with multiple active lesions were excluded as they lacked laboratory confirmation with intra-ocular fluid examination . Subclinical episodes were described based on sequenced retinographies but also by counting lesions , which brings a remote possibility of missed lesions in prior examinations being wrongly considered . What minimizes this possibility was the standardized indirect ophthalmoscopy examination performed by an uveitis experienced staff . Another aspect that should be considered concerning the subclinical episodes is that 3 subjects were diagnosed with active retinochoroidal lesions without new symptoms , these episodes may have been diagnosed before symptoms arose because of the programmed schedule of visits . The frequency of complications should also be carefully analyzed . If OCT and fluorescein angiography were routinely performed the complications rate could have been higher . Unfortunately , these ancillary tests were only requested to confirm clinical diagnosis of complications and investigation of vision loss and were performed in other health units because they were not available in our service by that time . For this reason they could not be included in our study protocol . The studied population showed a homogeneous distribution in relation to gender , and a larger number of cases among 20–29 years and 30–39 years ( 36 . 9% and 31 . 7% respectively ) was also observed in other series described in literature . [1 , 13 , 22] One hundred and sixty two episodes of recurrence in 104 ( 45 . 2% ) patients were observed during follow-up , and this high recurrence rate was also observed in other studies . [17–19 , 22–24] It was found that 53 ( 32 . 7% ) episodes of recurrence took place within the first year of follow-up and other 53 ( 32 . 7% ) in the second year , which suggest a high risk of recurrence in the first two years after an episode of active retinochoroiditis , however , this should be confirmed by other statistical analysis specially concerning survival time . The clustering pattern of recurrences in TRC has been described previously and although the causes of reactivation are not completely understood , factors such as age and host defenses have been proposed to influence this aspect of the disease . [25] If we consider that other publications involving subjects treated with different short term therapeutic regimens also had high recurrence rates , it can be assumed that the treatment used in this series of cases may have little influence on the frequency of recurrences , but this should be assessed by further studies , also concerning the influence of inflammation on recurrences . [13 , 16 , 17 , 26] Some representative series of outbreaks , such as those occurred in Canada and India , found much lower rates of recurrence; this may be related to the strain of T . gondii involved and the particular characteristics of the population evaluated , such as the time of acquiring the disease and even a possible re-infection as the cause of recurrence . [14 , 27] The occurrence of severe vision impairment in the affected eye was observed in 40 ( 17 . 4% ) of 230 patients , a slightly lower incidence of blindness when compared to other series . [13] As expected , severe visual impairment was associated with the location of the retinochoroidal lesions , recurrence and with the occurrence of complications . The association of individuals with more than 40 years of age and complications suggests a tendency of increased disease severity in patients of this age group that requires further studies especially concerning recurrences in this population . The rate of retinal detachment , the overall rate of complications and vision impairment was slightly smaller than that previously described in other series , which may be eventually related with the treatment regimen used but also with the exclusion of non-confirmed severe cases that lack intra-ocular fluid examination . [13 , 28 , 29] Several studies in the literature consider the visual outcome and number of lesions as a measure to assess the effectiveness of therapeutic regimens and prevention strategies , and if this is done without taking into consideration the location of the lesion , erroneous conclusions may be drawn , as the location of the lesion is not treatment related . [6 , 30] Involvement of the otherwise healthy eye developed in only 9 ( 3 . 9% ) subjects who started the study with unilateral disease and can be considered a rare event . The frequency of bilateral retinochoroidal scars has been described in 22–44% of subjects in other series , and it is similar to the frequency found in our study . [1 , 13 , 19 , 22] In summary , recurrences were observed in 45 . 2% of the followed subjects and the severe vision impairment found in less than 20% of subjects , and it was associated with the location of the retinochoroidal scar , recurrences and posterior segment complications . High recurrence rates were observed after an active episode of TRC in this case series . Subclinical episodes in adults were observed in this studied population and can be a cause of underestimation of recurrences in retrospective studies . It is crucial to consider the location of the retinochoroidal lesion in studies analyzing the visual outcome as a measure of the effectiveness of treatment and prevention strategies . | Ocular toxoplasmosis affects millions of people worldwide and is a cause of severe vision loss . Prospective studies on the disease are rare and require long and expensive follow-ups . A network of intricate host and parasite factors can influence its evolution and its treatment is still subject of discussion and controversy . This article describes clinical features in a series of cases prospectively followed , confirming findings of previous studies conducted in other continents , despite the probable genetic variability of the evaluated populations and parasite . In addition , brings insights into the course of the disease and causes of visual impairment that may help future studies on strategies for treatment and prevention . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"toxoplasma",
"gondii",
"social",
"sciences",
"parasitic",
"diseases",
"neuroscience",
"parasitic",
"protozoans",
"protozoans",
"signs",
"and",
"symptoms",
"toxoplasma",
"pharmaceutics",
"vision",
"eyes",
"lesions",
"protozoan",
"infections",
"head",
"toxoplasmosis",
"psychology",
"visual",
"impairments",
"diagnostic",
"medicine",
"anatomy",
"ophthalmology",
"biology",
"and",
"life",
"sciences",
"ocular",
"system",
"sensory",
"perception",
"drug",
"therapy",
"organisms"
] | 2016 | Toxoplasmic Retinochoroiditis: Clinical Characteristics and Visual Outcome in a Prospective Study |
Productive infection by herpesviruses involve the disabling of host-cell intrinsic defenses by viral encoded tegument proteins . Epstein-Barr Virus ( EBV ) typically establishes a non-productive , latent infection and it remains unclear how it confronts the host-cell intrinsic defenses that restrict viral gene expression . Here , we show that the EBV major tegument protein BNRF1 targets host-cell intrinsic defense proteins and promotes viral early gene activation . Specifically , we demonstrate that BNRF1 interacts with the host nuclear protein Daxx at PML nuclear bodies ( PML-NBs ) and disrupts the formation of the Daxx-ATRX chromatin remodeling complex . We mapped the Daxx interaction domain on BNRF1 , and show that this domain is important for supporting EBV primary infection . Through reverse transcription PCR and infection assays , we show that BNRF1 supports viral gene expression upon early infection , and that this function is dependent on the Daxx-interaction domain . Lastly , we show that knockdown of Daxx and ATRX induces reactivation of EBV from latently infected lymphoblastoid cell lines ( LCLs ) , suggesting that Daxx and ATRX play a role in the regulation of viral chromatin . Taken together , our data demonstrate an important role of BNRF1 in supporting EBV early infection by interacting with Daxx and ATRX; and suggest that tegument disruption of PML-NB-associated antiviral resistances is a universal requirement for herpesvirus infection in the nucleus .
Epstein-Barr virus ( EBV ) is a member of the human gammaherpesvirus subfamily that infects over 90% of the global adult population [1] , [2] . EBV preferentially establishes latent infection in B-lymphocytes but can also infect epithelial cells [3] , [4] . EBV primary infection is one of the main causes of infectious mononucleosis ( IM ) ; while EBV latent infection is associated with multiple malignancies such as nasopharyngeal carcinoma , Burkitt's lymphoma , and Hodgkin's lymphoma [3] , [4] . Furthermore , EBV is responsible for the majority of lymphoproliferative diseases associated with AIDS and immunosuppression following organ transplant [5] . Like all herpesviruses , EBV exists in a dynamic balance between productive and latent infection . The factors that regulate the fate decisions for lytic reactivation from latency have been investigated in some detail , but relatively little is known about the fate regulation during the earliest stages of primary infection . Upon entry into the nuclear compartment , herpesvirus DNA genomes must confront several intrinsic anti-viral resistances that restrict viral gene expression and replication . One prominent nuclear structure involved in antiviral resistances is the PML nuclear body ( PML-NB ) , also referred to as nuclear domain 10 ( ND10 ) . PML-NBs are nucleoplasmic protein aggregates mainly consisting of ( but not limited to ) the components PML , Sp100 , Daxx , and ATRX [6] , [7] . The size and abundance of PML-NB is interferon inducible [8] , [9] , [10] , and over-expression of the PML protein represses viral infection [11] . PML-NB is the nuclear localization site of many DNA viruses , including Herpes Simplex virus ( HSV-1 ) , Human Cytomegalovirus ( HCMV ) and Adenovirus ( Ad5 ) [12] , [13] . These viruses then modify the morphology and/or protein composition of PML-NBs shortly after infection [12] , [14] . The mechanism of PML-NB-mediated antiviral repression is not clearly determined . PML , Sp100 , and Daxx are all associated with transcription repression , and this function may act on viral genomes [15] . Daxx can act as a transcription co-repressor of many cellular transcription factors [16] , [17] , [18] , [19] , and forms repressive transcription complexes with histone deacetylases ( HDACs ) [20] , [21] and DNA methyltransferase I ( DNMT I ) [22] , [23] . Daxx has been shown to induce heterochromatin markers on the HCMV genome and repress viral gene expression in a HDAC dependent manor [24] , [25] . Daxx also forms a chromatin-remodeling complex with ATRX [26] and both can form a repression complex at heterochromatin [27] . Furthermore , RNA interference ( RNAi ) studies have shown that knockdown of Daxx or ATRX can result in a higher infection level of HCMV [28] , [29] , [30] and also relieve the infection defect of mutant HSV deficient in disrupting PML-NB [31] . Herpesviruses confront intrinsic anti-viral resistances immediately upon entering the host cell nucleus , and therefore must counteract these resistances at the earliest possible time points to initiate viral gene expression . Herpesvirus tegument proteins , which are pre-packaged and delivered with the infectious virion , are strategically positioned to counteract the intrinsic anti-viral defenses and support the early steps of infection [32] . Both alpha- and beta- herpesviruses encode tegument proteins that regulate early events during lytic replication , including the disruption of the PML-NBs . HSV-1 immediate early gene ICP0 , disrupts PML-NB structure by degrading the core component PML [33] , [34] , [35] and eliminating SUMO-modified Sp100 [36]; while HCMV tegument protein pp71 displaces ATRX and subsequently degrades Daxx [24] , [30] . Both ICP0-deficient HSV-1 and pp71-deficient HCMV mutants are deficient in infection , where viral gene expression is shutdown , resulting in a dormant viral genome [29] , [37] , [38] . Interestingly , it has been reported that disruption of PML-NB by ICP0 is mediated by de novo synthesized ICP0 , instead of tegument delivered ICP0 protein , suggesting that this event is coordinated with early viral gene activation or , perhaps , reactivation from latent infection [35] . We have previously shown that EBV genomes localize to and then disrupt PML-NB during lytic replication; while latent EBV episomes are segregated away from PML-NBs during latency [39] . EBV regulatory proteins , including the lytic cycle immediate early gene Zta ( also referred to as BZLF1 , ZEBRA , and Z ) , and latency associated EBNA1 and EBNA-LP , have been implicated in PML-NB interactions [40] , [41] , [42] . However , it remains unclear if PML-NBs regulate early events associated with viral gene expression upon EBV nuclear entry , and if an EBV tegument protein modulates this intrinsic defense . The EBV major tegument protein BNRF1 is one of the most abundant tegument proteins in the virion [43] and is essential for the establishment of viral latent infection [44] , yet its function is largely unknown . BNRF1 homologues are present in all gammaherpesviruses but absent in the alpha- and beta- herpesvirus subfamilies . All BNRF1 orthologues share regions homologous to the cellular enzymes Phosphoribosylformylglycineamide Amidotransferase ( FGARAT ) and Aminoimidazole ribonucleotide ( AIR ) synthetase , ATP-dependent enzymes in the 4th and 5th steps of the purine de novo biosynthesis pathway . However , no enzymatic activity has been found in any BNRF1 orthologues . In a knockout study , transfected BNRF1-deficient EBV genomes can reactivate from latency , produce morphologically normal virions , and the progeny can enter cells with little observed defects [44] . Yet , upon infection of B cells the mutant virus showed a 20-fold lower expression of a viral latency associated gene EBNA2 and failed to induce B cell transformation [44] . This suggests an important role of BNRF1 in supporting early infection . Furthermore , the BNRF1 orthologue encoded by murine herpes virus 68 ( MHV68 ) , tegument protein ORF75c , induces PML degradation and is essential for initiation of viral gene expression [45] , [46] . Here , we demonstrate that EBV BNRF1 is a novel PML-NB-interacting viral protein , and that this interaction is important for supporting EBV primary infection . We first show that Daxx is a primary cellular interaction partner of BNRF1 . BNRF1 co-localizes with Daxx at PML-NB foci while disrupting the Daxx-ATRX complex . Furthermore , we identify a novel Daxx interaction domain on BNRF1 . This domain is essential for BNRF1 to interact with Daxx , localize to PML-NB , and displace ATRX from Daxx . We then show that BNRF1 supports EBV primary infection and promotes the expression of viral genes soon after viral genomes enter the cell , and that the Daxx interaction domain contributes to these functions . Lastly , we show that knockdown of either Daxx or ATRX results in disruption of viral latency , suggesting that Daxx and ATRX play a role in the restriction of viral gene expression . Our study suggests that EBV tegument protein BNRF1 disassemble the Daxx-ATRX antiviral resistance complex to enable viral gene expression after cell invasion , and likely regulate the chromatin organization for the establishment of latent infection .
To characterize the biological properties of the EBV major tegument protein BNRF1 , we took a proteomic approach to screen for potential cellular interaction partners . BNRF1 was cloned into a 3x FLAG tag expression vector under the control of a CMV promoter . 293T cells were then stably transfected with either FLAG-vector or FLAG-tagged BNRF1 . Nuclear extracts from stable cell lines were subject to immunopurification ( IP ) with a FLAG antibody , and then analyzed by SDS-PAGE ( Fig . 1A ) . Bands unique to the BNRF1 lane ( B ) were cut out and analyzed by liquid chromatography-tandem mass spectrometry ( LC/MS/MS ) . The major identified species was BNRF1 , but substoichiometric proteins enriched in the BNRF1 IP were also identified , including Daxx , nucleophosmin ( NPM1 ) , and PARP1 ( Fig . 1C ) . We subsequently confirmed in BNRF1 transiently transfected 293T cells that Daxx co-precipitates with BNRF1 ( Fig . 1B ) by both FLAG pull-down and the Daxx reverse pull-down , indicating a stable in vivo interaction between BNRF1 and Daxx . Neither PARP1 nor NPM1 interaction with BNRF1 could be validated by subsequent co-IPs ( data not shown ) , we therefore focused our efforts on characterizing the interaction with Daxx . To further characterize the interaction between BNRF1 and Daxx , we introduced serial deletions on the FLAG-BNRF1 expression plasmid . We first made five deletion constructs of BNRF1 , sequentially deleting regions coding for 300 amino acids ( Fig . 2A , constructs d1 through d5 ) . We then performed IPs with either control IgG , αFLAG , or αDaxx on lysates of cells transfected with the BNRF1 deletion constructs . Daxx co-precipitated in the FLAG IP for all of the BNRF1 mutants with the exception of the BNRF1 300–600 aa deletion mutant ( d2 ) ( Fig . 2B , middle panels ) . Similarly , all of the FLAG-BNRF1 mutants , with the exception of d2 , co-precipitated with Daxx IP ( Fig . 2B , right panels ) . Since d2 was expressed and recovered by FLAG IP to similar levels as other BNRF1 mutants capable of interacting with Daxx , we conclude that a putative Daxx-interaction domain is located in the region between 300-600aa of BNRF1 . We then further made six serial deletions of 60 amino acids in the 300–600 aa region ( Fig . 3A , constructs d21 through d26 ) to narrow down the suspected Daxx-interaction domain to a smaller region . After a subsequent round of IP pull-downs , we found that all BNRF1 deletions , with the exception of d21 , were defective in binding Daxx ( Fig . 3B ) , suggesting that the 360–600 aa region of BNRF1 is responsible for interaction with Daxx . To determine if this region was sufficient for interaction with Daxx , we expressed only the 300–600 aa region in the FLAG-expression vector ( Fig . 3A , construct DID ) and performed IP pull-downs . We found that this region bound Daxx as efficiently as WT-BNRF1 , in both FLAG IP and in the reverse IP with anti-Daxx antibody ( Fig . 3C ) . Notably , we failed to find sequence homology of this Daxx interaction domain with any known protein motif , and this domain is also distinct from the FGARAT and AIR synthetase homology regions . These findings suggests that BNRF1 utilizes a previously unknown motif to bind Daxx , and that the Daxx interaction domain ( 300–600 aa ) may contain a complex protein fold sensitive to smaller truncation deletions . Daxx forms a chromatin remodeling complex with ATRX [26] and ATRX has been implicated in the transcriptional repression of both HSV-1 and HCMV during the early steps of infection [30] . Moreover , both HSV-1 and HCMV utilize viral encoded proteins that disrupt the interaction between Daxx and ATRX [30] , [31] . To determine if BNRF1 also disrupted the interaction between Daxx and ATRX , we assayed the effect of WT and mutant BNRF1 proteins on the co-IP of Daxx with ATRX . We observed that WT BNRF1 disrupted the interaction between Daxx and ATRX ( Fig . 3C , 2nd panel from top , right ) . However , deletion mutants d22 and d26 , which fail to interact with Daxx , did not disrupt ATRX binding in Daxx IP assays ( Fig . 3C , 2nd panel from top , right ) . Interestingly , the Daxx interaction domain by itself ( DID ) , which binds Daxx efficiently , could only partially disrupt ATRX binding . This suggests that Daxx binding by BNRF1 is necessary , but not sufficient for the disruption of ATRX with Daxx . We also found no evidence that BNRF1 co-IPs with PML ( Fig . 3C , 3rd panel from top ) . To determine whether any other domains of BNRF1 contribute to the disruption of ATRX from Daxx , we assayed FLAG-BNRF1 IPs for ATRX binding using the set of larger BNRF1 deletions examined in Figure 2 ( Fig . 3D ) . We found that WT BNRF1 did not co-IP with ATRX , although it efficiently pulled down Daxx . The BNRF1 d2 mutant failed to pull down Daxx or ATRX , as expected . In contrast , the BNRF1 d3 and d4 mutants , which disrupts most of the FGARAT and AIR synthetase homology regions , efficiently pulled down both ATRX and Daxx . The d1 and d5 truncations , which lie outside of the FGARAT and AIR synthase homology regions , pulled down only Daxx but not ATRX , suggesting it efficiently disrupted the ATRX-Daxx interaction similar to WT . These data suggest that the FGARAT and AIR synthetase homology regions of BNRF1 may contribute to the disruption of ATRX-Daxx complex . Daxx is a prominent component of PML nuclear bodies [47] , and Daxx localization at these nuclear bodies are disrupted by viral proteins of both HSV-1 and HCMV [24] , [34] . Thus , it is important to investigate the sub-cellular location of BNRF1-Daxx interaction , and check if BNRF1 disrupts Daxx localization to the nuclear bodies . For immunofluorescence ( IF ) microscopy studies , we selected Hep2 carcinoma cell lines because of their larger size and prominent PML nuclear bodies , and their common use in many previous studies with herpesvirus protein interactions with PML-NBs . Hep2 cells were transiently transfected with empty FLAG vector ( V ) or BNRF1 constructs WT , d26 , or DID . Cells were then fixed two days post transfection and subject to IF staining . We found that WT BNRF1 partially co-localized with nuclear foci containing Daxx ( Fig . 4A , S1A and Table S1 ) , PML ( Fig . 4B , S1B and Table S1 ) , and Sp100 ( Fig . S1D ) , suggesting that BNRF1 interacts with Daxx at the PML nuclear bodies . We also noticed that DID itself is sufficient for localizing to PML nuclear bodies , while the d26 deletion mutant showed a weak dispersed pattern in the cell ( Fig . 4A and B , S1 and Table S1 ) . The co-localizations were also confirmed by line scan analysis , where the BNRF1 WT and DID intensity peaks overlap with Daxx and PML peaks ( Fig . S2 ) . To ensure that the diffuse pattern of the d26 mutant is not due to deficient protein expression , the same set of transfected cells as used for IF were also assayed by Western blot for total expression levels of BNRF1 proteins ( Fig . 4D ) . We found that d26 protein was expressed at levels similar to that of WT , despite its diffuse staining in IF studies , confirming its protein expression in the cells used in our microscopy study . These findings indicate that the interaction with Daxx is necessary and sufficient for BNRF1 to localize to the nuclear bodies . To understand the BNRF1 disruption of Daxx-ATRX complex in a sub-cellular spatial context , we also examined ATRX by IF in BNRF1-transfected Hep2 cells ( Fig . 4C and S1C ) . Again , we found ATRX foci co-localizing with WT BNRF1 and DID but not d26 , which is also confirmed by line scan analysis ( Fig . S2C ) . However , we also found a substantial reduction in ATRX foci intensity when cells were transfected with WT BNRF1 , but no apparent reduction when transfected with d26 or DID-mutants ( Fig . 4C ) . The failure of DID to disperse ATRX is consistent with its only partial disruption of ATRX from Daxx IP ( Fig . 3C ) . Quantification of Daxx ( Fig . 4E ) , PML ( Fig . 4F ) , and ATRX ( Fig . 4G ) nuclear foci in BNRF1-expressing cells compared to non-expressing cells revealed that BNRF1-expressing cells contain a significantly lower ( p<0 . 0001 ) average number of ATRX nuclear foci than non-expressing cells ( Fig . 4G ) . In contrast , we found no significant difference in the number of Daxx ( Fig . 4E ) and PML nuclear foci ( Fig . 4F ) in BNRF1 transfected cells . Taken together , these results suggest that BNRF1 not only disrupts the Daxx-ATRX complex , but also actively disperses ATRX away from nuclear bodies . HSV-1 ICP0 and HCMV pp71 each induce the degradation of PML and Daxx proteins respectively , yet we did not observe any evidence of this with BNRF1 in our microscopy studies . To investigate the potential degradation of PML , Daxx and ATRX proteins by BNRF1 , we examined the stability of these proteins in BNRF1 stably transfected cells ( Fig . 5A ) . 293T cells stably transfected with control vector ( clone C ) or WT BNRF1 ( stable transfection clones 3 and 9 ) were lysed and subject to Western blot analysis . We found no evidence of degradation or gross post-translational modification of PML , Daxx , nor ATRX in BNRF1-expressing cell lines . We also analyzed the protein stability of PML and Daxx in Hep2 cells transiently transfected with control vector , WT BNRF1 or d26 , and again found no evidence of BNRF1-induced protein degradation ( Fig . S3 ) . This suggests that BNRF1 does not mimic the protein degradation function of HCMV pp71 or HSV-1 ICP0 , but rather , disrupts Daxx-ATRX interactions through alternative mechanisms . The diffuse distribution of the BNRF1-d26 mutant raised the question of whether the Daxx interaction domain of BNRF1 correlated with nuclear localization . To test this , we utilized biochemical fractionation methods to isolate nuclear and cytoplasmic proteins from 293T cells ( Fig . 5B ) . We found that WT BNRF1 localized to both cytoplasmic ( ∼60% ) and nuclear ( 40% ) fractions . The d26 mutant , which is deficient in both Daxx interaction and nuclear bodies localization , was expressed at lower amounts yet showed a cytoplasmic to nuclear distribution similar to WT ( Fig . 5B ) . This is consistent with d26 having a weak diffuse nuclear and cytoplasmic staining in IF ( Fig . 4 ) . Meanwhile , the DID mutant , which binds Daxx and co-localizes with PML nuclear bodies in the nucleus , was isolated at low , yet detectable levels in the nucleus; although substantially more was recovered in the cytoplasmic fraction . The efficiency of the fractionation was confirmed by the presence of PARP1 exclusively in the nuclear fractions , and α-tubulin exclusively in the cytoplasm . These findings suggest that BNRF1 can localize to both cytoplasmic and nuclear compartments , and that the Daxx interaction domain might contribute partially to the nuclear entry or stability of BNRF1 . A previous study using an EBV bacmid with a BNRF1-knockout demonstrated that BNRF1-mutant virions can be generated from producer 293 cells and can enter the cytosol of infected B-cells; yet mutant virus failed to express one of the first expressed latent genes , EBNA2 , upon primary infection of B cells , and were incapable of inducing B-cell proliferation [44] . To understand the role of BNRF1-Daxx interaction in primary infection , we took a complementation rescue approach with the BNRF1-mutant virus . 293 cells stably transfected with either wild type or BNRF1-knockout EBV bacmids were used for virus production ( Fig . 6 ) . As the EBV bacmids also encode GFP , cells infected with this bacmid-derived virus could be visualized by the presence of green fluorescence . To induce viral production , bacmid containing cells were co-transfected with the EBV transactivator Zta and BALF4 . To complement for BNRF1 deletion , production cells were also transfected with either control vector , WT BNRF1 ( WT ) , or the BNRF1 deletion mutant ( d26 ) which fails to interact with Daxx . Three days after transfection , the media was collected and used to infect primary B cells isolated from human peripheral blood mononuclear cells ( PBMCs ) . We detected high-levels of GFP positive proliferating B-cell clusters when infected with virus generated from wild type bacmid ( Fig . 6Ai ) , but no GFP positive or clumped cells were detected when infected with no virus ( Fig . 6Aii ) or virus from un-complemented ΔBNRF1 bacmids ( Fig . 6Aiii ) . However , when ΔBNRF1 virus was complimented with WT BNRF1 we were able to detect GFP positive cells and proliferating B-cell clusters ( Fig . 6Aiv ) . Notably , ΔBNRF1 virus complimented with the d26 mutant BNRF1 failed to express GFP or induce B cell proliferation ( Fig . 6Av ) , showing a similar defect as ΔBNRF1 virus with no complementation . Quantification of at least three independent infections confirmed that GFP positive and proliferating B-cells were detectable only when ΔBNRF1 bacmid virus was complemented with WT , but not with d26 mutant BNRF1 ( Fig . 6B ) . To ensure the infections between each complemented virus were comparable , virus titer was quantified by real time PCR for virion DNA . The viral titers of either empty vector or WT BNRF1 complemented virus was found to be similar , while some reduction in virus titer was observed with d26 virus ( Fig . 6C ) . We also tested by Western blotting for incorporation of FLAG-BNRF1 proteins in virions , and found that WT and d26 mutant BNRF1 proteins were both packaged into virions to similar per particle levels ( Fig . 6D ) . These findings confirm that BNRF1 is required for primary infection of B-cells , and suggests that the Daxx interaction domain of BNRF1 is important for this function . Other herpesvirus tegument proteins that interact with Daxx and ATRX have been shown to function in the transcription activation of viral genes during primary infection [25] . To investigate the role of BNRF1 on viral gene transcription early after primary infection , we infected human B-lymphocytes purified from PBMCs with the ΔBNRF1 virus complemented with empty FLAG-vector , WT BNRF1 or d26 mutant BNRF1 ( Fig . 7A ) . Viral gene expression in these newly infected cells was assayed at four days post infection using Reverse Transcription qPCR ( RT-qPCR ) . We found that WT BNRF1 complementation induced an up-regulation of EBNA1 , EBNA2 and BZLF1 mRNA expression compared with non-complemented virus or the d26-mutant complementation . Interestingly , background levels of BZLF1 expression were detectable in non-complemented and d26 mutant infections , suggesting that BNRF1 may only partly enhance BZLF1 expression , which can occur at low levels independently of BNRF1 . To investigate the potential mechanism of BNRF1 in viral gene regulation , we first tested the effect of BNRF1 on reporter plasmids using transient transfection assays , but found no consistent effect on candidate viral promoters ( data not shown ) . We reasoned that reporter plasmids may lack essential BNRF1 target elements or chromatin assembly , and therefore assayed BNRF1 activity on EBV bacmid genomes after transfection into 293 cells ( Fig . 7B ) . EBV bacmid DNA ( Bac36 ) and either empty FLAG-vector , WT BNRF1 , or the d26 mutant BNRF1 were co-transfected into 293 cells and assayed 3 days post transfection for viral gene expression using RT-qPCR . We found that WT BNRF1 promoted a robust expression of BZLF1 transcripts ( ∼20 fold ) , which was not observed in vector control or the d26 mutant ( Fig . 7B ) . BNRF1 also increased EBNA2 mRNA ( ∼3 fold ) relative to vector control , but this was not significantly increased relative to that of the d26 mutant . These studies suggest that BNRF1 can activate the expression of the EBV immediate early gene BZLF1 in the context of the viral genome , and in the absence of other virion-delivered tegument proteins . The previous experiments suggest that BNRF1 can function during tegument delivery in early infection , as well as after de novo synthesis , perhaps regulating the transition from latent to lytic infection . To explore the role of Daxx and ATRX in the context of EBV latent to lytic gene regulation , we test the effects of Daxx and ATRX knockdown on viral lytic gene expression in Mutu I cells , an EBV-latently infected Burkitt's lymphoma cell line ( Fig . 8 ) . Mutu I cells were transduced with puromycin resistant lentivirus carrying either non-targeting shRNA ( shNeg ) , shRNA against Daxx ( shDaxx ) , ATRX ( shATRX ) , or ZEB1 ( shZEB1 . 1 ) which acts as a positive control for reactivation . ZEB1 has been shown to repress Zta expression , and shRNA depletion of ZEB1 can reactivate lytic gene expression in several cell types [48] , [49] , [50] . Mutu I cells were harvested 9 days after shRNA transduction and selection , and then tested for viral reactivation by Western blot and FACS ( Fig 8 ) . Western blot analysis of whole cell lysates ( Fig . 8A and B ) revealed that knockdown of either Daxx or ATRX induced a reactivation of EBV early antigens , as shown by increased band intensities of both the immediate early gene Zta ( 2-fold ) and the lytic early antigen EA-D ( 3-fold ) . These induction levels are comparable to that observed with the shZEB1 . 1 positive control . The efficiency of shRNA-mediated knockdown was confirmed by the loss of Daxx , ATRX and ZEB1 bands in the corresponding lanes ( Fig . 8A ) . We also verified reactivation by flow cytometry quantification of the EBV viral capsid antigen VCA on cells from three independent shRNA-treatments ( Fig . 8C ) , where we observed an approximately 6-to-10-fold induction by either Daxx or ATRX depletion . These findings indicate that the depletion of either Daxx or ATRX can promote viral lytic gene expression from latently infected B-cells , and suggest that BNRF1 disruption of the Daxx-ATRX complex contributes to viral gene control during early infection and reactivation .
The specific class of antiviral defense dubbed the intrinsic immunity [51] , [52] plays a broad and general role in restricting viral infection . PML-NBs and its associated proteins such as PML , Sp100 , Daxx and ATRX , have been extensively studied as cellular defenses against herpesviruses , specifically with the alphaherpesvirus HSV-1 and betaherpesvirus HCMV . Upon the early stages of infection right after cell entry , HSV-1 and HCMV utilize viral proteins that effectively disrupt the structure and disable the function of the PML-NBs in restricting viral gene expression and replication . However , the gammaherpesvirus EBV has been relatively less studied in terms of how it counteracts these cellular resistances upon primary infection or reactivation . We show here that the major tegument protein of EBV , BNRF1 , interacts with Daxx ( Figs . 1–3 , S1–2 ) and disrupts its ability to form a complex with ATRX or recruit ATRX to PML-NBs ( Figs . 3–4 , S1–2 ) . Moreover , we show that BNRF1 functionally promotes viral early gene expression with a preference for the activation of the immediate early gene BZLF1 , and to a lesser extent the latent activator EBNA2 ( Figs . 6–7 ) . These findings indicate that EBV , like its relatives HSV1 and HCMV , encodes a viral tegument protein that targets PML-NB components to promote viral gene expression . Daxx is a prominent PML-NB component , but is also associated with a diverse , yet non-mutually exclusive variety of cellular functions , including the regulation of apoptosis , chromatin remodeling , gene repression , and antiviral resistance [47] , [53] . Daxx is a primary target of the HCMV pp71 protein , which both binds and induces the degradation of Daxx [30] . Like HCMV pp71 , BNRF1 binds Daxx and prevents the Daxx-interaction partner , ATRX , from associating with Daxx and localizing to PML-NBs . BNRF1 and pp71 are both tegument proteins , whose pre-made nature likely provides them with a temporal advantage to disarm cellular repression machinery without the prior need of viral gene transcription . However , unlike HCMV pp71 , BNRF1 does not induce Daxx degradation , which remains prominently associated with PML-NBs when BNRF1 is expressed ( Figs 2–5 ) . BNRF1 and pp71 share no obvious amino acid sequence similarity , and the Daxx interaction domains of these two proteins vary significantly in amino acid composition and size of the interaction domains . These findings suggest that BNRF1 is a functional homologue of pp71 , but utilizes a distinct mechanism for the dissociation of ATRX from PML-NBs . Herpesvirus tegument proteins have been implicated in the determination of viral lytic or latent gene expression programs . Restriction of tegument protein entry into the nucleus , as has been shown for HCMV pp71 and HSV VP16 , correlates with the establishment of latency [24] , suggesting that tegument proteins may play a critical role in determining lytic or latent gene expression programs . Interestingly , we found that the Daxx-interaction deficient BNRF1 mutant d26 , which fails to interact with Daxx , showed a weak diffuse subcellular distribution instead of the punctate nuclear dots of WT BNRF1 ( Fig . 4 and S1 ) . Similarly , biochemical fractionation studies ( Fig . 5B ) suggest that while BNRF1 can localize to both the cytoplasm and nucleus , it may require the Daxx-interaction domain to efficiently accumulate in the nucleus . Potentially related is the observation that HCMV pp71 translocation to PML-NBs is also dependent on its interaction with Daxx [54] . Selective cytoplasmic retention of several herpesvirus tegument proteins , including pp71 and VP16 , may play a critical role in determining lytic or latent gene expression programs [55] . We suspect that BNRF1 might be subject to similar regulation through its PML-NB localization . Daxx and ATRX are known to play a global role in the control of cellular and viral gene expression and chromosomal structure . Daxx itself has been shown to associate with HDACs and to function as a global repressor of transcription [20] , [21] . The Daxx-ATRX complex has in vitro chromatin remodeling activities [26] and can function as a histone H3 . 3 chaperone [56] . Recent studies suggest also that ATRX interacts with G-rich repeat chromatin regions [57] , and in collaboration with Daxx load histone variant H3 . 3 onto pericentromeric and telomeric chromatin [56] , [58] , [59] . H3 . 3 is generally associated with open chromatin and active transcription when loaded by the histone chaperone HIRA [60] . However , the Daxx-ATRX complex loaded H3 . 3 has been found to facilitate transcription from pericentromeric regions [58] but repress transcription from telomeric regions [59] . Interestingly , HIRA-loaded H3 . 3 can facilitate the lytic replication of HSV-1 during the early steps of infection [61] . Furthermore , the Daxx-degrading pp71 blocks the establishment of heterochromatin on the HCMV Major Immediate Early Promoter ( MIEP ) region [25] . These findings underscore the importance of host chromatin regulatory mechanisms in the control of herpesvirus infection . We suspect that the viral gene activation function of BNRF1 ( Fig . 7 ) is likely to be mediated by chromatin-dependent processes since we failed to observe consistent transcription activation when assayed in transient plasmid-based reporter assays using EBV promoters for BZLF1 ( Zp ) or EBNA2 ( Cp or Wp ) ( data not shown ) . We propose that BNRF1 stimulates EBV early gene activity through de-repression of the Daxx-ATRX mediated chromatin repression mechanism , perhaps similar to that of pp71 de-repression of the HCMV MIEP locus . However , the precise molecular mechanism through which BNRF1 activates early gene transcription through the disruption of ATRX-Daxx interaction remains to be investigated . While not explored yet , it is also not known if the FGARAT enzyme-homology domain of BNRF1 has any function in the context of supporting viral infection . This enzyme homology is conserved among all gammaherpesvirus orthologues of BNRF1 , including the KSHV and MHV68 ORF75 family members . Despite significant sequence similarity with BNRF1 , KSHV and MHV68 ORF75 proteins do not appear to interact with Daxx ( data not shown ) . However , MHV68 ORF75c targets PML-NBs through the degradation of PML [45] , [46] , an activity that we did not observe with BNRF1 . Thus , while these tegument family members share the FGARAT homology regions , and may similarly target components of the PML-NBs , they appear to target different proteins and utilize distinct mechanisms . It is also important to note that the disruption of ATRX by BNRF1 was partially dependent on the FGARAT domain , since the DID alone , which binds Daxx efficiently , only partially disrupt ATRX binding in IP assays ( Figs 2 and 3 ) while not causing any significant ATRX dispersion from PML-NBs in IF assays ( Fig . 4 ) . Also , deletions within the FGARAT domain ( d3 and d4 ) resulted in a mutant BNRF1 that co-precipitated with ATRX , creating a gain of function not seen with WT BNRF1 . All of this suggests that the FGARAT domain may play a regulatory role in BNRF1 interactions with Daxx and ATRX . In conclusion , our data demonstrates a novel example of herpesvirus tegument protein interacting with components of the cellular antiviral resistance . BNRF1 interaction with Daxx may provide several functions , including the establishment of a chromatin structure conducive to viral early gene activity . Our findings demonstrate that EBV , like other herpesviruses , confront the PML-NB associated intrinsic defenses through a viral factor that is available and active upon the early stages of infection , and shed light into the critical control mechanisms that govern the early events of EBV infection before the establishment of latency .
Human B-lymphocytes were obtained from the Wistar Institute phlebotomy lab . All samples were from anonymous adult donors and approved by the Wistar Institute Institutional Review Board . Written informed consent was provided by study participants . Hep2 and 293T cells were grown in Dulbecco's modified Eagle medium ( DMEM ) supplemented with 10% fetal bovine serum ( FBS ) , 20 mM GlutaMAX ( Gibco ) , 100 U/ml penicillin and 100 µl/ml streptomycin . 293HEK cells were grown in minimum essential medium Eagle ( MEM ) , supplemented with 10% FBS and 20 mM GlutaMAX ( Gibco ) . DG75 and Mutu I cells are EBV negative and positive ( respectively ) Burkitt's lymphoma cell lines , grown in RPMI 1640 medium supplemented with 10% FBS , 100 U/ml penicillin and 100 µl/ml streptomycin . Peripheral blood mononuclear cells ( PBMCs ) were isolated from fresh donated human blood by density gradient centrifugation with Ficoll-Paque Plus purchased from GE healthcare . Primary B cells were then isolated from PBMCs using Dynabeads Untouched Human B Cell isolation kit ( Invitrogen ) . All cells were grown in a 5% CO2 incubator at 37°C . Stable 293 cell lines expressing FLAG-BNRF1 ( clone 3 and clone 9 ) and empty FLAG vector ( clone C ) were grown in DMEM as described for 293T cells above , supplemented with 2 . 5 µg/ml Puromycin for selection . Viruses were produced using chloramphenicol and hygromycin resistant bacmids containing the EBV genome and the gene coding for green fluorescence protein ( GFP ) . 293/EBV-wt cells ( a gift from H . J . Delecluse ) are 293HEK cells stably transfected with the wild type EBV bacmid [62] . 293/△BNRF1 cells ( a gift from H . J . Delecluse ) are 293HEK cells stably transfected with an EBV bacmid with the BNRF1 gene deleted [44] . 293/EBV-wt and 293/△BNRF1 cells were grown in RPMI 1640 medium supplemented with 10% FBS and 100 µg/ml hygromycin . All restriction enzymes , T4 DNA ligase and associated buffers were purchased from New England Biolabs . Monoclonal mouse anti-FLAG antibody ( F1804 ) , Polyclonal rabbit anti-FLAG anibody ( F7425 ) , Polyclonal rabbit anti-Daxx antibody ( F7810 ) , Monoclonal Anti-α-Tubulin antibody ( T5168 ) , Monoclonal mouse anti-β-Actin-Peroxidase antibody ( A3854 ) , and Anti-mouse IgG R-Phycoerythrin ( PE ) conjugated antibody ( P8547 ) were purchased from Sigma-Aldrich . Monoclonal mice anti-PML ( PG-M3 , sc-966 ) , and polyclonal rabbit anti-ATRX ( H-300 , sc-15408 ) , and polyclonal rabbit anti-ZEB1 ( sc25388 ) were purchased from Santa Cruz Biotechnology . Polyclonal rabbit anti-PARP1 antibody ( ALX-210-895-R100 ) was purchased from Enzo Life Sciences . Mouse anti-EA-D antibody was purchased from Millipore . Anti-EBV-VCA ( 0231 ) antibody was purchased from Pierce Thermo Scientific . BNRF1 was cloned into the HindIII-SalI sites of the p3xFLAG-Myc-CMV-24 Expression Vector ( Sigma-Aldrich ) , using the PCR primers: gcgaagcttgaagagaggggcagggaaacgcaa and gcggtcgactcactcggaggggcgaccgtgcctg . BNRF1 deletion mutants were generated as follows . PCR Primers ( Table S2 ) were designed so that the front and rear halves of the DNA oligo each binds the 5′ or 3′ regions flanking the targeted deletion site on the BNRF1 template . PCR reactions were setup using iProof High-Fidelity DNA polymerase 2x master mix ( Bio-Rad ) , with primers at 1 µM concentration , and the FLAG-BNRF1 expression plasmid as the template at a concentration of 50 ng DNA in a 25 µl reaction setup . PCR was done with a Bio-Rad C1000 thermal cycler , thermal cycles setup according to DNA polymerase mix manufacturer suggested conditions . To clear out the wild type BNRF1 template , 15 µl of the PCR product were treated with 30 U DpnI ( New England Biolabs ) in a 20 µl reaction for 2 hours to over night at 37°C . 2 µl of DpnI-treated DNA were then transformed into 50 µl of Library Efficiency DH5α competent cells ( Invitrogen ) . Colonies were screened for the deletion by enzyme digestion analysis of miniprep DNA , and then confirmed by DNA sequencing of the expected deletion site . BNRF1 expression plasmids were transfected using Lipofectamine 2000 ( Invitrogen ) according to manufacturer instructions . Cells were harvested 2 days post transfection by washing cells off the plate with PBS . Harvested cells were further washed 3 times with cold PBS , and then subject to lysis with freshly prepared NET lysis buffer ( 50 mM Tris-HCl pH 7 . 5 , 150 mM NaCl , 5 mM EDTA , 0 . 5% NP-40 , and 0 . 1% mammalian protease inhibitor cocktail mix ( P8340 , Sigma-Aldrich ) , at 1 ml NET per IP pull-down . Cell lysates were homogenized by doing 10 strokes in a Dounce homogenizer . 60 µl of each lysate were isolated after this step as input control . The remaining lysates were incubated at 4°C rotating for 30 mins to fully solubilize proteins . Lysates were then spun at 13000 rpm 5 mins to remove insoluble cell debris , then antibodies were added ( 5 µl of each antibody per IP ) to the cleared lysates , and left rotating over night . 100 µl of 50% slurry of Protein A sepharose beads ( GE healthcare ) in NET buffer was added to each IP with rotating at 4°C for 2–3 hours , then washed three times with NET for 10 mins ( rotating at 4°C ) per wash . Pulled down proteins were released by adding 50 µl 2x Laemmli buffer ( 100 mM Tris-Cl pH 6 . 8 , 4% SDS , 0 . 2% Bromophenol Blue , 20% Glycerol ) , and boiling for 10 mins at 100°C . The resulting samples ( excluding beads ) were then loaded directly into protein gels and subject to Western blot analysis . For mass spectrometry identification of BNRF1 associated proteins , FLAG-BNRF1 expressing and FLAG-vector control stable cell lines were generated as mentioned above . Nuclear extracts from 5×107 cells were subject to immunopurification with anti-FLAG Sepharose beads ( A2220 , Sigma-Aldrich ) followed extensive washing with NET buffer , and FLAG peptide elution . Eluted protein was subject to precipitation with 10% trichloroacetic acid ( TCA ) followed by SDS-PAGE and colloidal blue staining . Sections of the gel with enriched polypeptides were subject to LC/MS/MS at the Wistar Proteomics Facility . Hep2 cells were transfected with BNRF1 expression plasmids using Lipofectamine 2000 ( Invitrogen ) according to the manufacturer's instructions , transfected cells were then reseeded at 2 . 7×104 cells/well in 24 well plates containing microscope coverslips 5 hours post transfection . 2 days post transfection , coverslips with cells attached were harvested , fixed with 1% paraformaldehyde at room temperature for 15 mins , then permeablized with 0 . 3% Triton-X 100 . Coverslips were then stained with the first antibodies over night at 4°C . First antibody dilutions used were as follows: mouse anti-FLAG at 1∶20000 , rabbit anti-Daxx at 1∶5000 , rabbit anti-FLAG at 1∶5000 , mouse anti-PML at 1∶250 , rabbit anti-ATRX at 1∶250 , all antibodies diluted in PBS . Second antibody stainings were carried out for 1 hour at room temperature with the red-fluorescent Alexafluor594 goat anti-rabbit antibody and green Alexafluor488 goat anti-mouse antibody ( both from Invitrogen ) each at 1/800 dilution in PBS . Coverslips were washed twice in PBS for 5 mins between each of the above treatments . Cell nuclei were stained briefly with DAPI ( diluted to a final concentration of 0 . 167 µg/ml in PBS ) for 2 mins , then washed with PBS , 70% EtOH , then 100% EtOH to wash out residual salts . Coverslips were air-dried briefly , and then mounted onto microscope slides with Vectasheld mounting media ( Vector Laboratories ) . Mounted slides were examined under a Nikon E600 upright microscope with a 100x oil objective . Photos for nuclear body quantification were took using a 40x objective to maximize the number of cells in each photo while retaining a clear view of PML bodies . Microscopy photos were analyzed using ImagePro Plus 6 . 2 software ( Media Cybernetics ) . Photos were pre-processed by subtracting out the background intensity using the operation function ( with a value of −30 ) , and passing through a flatten filter ( a value of 10 ) . A morphological ‘top hat’ filter was then applied to emphasize points or grains brighter then the background . The number of nuclear bodies in each cell nucleus was counted by quantifying the object numbers after applying the signal intensity threshold/segmentation tool to select the nuclear bodies as objects . Cell boundaries were defined by the outline from DAPI channel photos of the same field , while omitting all cells on the border of the image border . Resulting quantification numbers were then analyzed using Prism 4 software ( Graph Pad Software ) , statistical analysis did by Mann-Whitney U non-parametric , unpaired t test . 293T cells were transfected in 10 cm plates with 2 µg expression plasmids of either empty FLAG vector , WT-BNRF1 , BNRF1-DID , or 6 µg of BNRF1-d26 . Transfection was carried out using 10 µl Lipofectamine 2000 ( Invitrogen ) per transfection , following manufacturer instructions . Cells were harvested 24 hrs post transfection . 1/6 of cells isolated as input control . The rest of the cell pellets were fractionated with the Fermentas ProteoJET Cytoplasmic and Nuclear Protein Extraction Kit ( K0311 ) . The resulting cytoplasmic and nuclear fractions , along with the input samples , were analyzed by Western blot . To induce lytic virus production , 293/EBV-wt and 293/△BNRF1 cells were transfected in 10 cm plates with expression plasmids of 1 . 75 µg BALF4 , 3 . 25 µg BZLF1 or cDNA3 empty vector , and 3 µg of either empty FLAG vector or 3 µg BNRF1 or 7 . 5 µg BNRF1-d26 . Transfection was carried out using 15 µl Lipofectamine 2000 ( Invitrogen ) per transfection , following manufacturer instructions . The media of virus production cells were harvested 3 days post transfection , filtered through 0 . 45 µm filters , and added directly to freshly isolated primary B cells . B cells in virus containing media were centrifuged for 1200 rpm 90 mins at 25°C to enhance infection . For measuring infection by GFP levels , infected B cells were treated with 1 mM Sodium Butyrate and 20 ng/ml TPA 3 days post transfection to enhance GFP expression , and the number of GFP positive cells in each well were counted manually under a Nikon TE2000 microscope using a 20x objective . For measuring virus gene expression in infected B cells , cells were collected 4 days post transfection , and total RNA was purified using Trizol ( Invitrogen ) . The resulting RNA was then subject to DNase 1 treatment at 2 U/50 µl , 1 hour at 37°C , then DNase was heat inactivated by adding a final concentration of 5 mM EDTA and incubated at 70°C for 10 mins . cDNA was synthesized using the Super Script III first strand synthesis system reverse-transcription kit ( Invitrogen ) . The resulting cDNA was then subject to real time PCR analysis by ΔCt method and normalized to viral titers , measured as described below . Real time PCR primers used are listed in Table S3 . To measure the amount of complemented BNRF1 protein that were packaged into virions , △BNRF1 virions complemented with WT-BNRF1 , BNRF1-d26 , or un-complemented , were produced as mentioned above . 100 µl of the harvested and filtered virus-containing media were isolated for viral titer measurement as described below . The rest of the virus-containing media were concentrated by loading the media above a 5 ml layer of 22 . 5% sucrose in PBS , then centrifuged at 27000 rpm ( ∼100 , 000 g ) 4°C for 1 hour in a SureSpin 630 Rotor ( Thermo Scientific ) with a Sorvall WX 100 Ultra ultracentrifuge . The resulting virus pellet was then resuspended in PBS , and analyzed by Western blot . Protein gel loading volumes were normalized according to viral titers to ensure equal amounts of virion protein in each well . Viral DNA in media was extracted as described by C . Busse et al . [63] . Virus-containing media were treated with 5 U/50 µl of DNase I ( New England Biolabs ) for 1 hour at 37°C . DNase was then deactivated by adding EDTA to a final concentration of 5 mM , followed by 10 mins heat inactivation at 70°C . Samples were then mixed 1∶1 with 0 . 1 mg/ml of proteinase K in water , and incubated at 50°C for 1 hour , followed by 20 mins of heat inactivation at 75°C . The released viral DNA was measured by real time PCR analysis , using a serial dilution series of Namalwa cell lysate as the standard curve , which contain two copies of integrated EBV genome per Namalwa cell . EBV genomes were detected using primers specific to the OriLyt region: 5′- CGTCTTACTGCCCAGCCTACTC-3′ ( OriLyt-fwd ) , 5′- AGTGGGAGGGCAGGAAATG-3′ ( OriLyt-rev ) . Wild type EBV genome bacmids were prepared from 2 . 5 mls overnight LB culture using the Bacmax DNA purification kit ( Epicentre ) . 293HEK cells were seeded ( 2 . 3 million cells per plate ) the previous day in 10 cm plates , and transfected with 1 . 5 µg freshly prepared bacmids along with 0 . 5 µg of either empty FLAG vector , BNRF1 , or BNRF1-d26 mutant . Transfection was carried out using Effectene transfection reagents ( Qiagen ) , following manufacturer instructions . Cells were harvested three days post transfection , total RNA was purified using Trizol ( Invitrogen ) , and then subject to DNase 1 treatment at 2 U/50 µl , 1 hour at 37°C . DNase was heat inactivated by adding a final concentration of 5 mM EDTA and incubated at 70°C for 10 mins . cDNA was then synthesized using the Super Script III first strand synthesis system reverse-transcription kit ( Invitrogen ) . The resulting cDNA was then subject to real time PCR analysis by ΔCt method . Real time PCR primers used are listed in Table S3 . shNeg ( pLKO-shNeg ) , shDaxx ( pLKO-shDaxx-2 ) and shATRX ( pLKO-shATRX90 ) constructs in lentivirus production plasmid backbones were generous gifts from Roger Everett . shNeg ( sequence TTATCGCGCATATCACGCG ) was designed to poorly target the E . coli DNA polymerase and extensively screened to ensure that it does not affect human nor viral transcripts . Use of shDaxx and shATRX was previously described else where [30] , [31] . shZEB1 . 1 was obtained from the TRC library ( Sigma , Inc ) , with targeting sequence GCAACAATACAAGAGGTTAAACTCGAGTTTAACCTCTTGTATTGTTGC ) . Mutu I cells were infected with lentiviruses carrying pLKO . 1-puro vectors by spin-infection at 400 g for 45 minutes at room temperature . The pellets were resuspended in fresh medium and left growing overnight . The RPMI medium was replaced each day , with 2 . 5 ug/ml Puromycin added for selection for lentivirus transduced cells . The cells were collected after 9 days of puromycin selection , and subject to Flow cytometry quantification of EBV viral capsid antigen positive cells , and Western blot analysis . | Persistent infection by Epstein-Barr virus ( EBV ) is associated with a variety of diseases , including lymphoid and epithelial tumors . Despite a wealth of information on the mechanism of viral persistence , relatively little is known about the early steps of EBV infection and viral gene activation . Host cells actively mount resistances against viral infection , which viruses need to overcome to invade the cell . We have found that among the proteins packaged in the EBV viral particle , BNRF1 plays an important role of counteracting cellular defenses . We show that EBV protein BNRF1 binds to the cellular protein Daxx and disassembles the Daxx-ATRX complex , where both Daxx and ATRX are cellular proteins known to inhibit viral gene expression . We also confirm that BNRF1 can promote expression of early viral genes , and that Daxx-binding by BNRF1 is required for this function . Finally , we demonstrate that Daxx and ATRX repress viral gene expression during latency . We conclude that BNRF1 disassembles cellular antiviral defense machinery to promote expression of viral genes in the host cell . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"medicine",
"biology"
] | 2011 | EBV Tegument Protein BNRF1 Disrupts DAXX-ATRX to Activate Viral Early Gene Transcription |
Electrical activity plays a pivotal role in glucose-stimulated insulin secretion from pancreatic -cells . Recent findings have shown that the electrophysiological characteristics of human -cells differ from their rodent counterparts . We show that the electrophysiological responses in human -cells to a range of ion channels antagonists are heterogeneous . In some cells , inhibition of small-conductance potassium currents has no effect on action potential firing , while it increases the firing frequency dramatically in other cells . Sodium channel block can sometimes reduce action potential amplitude , sometimes abolish electrical activity , and in some cells even change spiking electrical activity to rapid bursting . We show that , in contrast to L-type -channels , P/Q-type -currents are not necessary for action potential generation , and , surprisingly , a P/Q-type -channel antagonist even accelerates action potential firing . By including SK-channels and dynamics in a previous mathematical model of electrical activity in human -cells , we investigate the heterogeneous and nonintuitive electrophysiological responses to ion channel antagonists , and use our findings to obtain insight in previously published insulin secretion measurements . Using our model we also study paracrine signals , and simulate slow oscillations by adding a glycolytic oscillatory component to the electrophysiological model . The heterogenous electrophysiological responses in human -cells must be taken into account for a deeper understanding of the mechanisms underlying insulin secretion in health and disease , and as shown here , the interdisciplinary combination of experiments and modeling increases our understanding of human -cell physiology .
Glucose-stimulated insulin secretion from human pancreatic -cells relies on the same major signaling cascade as their rodent counterparts , with electrical activity playing a pivotal role . Following metabolism of the sugar , ATP-sensitive potassium channels ( K ( ATP ) -channels ) close in response to the elevated ATP/ADP-ratio , which triggers action potential firing and -influx through voltage-gated calcium channels . The resulting increase in intracellular calcium leads to insulin release by -dependent exocytosis [1]–[4] . However , the electrophysiological properties of human and rodent -cells show important differences , e . g . , with respect to their palette of expressed -channels and the role of -channels , which contribute to electrical activity in human but not in rodent -cells [1] , [3] . Mathematical modeling has played important roles in studying the dynamics of electrical activity in rodent -cells [5] , [6] , and could plausibly aid in understanding the electrophysiological responses in human -cells , and how they might differ from rodent cells . Recently , the first model of electrical activity in human -cells [7] was constructed from careful biophysical characterizations of ion channels in human -cells , mainly from Braun et al . [3] . The model [7] included -channels , three types of -channels , an unspecified leak-current , and several -channels: delayed rectifier ( Kv ) -channels , large-conductance ( BK ) -sensitive -channels , human ether-a-go-go ( HERG ) -channels as well as K ( ATP ) -channels . Recently evidence for small conductance ( SK ) -sensitive -channels in human -cells was published [4] , [8] , a current not included in the mathematical model [7] . The model [7] was shown to reproduce , depending on parameter values , spiking or rapid bursting electrical activity , which could be modified in accordance with a series of experiments by simulating pharmacological interventions such as ion channel blocking . These experiments were in general straightforward to interpret , also without a model . For example , the facts that blocking depolarizing - or -currents slowed or abolished electrical activity [3] are as one would expect . Here , we extend the previous model for human -cells [7] by including SK-channels and dynamics , and show that the model now has reached a level of maturity that allows us to get insight in less intuitive experimental findings . We find experimentally that SK-channels in some cells play an important role in controlling electrical activity , while they have virtually no effect in other cells . Using the extended version of the model , we show that this difference can be explained by differences in the excitability of the cells . Moreover we find that SK-channels can substitute for HERG-channels in controlling rapid bursting . We also show that blocking -channels in some cells can transform spiking behavior into rapid bursting , in contrast to the usual effect of -channel blockers , which in general reduce or abolish spiking behavior [3] , [9] , [10] . Using our model we suggest that this happens in cells with a large -current and that BK-channels play a prominent role . In addition , we suggest that SK-channels might underlie the surprising result that blocking depolarizing P/Q-type -channels enhances electrical activity , in contrast to the effect of L- or T-type -channel antagonists , which reduce excitability and electrical activity [3] . Our model is then used to investigate paracrine effects of -aminobutyric acid ( GABA ) and muscarinic signaling on electrical activity . Finally , we show experimentally slow oscillations in electrical activity that might underlie pulsatile insulin secretion from human pancreatic islets , and by adding an oscillatory glycolytic component [11] to the electrophysiological model , we simulate such slow bursting patterns .
When stimulated by glucose , human -cells show electrical activity [1] , [3] . Human -cells express SK-channels [4] , [8] , which might participate in controlling electrical activity . To study the role of SK-channels in human -cells , we included SK-channels and dynamics in our previous model [7] . The new model with standard parameters produces spiking electrical activity ( Fig . 1A ) , which is virtually unaffected by setting the SK-conductance nS/pF simulating SK-channels block . This model prediction was confirmed by our experimental data , and was also observed in at least one cell by Jacobson et al . [8] . Fig . 1B shows an example of spiking electrical activity in a human -cell stimulated by 6 mM glucose , where addition of the SK1-3 channel blocker UCL 1684 ( 0 . 2 µM ) did not affect the spiking pattern . Unchanged or marginal effects on electrical activity were also seen with a specific SK4 channel antagonist , TRAM-34 ( 1 µM , Fig . 1C ) . However , in some cells TRAM-34 application increased the action potential dramatically ( Fig . 1D ) in agreement with observations with the SK-channel antagonist apamin [8] . Note that before SK-channel block , the cell in Fig . 1D was almost quiescent , and fired action potentials very infrequently and randomly . This increase in spike frequency can be simulated by a stochastic version of the model . By including noise in the K ( ATP ) current , an otherwise silent cell produces infrequent action potentials evoked by random perturbations ( Fig . 1E ) . When the SK-conductance is set to 0 nS/pF , the cell starts rapid action potential firing driven by the underlying deterministic dynamics . The model analysis indicates that this mechanism only works if the cell is very near the threshold for electrical activity in the absence of the SK-channel antagonist . Düfer et al . [15] suggested a similar , important role for SK4 channels in promoting electrical activity in murine -cells at subthreshold glucose concentrations . Summarizing , cell-to-cell heterogeneity can explain the differences seen in the electrophysiological responses to SK-channel antagonists . In addition to spiking electrical activity , human -cells often show rapid bursting , where clusters of a few action potentials ( active phases ) are separated by hyperpolarized silent phases [1] , [4] , [9] , [10] , [16] ( Fig . 2A ) . The extended model presented here can also reproduce this behavior ( Fig . 2B ) as could the previous version of the model [7] , where the alternations between silent and active phases were controlled by HERG-channels . In contrast , in the present version of the model the rapid burst pattern ( Fig . 2B , upper trace ) can be controlled by SK-channels , which in turn are regulated by and ultimately by bulk cytosolic levels ( ) . The simulated cytosolic concentration shows the characteristic sawtooth pattern ( Fig . 2B , lower trace ) of a slow variable underlying bursting [17] , [18] . Thus , as in the pioneering model by Chay and Keizer [19] , increases during the active phase and activates SK-channels , which eventually repolarize the cell . During the silent phase decreases and SK-channels close , allowing another cycle to occur . Blocking voltage-dependent -channels in human -cells showing spiking electrical activity with tetrodotoxin ( TTX ) typically reduces the action potential amplitude by ∼10 mV , and broadens its duration [3] , [9] , [10] ( Fig . 3A ) . The previous version of the model [7] could reproduce these results , though the reduction in peak voltage was slightly less than observed experimentally . The inclusion of SK-channels in the model leads to a greater reduction in the spike amplitude ( Fig . 3B , upper trace ) when -channels are blocked . This improvement is because of a mechanism where the slower upstroke in the presence of -channel blockers allows submembrane to build up earlier and to higher concentrations ( Fig . 3B , lower trace ) , and consequently to activate more SK-channels , which in turn leads to an earlier repolarization reducing the action potential amplitude . In other experiments ( Fig . 3C ) [16] , TTX application suppresses action potential firing . In agreement , simulated spiking electrical activity can be suppressed by TTX application if the cell is less excitable because of , for example , smaller -currents ( Fig . 3D , upper , black trace ) . Before TTX application , the simulated cell had less hyperpolarized inter-spike membrane potential ( mV; Fig . 3D ) compared to the simulation with default parameters ( mV , Fig . 3B ) . This finding is in accordance with experimental recordings ( compare Fig . 3A and 3C ) . The cessation of action potential firing leads to a reduction in simulated ( Fig . 3D , lower , black trace ) . The model predicts that spiking , electrical activity can continue in presence of TTX even in less excitable cells , e . g . , with lower depolarizing -currents , if the hyperpolarizing K ( ATP ) -current is sufficiently small ( Fig . 3D , upper , gray trace ) . In this case , is nearly unchanged ( Fig . 3D , lower , gray trace ) . Hence , it is the relative sizes of the depolarizing and hyperpolarizing currents that determine whether TTX application silences the cell or allows the cell to remain in a region where action potential firing continues . The model thus predicts that in some cells , which stop firing action potentials in the presence of TTX , increased glucose concentrations or sulfonylureas ( K ( ATP ) -channel antagonists ) could reintroduce spiking electrical activity . More surprisingly , TTX application can change spiking electrical activity to rapid bursting in some cells ( Fig . 3E ) . This behavior can also be captured by the model ( Fig . 3F ) . To simulate this behavior it was necessary to increase the size of the -current . Without TTX , the big -current leads to large action potentials , which activate sufficient BK-current to send the membrane potential back to the hyperpolarized state , allowing a new action potential to form . With -channels blocked , there is insufficient depolarizing current to allow full action potentials to develop . In consequence , less BK-current is activated ( Fig . 3F , lower trace ) , and the membrane potential enters a regime with more complex dynamics where smaller spikes appear in clusters from a plateau of ∼−40 mV . The change to bursting activity leads to a notable increase in simulated ( Fig . 3F , middle trace ) . High-voltage activated L- and P/Q-type -currents are believed to be directly involved in exocytosis of secretory granules in human -cells [1] , [3] , [4] , [20] , [21] . Blocking L-type -channels suppresses electrical activity [3] , which is reproduced by the model ( Fig . 4A ) [7] , and the lack of electrical activity is likely the main reason for the complete absence of glucose stimulated insulin secretion in the presence of L-type -channel blockers [3] . Thus , L-type -channels participate in the upstroke of action potentials and increases excitability of human -cells . In contrast , and surprisingly , application of the P/Q-type -channel antagonist -agatoxin IVA does not block or slow down electrical activity , but leads to an increased spike frequency ( Fig . 4B ) . Electrical activity continues also in our model simulations of P/Q-type channel block with slightly increased spike frequency ( Fig . 4C ) . Reduced entry leads to lower peak concentrations in the submembrane space ( ; Fig . 4D ) . As a consequence , less hyperpolarizing SK-current is activated ( Fig . 4E ) , which leads to an increase in spike frequency ( Fig . 4C ) . Hence , the reduction in excitability caused by blockage of the P/Q-type -current can be overruled by the competing increase in excitability due to the smaller SK-current . Experimentally , -agatoxin IVA application reduced the action potential amplitude slightly in 3 of 4 cells ( by 2 . 0–4 . 3 mV ) , a finding that was quantitatively reproduced by the model , although the reduction was larger ( ∼7 . 5 mV in Fig . 4C ) . A direct conclusion from Fig . 4B is that the P/Q-type -current is not needed for the action potential upstroke , unlike the L-type current , probably because of the fact that P/Q-type channels activate at higher membrane potentials than L-type channels . The fact that electrical activity persists with P/Q-type -channels blocked , albeit with lower peak , could underlie the finding that -agatoxin IVA only partly inhibits insulin secretion [3] . The neurotransmitter -aminobutyric acid ( GABA ) is secreted from pancreatic -cells , and has been shown to stimulate electrical activity in human -cells [12] . In human -cells , GABA activates GABAA receptors , which are ligand-gated channels , thus creating an additional current . Notably , the reversal potential in human -cells is less negative than in many neurons , and positive compared to the -cell resting potential , which means that currents , such as the GABAA receptor current , stimulate action potential firing in -cells . Hence , GABA is a excitatory transmitter in -cells , in contrast to its usual inhibitory role in neurons . We simulate the addition of GABA by raising the GABAA receptor conductance . In a silent model cell with a rather large K ( ATP ) -conductance , simulated GABA application leads to a single action potential whereafter the membrane potential settles at ∼−45 mV ( Fig . 5A ) , in close correspondence with the experimental results [12] . In an active cell , the simulation of activation of GABAA receptors leads to a minor depolarization and increased action potential firing ( Fig . 5B ) , as found experimentally [12] . Another neurotransmitter , acetylcholine , might also play a paracrine role in human pancreatic islets , where it is released from -cells , and activates muscarinic receptors in -cells [22] . Muscarinic receptor activation by acetylcholine triggers a voltage-insensitive -current in mouse pancreatic beta-cells [23] , and similarly , the muscarinic agonist carbachol activates nonselective leak channels ( NALCN ) in the MIN6 -cell line [24] . Based on these findings , it was speculated that muscarinic activation of NACLN currents in human -cells might participate in the positive effect of acetylcholine and carbachol on insulin secretion [4] . Experimentally , we found that carbachol ( 20 µM ) accelerates action potential firing ( Fig . 5C ) . We tested the hypothesis of a central role of leak current activation by increasing the leak conductance in the model to simulate carbachol application , which caused accelerated action potential firing . The simulation thus reproduced the experimental data , and lends support to the hypothesis that carbachol and acetylcholine can accelerate action potential firing via muscarinic receptor-dependent stimulation of NALCN currents [4] . We finally use our model to address the origin of slow rhythmic patterns of electrical activity in human -cells ( Fig . 6A ) [4] , [25] , which likely underlie slow oscillations in intracellular [26] , [27] and pulsatile insulin release [28] , [29] . Based on accumulating evidence obtained in rodent islets [5] , [30] , we have previously speculated that oscillations in metabolism could drive these patterns [7] . In support of this hypothesis , oscillations in ATP levels with a period of 3–5 minutes have been observed in human -cells [13] , [14] . By adding a glycolytic component [11] , which can oscillate due to positive feedback on the central enzyme phosphofructokinase ( PFK ) , our model can indeed simulate such periodic modulation of the electrical pattern , where action potential firing is interrupted by long silent , hyperpolarized periods , which drives slow oscillations ( Fig . 6 ) .
Human -cells show complex and heterogeneous electrophysiological responses to ion channel antagonists . It can therefore sometimes be difficult to reach clear conclusions regarding the participation of certain ion channels in the various phases of electrical activity , in particular since some of the electrophysiological responses are nonintuitive as shown here . A deeper understanding of the role of ion channels in electrical activity and insulin secretion could have important clinical benefits , since it might help in the development of new anti-diabetic drugs . We have here shown how mathematical modeling can help in interpreting various electrophysiological responses , and in particular , to study the effect of competing effects and cell heterogeneity . The role of SK-channels in human -cells is still not clear . We ( Fig . 1 ) and others [8] have found heterogeneous electrophysiological responses to SK-channel antagonists . Our model suggests that these differences can be caused by underlying variations in cell excitability: Less excitable -cells that produce action potentials evoked mostly by stochastic channel dynamics show a clear increase in action potential frequency when SK-channels are blocked ( Fig . 1DE ) . In contrast , spiking electrical activity in very active cells is driven by the deterministic dynamics caused by ion channel interactions , and is nearly unchanged by SK-channel blockers ( Fig . 1A–C ) . We showed also that rapid bursting activity can be driven by and SK-channels ( Fig . 2 ) , which could add a complementary mechanism to HERG-channel dynamics [7] for the control of rapid bursting . The wide range of responses to TTX could be accounted for by a single model but with different parameters , i . e . , differences in the relative size of the various currents . A peculiar finding is the qualitative change from spiking to rapid bursting seen in some cells ( Fig . 3E ) . We suggest that this happens in human -cells with large -currents . The blockage of this depolarizing current reduces the amplitude of the action potentials , and as a consequence , the size of the hyperpolarizing BK-current . Under the right conditions , the combination of these competing events allows the membrane potential to enter a bursting regime controlled by SK- and/or HERG-channels ( Fig . 3F ) . Interestingly , it has been found that TTX reduces insulin secretion evoked by 6 mM glucose greatly , but at glucose levels of 10–20 mM , the effect of TTX on secretion is smaller [3] , [9] , [10] . Based on our simulations showing that less excitable cells cease to fire in the presence of TTX ( Fig . 3D , black traces ) , but that lower can reintroduce spiking activity ( Fig . 3D , gray traces ) , we suggest that at low , near-threshold glucose levels TTX abolishes electrical activity in many cells , which reduces the and consequently insulin secretion greatly ( Fig . 3D , black traces ) . At higher glucose concentrations , -cells have lower K ( ATP ) -conductance and in some of the cells that stop firing in low glucose concentration the effect of TTX on electrical activity and is smaller ( Fig . 3D , gray traces ) . Hence , more -cells remain active in the presence of TTX at high than at low glucose levels . Consequently , insulin secretion is more robust to TTX at higher glucose concentrations . Similarly , insulin release is more affected by the P/Q-type -channel blocker -agatoxin IVA at 6 mM ( −71% ) than at 20 mM ( −31% ) glucose [3] . This is in contrast to L-type -channel antagonists , which abolish insulin secretion at both high ( 15–20 mM ) and low ( 6 mM ) glucose concentrations [1] , [3] , [10] . These results concerning L-type -channel block are easily explained by the fact that L-type channel activity is necessary for action potential generation [3] ( Fig . 4A ) . In contrast , we showed that electrical activity in human -cells not only persists , but is accelerated by -agatoxin IVA ( Fig . 4B ) . The counter-intuitive finding of increased excitability and electrical activity when the depolarizing P/Q-type -current is blocked by -agatoxin IVA can be accounted for by an even greater reduction in the hyperpolarizing SK-current due to reduced -influx and consequently lower . Our mathematical modeling confirmed that GABA released from human -cells can have a role as a positive feedback messenger . GABA application has been shown to depolarize both silent and active human -cells [12] , which was reproduced here . A detailed characterization of GABAA receptor currents would refine the analysis presented here . Data from mouse -cells [23] and the MIN-6 -cell line [24] suggest that muscaric agonists such as carbachol and acetylcholine stimulate insulin secretion partly by activating NACLN currents . Using our model we could translate this finding to the human scenario , thus testing the hypothesis that this mechanism is also operating in human -cells [4] . Our simulations confirmed that increased leak currents can underlie the change in electrical activity found experimentally ( Fig . 5C ) . The incretin hormone glucagon-like peptide 1 ( GLP-1 ) has also been shown to act partly via activation of leak channels [31] , a mechanism which might be involved in activating otherwise silent -cells [7] , [32] , [33] . These results suggest that leak currents could play important roles in controlling electrical activity in -cells , and potentially be pharmacological targets . Further studies are clearly needed to investigate these questions . We were also able to simulate slow rhythmic electrical activity patterns by adding an oscillatory glycolytic component to the model . To date , there is to our knowledge no evidence of oscillations in glycolytic variables in human ( or rodent ) -cells or islets , but ATP levels have been found to fluctuate rhythmically also in human -cells [13] , [14] , supporting the idea of metabolism having a pacemaker role . In agreement , data from rodent -cells show accumulating evidence for oscillations in metabolism playing an important role in controlling pulsatile insulin secretion [5] , [30] . It will be interesting to see if these findings in rodents are applicable to human -cells . Regarding the model development , the inclusion of SK-channels in the model provided insight that was not within reach with the previous version of the model [7] . Besides the direct investigation of the role of SK-channels , the acceleration in action potential firing seen with P/Q-type channel blockers ( Fig . 4 ) can not be reproduced by the older version of the model without SK-channels [7] . Moreover , considering the effect of TTX on spike amplitude , a better correspondence between experiments and simulations was found with SK-channels included in the model . To model SK-channel activation accurately , we made a special effort to describe carefully . Submembrane responds rapidly to an action potential , while integrates many action potentials . The rapid submembrane dynamics has important consequences for the study of the role of SK-channels in spiking electrical activity , e . g . , it was crucial for explaining the larger effect of TTX on spike amplitude in this version of the model . Most models of electrical activity in rodent -cells do not include a submembrane compartment , but these models were typically built to explain the slow bursting patterns seen in rodent islets with a period of tens of seconds . For these long time scales , the rapid dynamics in the submembrane compartment is not important . In contrast , the situation is different in human -cells with their faster dynamics .
We build on the previously published Hodgkin-Huxley type model for human -cells [7] , which was mainly based on the results of Braun et al . [3] , who carefully assured that investigated human islet cells were -cells . We include SK-channels in the model . Since these channels are -sensitive and located at some distance from -channels [34] we also model -dynamics in a submembrane layer controlling SK-channel activity . The membrane potential ( measured in mV ) develops in time ( measured in ms ) according to ( 1 ) All currents ( measured in pA/pF ) , except the SK-current and the GABAA receptor mediated current , are modeled as in [7] . Expressions and parameters are given below . For the stochastic simulation in Fig . 1E , we included “conductance noise” [35] in the K ( ATP ) current by multiplying by a stochastic factor , where is a standard Gaussian white-noise process with zero mean and mean square , see also [36]–[38] . SK-channels are assumed to activate instantaneously in response to elevations at the plasma membrane but away from channels [34] , and are modeled as [39] ( 2 ) In human -cells , flash-released triggered a ∼10 pA current at a holding current of mV , presumably through SK-channels [8] . Assuming that SK-channels were nearly saturated by , the maximal SK-conductance is estimated to be nS/pF . Here , = 10 pF is the capacitance of the plasma membrane [3] . In Eq . 2 , is the submembrane concentration ( ; measured in µM ) , which is described by a single compartment model [21] ( 3 ) where is the ratio of free-to-total , µmol/pA/ms changes current to flux , and and are the volumes of the submembrane compartment and the bulk cytosol , respectively . describes the flux of from the submembrane compartment to the bulk cytosol , is the flux through plasma membrane -ATPases , and represents flux through the - exchanger . Cytosolic ( ; measured in µM ) follows ( 4 ) where describes SERCA pump-dependent sequestration of into the endoplasmic reticulum ( ER ) , and is a leak flux from the ER to the cytosol . Expressions and parameters for the fluxes are taken from [40] . The submembrane compartment volume is estimated based on the considerations of Klingauf and Neher [41] , who found that a shell model ( in contrast to a domain model ) describes submembrane satisfactorily when the shell-depth is chosen correctly . The dynamics between channels can be estimated from a shell model at a depth of ∼23% of the distance to a -channel . In mouse -cells the interchannel distance has been estimated to be nm [42] . Moreover , SK-channels are located nm from channels [34] . Based on these considerations , we modeled the submembrane space controlling SK-channels as a shell of depth nm . The radius of a human -cell is µm , which gives cell volume ( ) , shell volume ( ) and internal surface area ( ) of the shell , of ( 5 ) The flux-constant can then be calculated as [43] ( 6 ) where is a typical length scale . We set to 1 µm , which together with the diffusion constant for , [41] , [44] , gives . In human -cells , GABA activates GABAA receptors , which are ligand-gated channels . We model the current carried by GABAA receptor as a passive current with the expression ( 7 ) where is the GABAA receptor conductance , and mV is the chloride reversal potential [4] . We estimate from the findings that 1 mM GABA evokes a current of pA/pF ( but with substantial cell-to-cell variation ) at a holding potential of −70 mV [12] , which yields a conductance of nS/pF . To simulate the changes in firing patterns evoked by lower GABA concentrations ( 10 or 100 µM ) [12] , we take into consideration the does-response curve [45] for the subunits , which are the most highly expressed subunits in human -cells [12] . At 10 µM the GABA-evoked current is -fold smaller compared to 1 mM GABA , and we set nS/pF . At 100 µM , the reduction is about 2-fold compared to 1 mM . We used nS/pF to simulate application of 100 µM GABA . To investigate slow electrical patterns ( Fig . 6 ) we added a glycolytic component [11] , which drives ATP levels and K ( ATP ) channel activity . The glycolytic subsystem can oscillate due to positive feedback on the enzyme phosphofructokinase ( PFK ) from its product fructose-1 , 6-bisphosphate ( FBP ) . The glycolytic equations are ( 8 ) ( 9 ) ( 10 ) where is the rate of glucokinase , which phosphorylates glucose to glucose-6-phosphate ( G6P ) . G6P is assumed to be in equilibrium with fructose-6-phosphate ( F6P ) , the substrate for PFK , and is the sum of G6P and F6P . is the rate of PFK producing FBP , which is subsequently removed by fructose-bisphosphate aldolase ( FBA ) , which produces glyceraldehyde-3-phosphate ( G3P ) and dihydroxyacetone-phosphate ( DHAP ) with rate . DHAP and G3P are assumed to be in equilibrium , and indicates their sum . Finally , G3P serves as substrate for glyceraldehyde-3-phosphate dehydrogenase ( GAPDH with rate ) , which via the lower part of glycolysis eventually stimulates mitochondrial ATP production . We introduce a phenomenological variable that mimics ATP levels , and is model by ( 11 ) The K ( ATP ) conductance depends inversely on , and is modeled as ( 12 ) Expressions and parameters are given below . Simulations were done in XPPAUT [46] with the cvode solver , except the stochastic simulation in Fig . 1E , which was performed with the implicit backward Euler method . Computer code can be found as supplementary material , or downloaded from http://www . dei . unipd . it/ pedersen . Human pancreatic islets were obtained with ethical approval and clinical consent from non-diabetic organ donors . All studies were approved by the Human Research Ethics Board at the University of Alberta . The islets were dispersed into single cells by incubation in free buffer and plated onto 35 mm plastic Petri dishes . The cells were incubated in RPMI 1640 culture medium containing 7 . 5 mM glucose for >24 h prior to the experiments . Patch-pipettes were pulled from borosilicate glass to a tip resistance of 6–9 MΩ when filled with intracellular solution . The membrane potential was measured in the perforated-patch whole-cell configuration , using an EPC-10 amplifier and Patchmaster software ( HEKA , Lambrecht , Germany ) . The cells were constantly perifused with heated bath solution during the experiment to maintain a temperature of C . The extracellular solution consisted of ( in mM ) 140 NaCl , 3 . 6 KCl , 0 . 5 , 1 . 5 , 10 HEPES , 0 . 5 , 5 and 6 glucose ( pH was adjusted to 7 . 4 with NaOH ) . The pipette solution contained ( in mM ) 76 , 10 KCl , 10 NaCl , 1 , 5 HEPES ( pH 7 . 35 with KOH ) and 0 . 24 mg/ml amphotericin B . -cells were identified by immunostaining ( 18 out of 28 cells ) or by size when immunostaining was not possible ( cell capacitance >6 pF , [3] ) . Tetrodotoxin ( TTX ) and -agatoxin IVA were purchased from Alomone Labs ( Jerusalem , Israel ) , UCL-1684 was obtained from R&D Systems ( Minneapolis , MN ) , TRAM-34 from Sigma-Aldrich ( Oakville , ON , Canada ) . Figures with experimental responses to ion channel antagonists ( Figs . 1 , 3 , 4 and 5 ) show recordings from the same cell before ( ctrl ) and after application of the blocker . For completeness , we report all expressions and parameters of the mathematical model here . For details , please refer to the Modeling section above and the previous article [7] . The main variables , membrane potential , , submembrane , , and cytosolic , , are described by ( 13 ) ( 14 ) ( 15 ) The currents are ( 16 ) ( 17 ) ( 18 ) ( 19 ) ( 20 ) ( 21 ) ( 22 ) ( 23 ) ( 24 ) ( 25 ) ( 26 ) where ( 27 ) and activation variables ( and similarly inactivations variabels , , where denotes the type of current ) follow ( 28 ) where ( respectively ) is the time-constant of activation ( respectively inactivation for ) , and ( respectively ) is the steady-state voltage-dependent activation ( respectively inactivation ) of the current . The steady-state activation ( and inactivation ) functions are described with Boltzmann functions , ( 29 ) except ( 30 ) for -dependent inactivation of L-type channels . The time-constant for activation of Kv-channes is assumed to be voltage-dependent [3] , [7] , ( 31 ) -fluxes are [40] ( 32 ) ( 33 ) ( 34 ) Glycolysis was modeled by [11] ( 35 ) ( 36 ) ( 37 ) which controls the electrophysiological subsystem via the “ATP-mimetic” and K ( ATP ) -channels , as described by ( 38 ) ( 39 ) Here ( 40 ) ( 41 ) ( 42 ) ( 43 ) where ( 44 ) ( 45 ) ( 46 ) and ( 47 ) Default parameters are given in Table 1 . | Insulin is a glucose-lowering hormone secreted from the pancreatic -cells in response to raised plasma glucose levels , and it is now well-established that defective insulin secretion plays a pivotal role in the development of diabetes . The -cells are electrically active , and use electrical activity to transduce an increase in glucose metabolism to calcium influx , which triggers insulin release . Experimental and theoretical studies on -cells from rodents have provided valuable insight in their electrophysiology . However , human -cells differ from their rodent counterparts in several aspects including their electrophysiological characteristics . We show that the electrophysiological responses in human -cells to a range of experimental manipulations are heterogeneous . We extend a previous mathematical model of electrical activity in human -cells to investigate such heterogeneous and nonintuitive electrophysiological responses , and use our findings to obtain insight in previously published insulin secretion measurements . By adding a glycolytic component to the electrophysiological model , we show that oscillations in glucose metabolism might underlie slow oscillations in electrical activity , calcium levels and insulin secretion observed experimentally . We conclude that the interdisciplinary combination of experiments and modeling increases our understanding of human -cell physiology and provides new insight in -cell heterogeneity . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"medicine",
"insulin",
"endocrine",
"system",
"biology",
"anatomy",
"and",
"physiology",
"electrophysiology",
"computational",
"biology",
"endocrine",
"physiology"
] | 2014 | Mathematical Modeling of Heterogeneous Electrophysiological Responses in Human β-Cells |
Herpes simplex virus type 2 ( HSV-2 ) glycoprotein D ( gD2 ) subunit antigen is included in many preclinical candidate vaccines . The rationale for including gD2 is to produce antibodies that block crucial gD2 epitopes involved in virus entry and cell-to-cell spread . HSV-2 gD2 was the only antigen in the Herpevac Trial for Women that protected against HSV-1 genital infection but not HSV-2 . In that trial , a correlation was detected between gD2 ELISA titers and protection against HSV-1 , supporting the importance of antibodies . A possible explanation for the lack of protection against HSV-2 was that HSV-2 neutralization titers were low , four-fold lower than to HSV-1 . Here , we evaluated neutralization titers and epitope-specific antibody responses to crucial gD2 epitopes involved in virus entry and cell-to-cell spread as correlates of immune protection against genital lesions in immunized guinea pigs . We detected a strong correlation between neutralizing antibodies and protection against genital disease . We used a high throughput biosensor competition assay to measure epitope-specific responses to seven crucial gD2 linear and conformational epitopes involved in virus entry and spread . Some animals produced antibodies to most crucial epitopes while others produced antibodies to few . The number of epitopes recognized by guinea pig immune serum correlated with protection against genital lesions . We confirmed the importance of antibodies to each crucial epitope using monoclonal antibody passive transfer that improved survival and reduced genital disease in mice after HSV-2 genital challenge . We re-evaluated our prior study of epitope-specific antibody responses in women in the Herpevac Trial . Humans produced antibodies that blocked significantly fewer crucial gD2 epitopes than guinea pigs , and antibody responses in humans to some linear epitopes were virtually absent . Neutralizing antibody titers and epitope-specific antibody responses are important immune parameters to evaluate in future Phase I/II prophylactic human vaccine trials that contain gD2 antigen .
Nearly one half-billion people worldwide are infected with herpes simplex virus type 2 ( HSV-2 ) and another one-quarter billion have genital infection caused by herpes simplex virus type 1 ( HSV-1 ) [1 , 2] . The disease manifestations vary from asymptomatic genital infection to severe , recurrent ulcerative genital disease [3–5] . Genital herpes poses major risks to newborns delivered through an infected birth canal with an estimated 14 , 000 annual cases globally and a mortality of 60% if untreated [2 , 6] . HSV-2 genital infection carries a three-fold increased risk for acquisition and transmission of HIV [7] . A vaccine to prevent genital herpes is urgently needed , yet none has been approved despite several clinical trials evaluating the efficacy of potential vaccine candidates [8–10] . The Herpevac Trial for Women was the most recent large prophylactic genital herpes human vaccine trial that involved immunizing HSV-1 and HSV-2 doubly seronegative women with gD2 subunit antigen [10] . The vaccine protected against HSV-1 genital infection , but not HSV-2 . Higher ELISA antibody titers were associated with reduced acquisition of HSV-1 , highlighting the importance of antibodies for protection against genital herpes [10] . Vaccine-induced antibodies correlate with protection against other sexually transmitted infections , including human papilloma virus ( HPV ) and hepatitis B virus ( HBV ) [11] . Therefore , our strategy has focused on the generation of robust antibody responses for a prophylactic HSV-2 vaccine . HSV-2 prophylactic vaccine antibody responses have generally targeted virus entry molecules [8–10] . Glycoprotein D is one of four glycoproteins essential for virus entry into cells . The other glycoproteins are gB , gH , and gL [12 , 13] . Glycoprotein B , along with gC facilitates the initial attachment of virus to host cells by interaction with heparan sulfate proteoglycans . Glycoprotein D then binds to cell receptors nectin-1 , herpesvirus entry mediator ( HVEM ) , or 3-O-sulfotransferase heparan sulfate an event that induces conformational changes in gD , allowing it to activate the gH/gL heterodimer complex [14 , 15] . Activation of gH/gL transitions gB into a fusogenic state that facilitates fusion of the virus and host cell membranes and subsequent entry of the virus capsid into the cell [16–19] . Glycoprotein D is also required for efficient spread of virus from one cell to another [20 , 21] . Domains on gD involved in cell-to-cell spread are thought to differ from those involved in virus entry [22 , 23] . Identifying the location on gD of these important functional domains was aided by solving the atomic structure of gD [24] . We have assembled a large panel of anti-gD MAbs that we used to assess antibody responses to specific gD2 linear and conformational epitopes elicited by the gD2 vaccine in a subset of 29 human volunteers in the Herpevac trial [25] . Individuals that produced antibodies to the greatest number of neutralizing gD2 epitopes , as measured by antibody competition assays , had the highest neutralizing antibody titers , suggesting that a diverse antibody response correlates with neutralizing activity [25] . The Herpevac trial did not protect against HSV-2 genital infection; therefore , we were not able to assess whether antibody responses to specific gD2 epitopes correlate with protection against genital herpes . We recently evaluated a trivalent vaccine consisting of HSV-2 glycoproteins C , D and E ( gC2 , gD2 , and gE2 ) aimed at improving antibody responses by blocking virus entry mediated by gD2 and immune evasion from antibody and complement mediated by gE2 and gC2 , respectively [26] . The trivalent vaccine was 97% efficacious in reducing the number of days with genital lesions compared with mock-immunized animals [26] . Some guinea pigs received gD2 alone as comparison for vaccine efficacy with the trivalent vaccine . The gD2 alone group was 80% efficacious in reducing the number of days with genital lesions compared with mock-immunized animals , and 33% efficacious in totally preventing genital lesions . The observation that some animals in the gD2 alone group were fully protected while others had lesions on one or more days provided an opportunity to correlate responses to specific gD2 epitopes with the level of protection against disease . Passive antibody transfer studies have been used to evaluate protection provided by antibodies against HSV-2 intravaginal infection [27 , 28] . We used the murine genital infection model to demonstrate that passive transfer of individual gD type common and type specific MAbs that bind to crucial gD2 epitopes required for virus entry or cell-to-cell spread in vitro provide protection against intravaginal challenge in vivo . We used the guinea pig model of genital infection to correlate antibodies produced to specific epitopes with protection against genital lesions after intravaginal challenge in gD2-immunized guinea pigs . We noted that neutralizing and gD2 ELISA antibody titers correlated with protection as did producing antibodies to multiple crucial gD2 epitopes . We used this information to re-evaluate the gD2 immune responses of humans immunized with the HSV-2 gD2 vaccine in the Herpevac Trial . Our results demonstrate that antibody responses to some crucial gD2 epitopes that are protective in animal models were absent in the human vaccine trials .
Twenty-five animals immunized with gD2 subunit antigen with CpG and alum as adjuvants were evaluated in two prophylactic vaccine studies [26] . All 25 animals survived intravaginal challenge; however , 16/25 animals developed genital lesions [26] . We performed a secondary analysis of correlates of protection against genital lesions based on the previously collected neutralizing antibody and genital lesions data . In addition , we generated new results that assessed endpoint ELISA antibody titers , monoclonal antibody ( MAb ) passive transfer , and high throughput biosensor assays . We restricted our analysis to protection against genital lesions since too few animals , only 4 of 25 ( 16% ) , were totally protected against any evidence of infection , defined as no genital lesions and no subclinical shedding of HSV-2 DNA in genital secretions [26] . The previously reported serum neutralizing antibody titers without complement are aligned from lowest to highest neutralizing antibody titers and ranged from 1:160 to 1:5120 in these outbred animals ( geometric mean titer 1:893 ) ( Fig 1A ) [26] . We listed the animals in the same order and plotted cumulative days with genital lesions over a period of 60 days post-challenge . Nine animals had no genital lesions , while 16 had genital lesions on one to 14 days ( Fig 1B ) . We performed regression analysis and found that higher neutralizing titers correlated significantly with fewer lesions days ( Fig 1C ) . We conclude that neutralizing antibodies are important for protection of these gD2-immunized animals; however , other immune responses associated with gD2 immunization also likely contribute to protection since the correlation was strong , but not perfect . For example , animals #19 and 25 with neutralizing titers of 1:2560 and 1:5120 , respectively developed genital lesions . We performed a similar analysis comparing gD2 ELISA endpoint titers with genital lesions , since a correlation between ELISA titers and protection against genital HSV-1 infection was noted in the Herpevac Trial for Women [10 , 29] . Each animal was assigned the same number as in Fig 1 , although the order of the animals changed based on lowest to highest ELISA titers that ranged from 1:3 , 000 to 1:25 , 000 , with a geometric mean titer of 1:8922 ( Fig 2A ) . We noted a significant correlation between higher ELISA titers and fewer cumulative lesion days ( Fig 2B and 2C ) , which is consistent with the HSV-1 results in the Herpevac Trial for Women [10 , 29] . Neutralizing antibodies reflect a mechanism of protection , such as blocking virus entry , while ELISA titers indicate antibody binding but do not define a specific mechanism by which the antibodies protect . We analyzed whether neutralizing antibody and ELISA titers correlate with one another ( Fig 2D ) . The relatively strong correlation between these two assays for measuring antibody responses indicates that neutralizing antibodies are an important mechanism for protection . Since the correlation was not perfect ( that is , it was <1 . 0 ) , it suggests that other immune mechanisms also contribute to overall protection . We used a high throughput biosensor-based MAb competition assay to assess the production of epitope-specific antibodies in immunized guinea pigs . The MAbs selected for evaluation were based on their assignment to competition-based communities and subcommunities ( Fig 3A ) [30] . We assessed whether antibodies are produced by immunization to specific gD2 epitopes using the approach shown schematically in Fig 3B . First , gD MAbs from the different communities were printed on the biosensor chip . Purified gD2 ( 285t ) was pre-incubated with IgG isolated from individual immunized guinea pigs and each gD2-IgG mix was flowed separately over the printed MAbs on the chip . The rationale for using gD2 ( 285t ) rather than gD2 ( 306t ) is explained in Methods . We evaluated whether pre-incubation of gD2 ( 285t ) with IgG from individual guinea pigs blocked gD2 ( 285t ) binding to the printed MAbs . Blocking gD2 ( 285t ) binding to the printed MAb indicates that the guinea pig IgG contains antibodies that compete for binding with the printed MAb ( Fig 3C–3I ) . Guinea pigs were assigned the same numbers as in Figs 1 and 2 . Twenty-three of 25 guinea pigs produced antibodies that competed with the prototype MAb 77S in the pink community . 77S was used as the prototype MAb rather than DL11 because DL11 binding activity tends to be lost when the chip is repeatedly regenerated during antibody blocking studies . Only guinea pigs #1 and #2 failed to produce antibodies that blocked gD2 binding to prototype MAb 77S ( Fig 3C ) . Twenty-four of 25 immunized animals produced antibodies to epitopes recognized by the blue prototype MAb MC5 , with only guinea pig #2 failing to block ( Fig 3G ) . Twenty-two of 25 guinea pigs produced antibodies that competed with the red prototype MAb MC23 , with animals #2 , #4 and #10 failing to block ( Fig 3H ) . Eighteen of 25 animals produced antibodies that blocked the brown prototype MAb DL6 ( Fig 3F ) , while many fewer animals ( 7/24 ) produced antibodies to epitopes recognized by the yellow prototype MAb 1D3 ( Fig 3I ) , green prototype MAb MC2 ( 8/25 ) ( Fig 3D ) or brown prototype MAb MC14 ( 3/24 ) ( Fig 3E ) . For most communities , we evaluated two or more MAbs and obtained highly concordant results regarding the percent blocking and whether guinea pig IgG blocked gD2 binding to the MAbs in the community . The one exception was in the brown community where DL6 ( Fig 3F ) had different performance characteristics than the other MAbs in that community ( Fig 3E ) , supporting prior mapping studies that placed the DL6 epitope in group IIb , while the other brown community MAbs were grouped as IIa [25 , 31] . We conclude that some crucial gD2 epitopes , such as those recognized by MAbs in the pink , blue and red communities are highly immunogenic in almost all immunized animals , while others , such as those recognized by the green , yellow and brown communities are considerably less immunogenic . We next evaluated whether immunization in guinea pigs that produced antibodies to multiple crucial gD2 epitopes offered added protection against genital lesions compared with immunization that produced antibodies to few crucial epitopes . An epitope was considered blocked if it showed any competition ( ≥1% ) with the community prototype MAb for binding to gD2 . We chose ≥1% as positive for blocking , because the biosensor value is generally a negative number when IgG fails to block gD2 binding . Animals were assigned the same numbers as in Figs 1–3 . The number of separate epitopes blocked ranged from 2 to the maximum of 7 with a mean of 4 . 2 ( Fig 6A ) . The genital disease scores in animals were arranged in the same order as Fig 6A now according to the number of epitopes blocked ( Fig 6B ) . We detected a highly significant correlation between the number of epitopes blocked and cumulative lesion days ( Fig 6D ) . A breakpoint appears at ≥4 epitopes . The 18 animals that produced antibodies that blocked ≥4 epitopes had genital lesions on a mean of 1 . 9 ± 3 . 4 days compared with 7 . 3 ± 5 . 2 days for the seven animals that blocked ≤3 epitopes ( p = 0 . 006 , by Student’s t-test ) . We performed a sensitivity analysis in which we used ≥5% blocking as the cutoff to consider an epitope blocked rather than ≥1% . The correlation between number of epitopes blocked and cumulative lesion days remained highly significant ( r = -0 . 52 , p = 0 . 008 ) . Therefore , we detected a significant correlation between the epitopes blocked and genital lesions when we used 1% or 5% as the cut-off for blocking . We evaluated the cumulative percent of epitopes blocked as another approach to assess the correlation between epitopes blocked and cumulative lesion days . We determined the cumulative percent of epitopes blocked by calculating the sum of blocking by each prototype MAb . For example , guinea pig #2 had a cumulative disease score of 13 and cumulative percent blocking of 37% that was based on blocking the MC2 epitope by 7% and the 1D3 epitope by 30% , but not blocking any other epitope ( Figs 3 and 6C ) . The correlation between cumulative lesion days and cumulative percent epitope blocking was r = -0 . 42 ( p = 0 . 04 ) ( Fig 6F ) , which was less robust than the correlation between cumulative lesion days and number of epitopes blocked ( r = -0 . 57 ) . The results indicate that the number of epitopes blocked is a stronger correlate of cumulative disease days than the cumulative percent of epitopes blocked . In human vaccine trials , the number of subjects that develop disease ( incidence of disease ) is generally used as an endpoint rather than the number of days with lesions ( a measure of disease severity ) [9 , 10] . Therefore , we extended our analysis to evaluate the guinea pig results based on the incidence of genital disease , defined as one or more lesion days . Animals with no genital lesions produced antibodies that blocked significantly more epitopes compared with animals that developed genital lesions ( Fig 6E ) , which is consistent with results using the cumulative number of lesion days as the endpoint . In addition , we evaluated the antibody responses to each crucial epitope to determine whether the lack of a response to a particular epitope was associated with significantly more lesion days . Two animals failed to produce antibodies that blocked the pink DL11/77S epitope . These two animals had significantly more lesion days than the 23 animals that produced antibodies to that epitope . The lack of response to no other crucial epitope was significantly associated with more lesion days , although animals that failed to produce antibodies that blocked the brown MC14 and red MC23 epitopes had more lesion days that approached statistical significance ( Table 2 ) . These results suggest that antibodies to the receptor binding epitopes , represented by DL11/77S and MC23 , and to a cell-to-cell spread epitope , represented by MC14 , are particularly important components of the multiple epitope response required for protection . Whitbeck et al reported the epitope-specific antibody responses of 29 human volunteers immunized with the GSK gD2 vaccine in the Herpevac Trial [10 , 25] . The samples selected for evaluation were obtained one-month after the third ( final ) immunization and were approximately equally represented by subjects that had high ( >8 , 061 ) , intermediate ( 2 , 463–8 , 060 ) or low ( <2 , 463 ) gD2 ELISA titers [25] . Most subjects produced antibodies to the epitopes recognized by the pink , blue , green and red communities; however , few or none produced antibodies to the yellow or brown communities represented by 1D3 , DL6 and MC14 ( Table 3 ) . The gD2 vaccine provided no protection against genital HSV-2 infection in the Herpevac Trial; therefore , we were not able to assess epitope specific correlates of protection . An interesting analysis that we were able to perform was to compare the epitope-specific antibody responses in gD2 immunized humans and guinea pigs ( Table 3 ) . The gD2 vaccine used in the guinea pig study differed from the Herpevac Trial in that gD2 containing 306 amino acids was used in guinea pigs , while 281 amino acids was used in humans . In addition , CpG and alum were used as adjuvants in guinea pigs compared with MPL and alum in humans . Importantly , both vaccines included all the epitopes evaluated with our gD2 MAb panel [10 , 26] . Table 3 includes the epitope specific antibody responses in humans ( results from [25] ) , and guinea pigs ( results from Fig 3 ) , and the protection provided by MAb passive transfer in mice ( results from Fig 5 ) . Yellow ( 1D3 ) and brown ( DL6 ) community antibodies were detected in 0% and 3% , respectively in humans immunized with gD2 . In contrast , 29% of guinea pigs produced antibodies that blocked gD2 binding to 1D3 and 72% produced antibodies that blocked binding to DL6 ( Table 3 ) . MAbs 1D3 and DL6 protected 80% and 70% respectively of mice against death . Few brown community antibodies ( MC14 ) were detected in either guinea pigs ( 13% ) or humans ( 3% ) , yet MC14 protected 50% of mice ( Fig 5 ) and each animal that produced antibodies that blocked this epitope was totally protected against genital lesions in guinea pigs ( Table 2 ) . Our guinea pig results suggest that blocking 4 or more epitopes protects against genital lesions . The mean number of epitopes blocked by gD2 immunization in guinea pigs was 4 . 2 ± 1 . 3 compared with 2 . 9 ± 1 . 4 in humans ( p = 0 . 001 , by Student’s t-test ) ( human data derived by re-analysis of [25] ) . Nineteen of 25 guinea pigs immunized with gD2 produced antibodies that blocked one or more of the DL6 , 1D3 and MC14 epitopes , compared with 2/29 humans ( p = 0 . 0001 , Fisher’s Exact Test ) . These results suggest that some important gaps may exist in the antibody responses produced by the gD2 vaccine in humans , particularly to the DL6 , 1D3 and MC14 linear epitopes , while antibody responses to the conformational epitopes were robust in humans and guinea pigs .
HSV-2 gD2 mediates virus receptor binding , activation of gH/gL and cell-to-cell spread , which are crucial viral functions [12 , 37] . We hypothesized that an optimal gD2 vaccine is one that produces antibodies to multiple crucial epitopes involved in these functions . The high throughput biosensor technology provides a powerful new tool for measuring epitope-specific antibody responses to linear and conformational epitopes . We used the biosensor technology to demonstrate that some epitopes were immunogenic in almost all guinea pigs , while others were immunogenic in only a few . We determined in guinea pigs that the greater the number of crucial gD2 epitopes blocked by immunization , the better the protection was against genital lesions . We used MAb passive antibody studies in mice to demonstrate that MAbs to each of seven crucial gD2 epitopes provided significant , although partial protection against death and genital disease . Additional passive antibody studies in mice will be required to evaluate whether combinations of MAbs will provide maximum protection and to assess the extent of protection provided by the MAb IgG Fab domain compared with the Fc domain . We demonstrated a strong correlation in guinea pigs between both neutralizing and ELISA titers with protection against genital lesions . In the Herpevac Trial , ELISA titers correlated with protection again HSV-1 genital infection [10 , 29] . The geometric mean gD2 ELISA titer in the Herpevac Trial was 1:6809 , which is not very different from the geometric mean ELISA titer of 1:8922 noted in our guinea pig studies [10] . In contrast , after the final immunization , the mean HSV-2 neutralizing titer in the Herpevac Trial for Women was 1:29 , which is radically different from the geometric mean neutralizing titer of 1:893 in our guinea pig studies [10 , 38] . In addition , in the Herpevac Trial , the mean neutralizing antibody titers fell to undetectable levels by study month 16 [10] . The strong correlation in guinea pigs between neutralizing antibodies and protection against genital lesions raises important concerns about the low neutralizing antibody response in the Herpevac Trial . Some researchers have proposed an important role for ADCC in protecting vaccinated mice against HSV , although the target antigens for the ADCC antibodies have not be defined [33 , 34 , 39] . The single cycle live virus vaccine candidate in those studies lacks gD2 , which is a major target of neutralizing antibodies; therefore , it is not surprising that alternate mechanisms of protection assume greater importance for that vaccine in murine models [40] . Our guinea pig results provide strong evidence that neutralizing antibodies contribute significantly to protection in gD2-immunized animals . Two other studies have also reported a correlation between neutralizing antibodies and protection against genital disease in gD2-immunized guinea pigs; however , a third report failed to detect this correlation [41–43] . In humans , studies of neonatal herpes have demonstrated that neutralizing antibodies in the newborn protect against the disseminated form of the disease [44 , 45] . A human neonatal herpes study evaluated both neutralizing antibodies and ADCC titers in newborns and noted that both independently correlate with protection against neonatal HSV infection [46] . Similar findings were reported in mice [47] . In addition , passive transfer studies in adult mice demonstrated that protection against intravaginal infection was mediated by neutralization and ADCC [48] . Therefore , both neutralizing antibodies and ADCC contribute to protection , with one or the other assuming greater importance depending on the immunogen used and both contributing to protection during natural infection . Neutralization in the absence of complement is mediated by the IgG Fab domain , while ADCC requires the IgG Fc domain . Our guinea pig results demonstrated a significant correlation between neutralizing titers and protection , indicating a role for the IgG Fab domain . Evidence to support a possible contribution of the IgG Fc domain to protection comes from some passive transfer experiments . MAb MC23 from the red community is an IgG1 MAb that blocks the nectin-1 receptor . It protected only 30% of mice against death despite being a potent neutralizing antibody with an endpoint titer of 0 . 26 μg/ml . DL11 from the pink community is an IgG2a MAb that blocks both the nectin-1 and HVEM receptors [37] . It has a similar neutralizing antibody titer as MC23 and protected 80% of mice , which was significantly better than MC23 . MC23 binds only to FcγRIII , while DL11 binds to FcγRIII at a higher endpoint titer than MC23 and also binds to FcγRIV . These results support a possible contribution of ADCC to protection mediated by DL11 . An alternate explanation , however , is that protection is mediated by the Fab domain based on DL11 binding to gD2 with 81-fold higher avidity ( slower off rate in Table 1 ) than MC23 . Additional passive transfer studies will be required to assess the relative contributions of the Fab and Fc domains of antibodies targeting crucial gD2 epitopes in protecting animals against genital HSV-2 . Approaches include passive transfer with MAb Fab domains , using antibodies to various murine FcγRs to block their activity , and performing experiments in IgG Fc receptor knockout mice [35 , 39] . To fully evaluate the contribution of the IgG Fc domain in antibody protection , experiments will require assessing the impact of HSV-1 and HSV-2 gE . These glycoproteins function as IgG Fc receptors that greatly diminish the ability of the IgG Fc domain to activate complement and mediate ADCC [32 , 49 , 50] . Greater protection may have been observed during our passive transfer studies if the gE2 Fc binding domains were non-functional . A goal of our gC2/gD2/gE2 trivalent subunit antigen vaccine is to produce antibodies that target crucial functional epitopes with their Fab domains while improving IgG Fc activity by generating antibodies that block the immune evasion properties of gE [26] . The comparison of epitope specific antibody responses in guinea pigs and humans is instructive . The gD2 vaccine protein that we used in guinea pigs has 306 amino acid residues and includes the 281 amino acids used in the gD2 human Herpevac trial [10] . The gD2 vaccine in the Herpevac trial failed to produce significant levels of antibodies that blocked gD2 binding to yellow ( 1D3 ) and brown ( DL6 and MC14 ) epitopes [25] . The response in guinea pigs was more robust than in humans , yet many guinea pigs also failed to produce antibodies to these epitopes . The observation that passive transfer of MAbs to these epitopes partially protected mice against a lethal HSV-2 infection suggests that this deficiency in coverage is important . The 1D3 , DL6 and MC14 epitopes are linear . Therefore , one approach to improve the immunogenicity of these epitopes is to systematically substitute amino acid sequences along the 8–11 amino acids that comprise the epitopes and then assess the immunogenicity of the mutated epitope . Another option to improve epitope immunogenicity is to administer peptides with adjuvants , a third option is to remove glycosylation sites and a fourth is to evaluate different adjuvants [41] . Many vaccine trials in animal models have used gD2 as an immunogen and numerous studies have assessed protection by antibody passive transfer in mice [27 , 28 , 41–43 , 51–59] . Some animal and human studies have also evaluated epitope-specific antibody responses to gD2 immunization [25 , 42] . Our report differs from prior studies in that we correlated antibody responses to crucial gD2 epitopes with protection against genital lesions . Measuring epitope specific antibody responses to crucial glycoprotein epitopes provides a novel approach to improve the accuracy of animal models in predicting outcomes in human trials . The disappointing results of prior human trials suggest that efforts at improving animal models are important . We propose that future phase I/II human immunogenicity trials that include gD2 as an immunogen assess whether high titers of neutralizing antibodies are produced and determine the breadth of the antibody response to an array of four or more crucial epitopes . Preferably , the antibodies produced will include one or more of the linear 1D3 , DL6 and MC14 epitopes that were largely absent in the Herpevac Trial for women . Attempts to modify the immunogen or adjuvant are warranted prior to large phase III human trials if deficiencies are noted .
Female Hartley strain guinea pigs and female BALB/c mice were obtained from Charles River Laboratories ( Wilmington , MA ) . The animals were used for studies in accordance with protocol No . 805187 approved by The Institutional Animal Care and Use Committee of the University of Pennsylvania . The protocol followed recommendations in the Institute for Laboratory Animals Research’s “Guide for the Care and Use of Laboratory Animals . ” Subcutaneous saline was given to animals that appeared dehydrated . Meloxicam was used to control pain once genital lesions developed . For euthanasia , Euthasol was administered to guinea pigs while CO2 was administered to mice according to the recommendations of the Panel on Euthanasia of the American Veterinary Medical Association . Sera were obtained from two prior experiments performed in gD2 immunized guinea pigs [26] . In that study , animals were bled two weeks after the third immunization and challenged intravaginally with 5x105 PFU of HSV-2 strain MS . Days with genital lesions were recorded post infection for 60 days . HSV-2 strain MS was grown and titered on African green monkey kidney ( Vero , ATCC CRL-1586 ) cells [60] . Baculovirus gD2 ( 285t ) and gD2 ( 306t ) were produced in insect sf9 cells and purified from the supernatant fluids [31 , 60] . Glycoprotein D-specific MAbs have been previously described [30] . IgG was purified from guinea pig sera taken two weeks after the final immunization using a Protein A Spin Plate for IgG according to the manufacturer's instructions ( Thermo Scientific , Pierce Biotechnology , Rockford , IL ) . IgG-containing fractions were stored at −80°C . The protein concentration of purified IgG was determined using the Micro BCA Protein Reagent Kit ( Thermo Scientific , Pierce Biotechnology , Rockford , IL ) . Neutralizing antibody titers were measured in the absence of complement as described in our prior publication [26] . Endpoint ELISA titers were performed as previously described except serial two-fold dilutions of purified IgG were added to the plates instead of a 1:1000 dilution of serum [26] . Endpoint titers were calculated by regression analyses as the IgG dilution giving an OD reading 2-fold higher than background . Human epidermal cells ( HaCaT ) ( provided by H . Diggelmann , ISREC , Switzerland ) were grown to confluence on 48-well tissue culture plates ( Costar ) [61] . Each well was inoculated with approximately 30 PFU of HSV-2 strain MS . Virus infection was allowed to proceed for 1 h at 37°C , then virus that had not entered cells was removed by citric acid wash at pH3 . 0 for 1 min at RT followed by a PBS wash . Serial 4-fold dilutions ranging from 40 μg/ml to 0 . 67 μg/ml of IgG purified from MAbs DL6 , MC14 , MC16 , and from rabbit anti-gE2 polyclonal serum R265 were added to wells in 250 μl of DMEM containing 5% fetal bovine serum and incubated at 37°C in 5% CO2 for 48 h [32] . Monolayers were stained with crystal violet and plaque size measured using an inverted microscope fitted with an eyepiece micrometer [61] . Antibody-dependent cellular cytotoxicity ( ADCC ) endpoint titers were determined using murine FcγRIII and FcγRIV ADCC Reporter Bioassays according to the manufacturer’s instructions ( Promega Corporation , Madison , Wisconsin ) . Briefly , white with clear , flat-bottomed 96-well plates ( Falcon REF 353377 , Corning Incorporated , Big Flats , NY ) containing approximately 1 . 2X104 Vero cells/well were infected with HSV-2 strain MS at an MOI of 2 for 16h . At the time of the assay , 95% of the culture medium was removed from each well and replaced with 25μl of RPMI 1640 supplemented with 4% low IgG serum ( RPMI low IgG serum ) and 25 μl purified gD2-specific MAb or control antibodies that had been serially diluted 5-fold in RPMI low IgG serum , and incubated at 37°C in 5% CO2 for 30 min . Jurkat cells stably expressing murine FcγRIII or FcγRIV were added as effector cells at an effector to target ratio of 6:1 in a volume of 25 μl and incubated at 37°C in 5% CO2 for 6 h . 75 μl of Bio-Glo Reagent was then added to each well , and luminescence was determined using a Luminoskan Ascent Luminometer ( Thermo Scientific ) . Fold-induction was calculated using the following equation: Fold induction = RLU ( induced—background ) / RLU ( no antibody control—background ) . Endpoint titers were calculated as the IgG dilution yielding a ≥2-fold induction . Seven- to eight-week old female BALB/c mice ( Charles River Laboratories , Wilmington , MA ) were treated subcutaneously with 2 mg of medroxyprogesterone 5 days prior to challenge to induce the diestrus stage of the estrus cycle that increases susceptibility to HSV-2 intravaginal infection [62] . One day prior to challenge , mice received a passive transfer of 100 μg of MAb IgG/mouse intraperitoneally ( IP ) . On the day of challenge , the vagina of each mouse was cleared using a sterile cotton swab moistened with PBS , followed by an intravaginal challenge with HSV-2 MS ( 5X103 PFU/mouse , approximately 500 LD50 ) . The virus used for inoculation was titered to ensure accurate dosage ( 5X103 PFU/mouse ) . Mice were followed for survival for 10 days post-infection , weighed daily and scored for genital disease on a scale of 0–4 by assigning one point each for erythema , exudate , hair loss and necrosis of genital tissues [32] . Antibody competitions were performed at RT using the Carterra Microfluidics continuous flow microspotter surface plasmon resonance imaging ( CFM/SPRi ) system [30] . Briefly , a panel of distinct MAbs recognizing gD were amine-coupled to a CDM200M sensor chip ( Xantec GmbH ) in a 96-spot format . The sensor chip was then moved to the SPR imager ( IBIS MX96 ) , blocked with ethanolamine , and primed with a running buffer of PBS-0 . 01% Tween 20 . Antibody competitions were performed in a pre-mix assay format by saturating soluble gD2 ( 285t ) at a concentration of 75 ng/reaction with purified guinea pig IgG at 2 μg/reaction in 200 μl total volume . Each gD2 ( 285t ) :guinea pig IgG mix was flowed across the sensor chip spotted with the gD MAbs . After each gD2 ( 285t ) :guinea pig IgG mix , the chip surface was regenerated using 10 mM glycine pH 2 . 0 prior to flowing the next gD2:guinea pig IgG . The starting and ending ( post-regeneration ) response unit ( RU ) value of zero for each MAb spot indicates that both the guinea pig IgG and the gD2 were removed . Every 10th cycle gD2-only ( without IgG ) was flowed across the chip . The RU value obtained was used to establish the background binding of gD2 to compare with gD2 plus guinea pig IgG . Binding of gD2 establishes that the MAb is still present on the chip and has not lost its capacity to bind gD2 during regeneration . The blocking activity of the guinea pig IgG was calculated for each MAb as a percentage using the formula: [1- ( RU gD2 ( 285t ) + GP IgG ) / ( RU gD ( 285t ) alone ) ]*100 . Purified gD2 ( 285t ) was used to screen for antibodies although gD2 ( 306t ) was used to immunize the guinea pigs . The rationale was based on MAbs in the pink community binding better to gD2 ( 285t ) than to gD2 ( 306t ) , thus making the evaluation more reliable on gD2 ( 285t ) [30] . In addition , the samples from gD2 immunized humans were evaluated using gD2 ( 285t ) , which facilitates comparisons between human and guinea pig responses [25] . | Herpes simplex virus type 2 ( HSV-2 ) glycoprotein D ( gD2 ) mediates virus entry and cell-to-cell spread . We hypothesized that an effective gD2 vaccine needs to block these activities . Neutralizing titers , which assess blocking virus entry , correlated with protection against HSV-2 genital lesions in gD2-immunized guinea pigs . We used a high throughput biosensor assay to evaluate whether gD2-immunized guinea pigs produce antibodies that block the activities of gD2 epitopes involved in virus entry and cell-to-cell spread . We detected a strong correlation between the number of epitopes blocked by immune serum and protection against genital lesions . Passive transfer of monoclonal antibodies in mice infected intravaginally with HSV-2 confirmed the in vivo importance of antibodies to these epitopes . We re-evaluated our prior report of epitope-specific antibody responses in the gD2 Herpevac Trial , a trial that failed to protect subjects against HSV-2 genital infection . We detected antibody responses to significantly fewer epitopes in immunized humans than guinea pigs . We conclude that neutralizing antibodies and antibodies to multiple gD2 epitopes involved in virus entry and cell-to-cell spread represent important immune correlates of protection . Future human trials that contain gD2 as an immunogen need to assess these responses and correct deficiencies prior to embarking on large trials . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"medicine",
"and",
"health",
"sciences",
"immune",
"physiology",
"enzyme-linked",
"immunoassays",
"viral",
"transmission",
"and",
"infection",
"engineering",
"and",
"technology",
"immunology",
"microbiology",
"vertebrates",
"animals",
"mammals",
"animal",
"models",
"vaccines",
"experimental",
"organism",
"systems",
"infectious",
"disease",
"control",
"antibodies",
"immunologic",
"techniques",
"antibody",
"response",
"prototypes",
"research",
"and",
"analysis",
"methods",
"immune",
"system",
"proteins",
"infectious",
"diseases",
"viral",
"entry",
"antibody",
"production",
"proteins",
"technology",
"development",
"immunoassays",
"immune",
"response",
"guinea",
"pigs",
"biochemistry",
"rodents",
"eukaryota",
"virology",
"physiology",
"biology",
"and",
"life",
"sciences",
"amniotes",
"organisms"
] | 2018 | Vaccine-induced antibodies to herpes simplex virus glycoprotein D epitopes involved in virus entry and cell-to-cell spread correlate with protection against genital disease in guinea pigs |
Accurate genome duplication underlies genetic homeostasis . Metazoan Mdm2 binding protein ( MTBP ) forms a main regulatory platform for origin firing together with Treslin/TICRR and TopBP1 ( Topoisomerase II binding protein 1 ( TopBP1 ) –interacting replication stimulating protein/TopBP1-interacting checkpoint and replication regulator ) . We report the first comprehensive analysis of MTBP and reveal conserved and metazoa-specific MTBP functions in replication . This suggests that metazoa have evolved specific molecular mechanisms to adapt replication principles conserved with yeast to the specific requirements of the more complex metazoan cells . We uncover one such metazoa-specific process: a new replication factor , cyclin-dependent kinase 8/19–cyclinC ( Cdk8/19-cyclin C ) , binds to a central domain of MTBP . This interaction is required for complete genome duplication in human cells . In the absence of MTBP binding to Cdk8/19-cyclin C , cells enter mitosis with incompletely duplicated chromosomes , and subsequent chromosome segregation occurs inaccurately . Using remote homology searches , we identified MTBP as the metazoan orthologue of yeast synthetic lethal with Dpb11 7 ( Sld7 ) . This homology finally demonstrates that the set of yeast core factors sufficient for replication initiation in vitro is conserved in metazoa . MTBP and Sld7 contain two homologous domains that are present in no other protein , one each in the N and C termini . In MTBP the conserved termini flank the metazoa-specific Cdk8/19-cyclin C binding region and are required for normal origin firing in human cells . The N termini of MTBP and Sld7 share an essential origin firing function , the interaction with Treslin/TICRR or its yeast orthologue Sld3 , respectively . The C termini may function as homodimerisation domains . Our characterisation of broadly conserved and metazoa-specific initiation processes sets the basis for further mechanistic dissection of replication initiation in vertebrates . It is a first step in understanding the distinctions of origin firing in higher eukaryotes .
Eukaryotic cells must faithfully replicate their genomic DNA exactly once before each cell division in order to ensure genetic homeostasis through successive cell generations . Initiation of DNA replication is a major step of replication control by the cell cycle and the DNA damage checkpoint , ensuring faithful genome duplication in normal and adverse conditions [1 , 2] . In higher eukaryotic cells , we know little about the molecular mechanisms and regulation of replication initiation . Yeast cells have served as a model for understanding initiation in eukaryotes . In the first initiation step , origin licensing , that occurs in the G1 cell cycle phase , prereplicative complexes ( pre-RCs ) form dependently on the origin recognition complex ( ORC ) , cell division cycle 6 ( Cdc6 ) , and Cdc10-dependent transcript ( Cdt1 ) . Pre-RCs are inactive double hexamers of the Mcm2-7 helicase , comprising the six minichromosome maintenance ( Mcm ) proteins 2–7 , loaded on origin DNA . In S phase , high activity of S phase cyclin-dependent kinase ( S-CDK ) and dumbbell former 4 ( Dbf4 ) –dependent kinase ( DDK ) induces origin firing , the conversion of pre-RCs into two bidirectional replisomes , each containing a Mcm2-7 hexamer , Cdc45 , and the go-ichi-ni-san ( GINS ) complex . This Cdc45-Mcm2-7-GINS ( CMG ) complex is the active replicative helicase [3] . DDK phosphorylates pre-RCs , which recruits the synthetic lethal with Dpb11 3/7 ( Sld3-Sld7 ) complex to pre-RCs , which in turn interacts with and recruits Cdc45 [4–6] . S-CDK facilitates origin firing by phosphorylating Sld3 and Sld2 [7–9] . Phospho-Sld3 interacts with DNA polymerase binding 11 ( Dpb11 ) , which is part of the preloading complex ( pre-LC ) [10] . This interaction is thought to locate pre-LCs at origins . Pre-LCs also contain Sld2 , polymerase epsilon , and the GINS tetramer . Pre-LC formation depends on S-CDK that phosphorylates Sld2 to induce the interaction of Sld2 with Dpb11 [11] . Formation of active replisomes further requires Mcm10 , whose function is less clear [12 , 13] . Biochemical reconstitution demonstrated that these yeast core initiation factors are sufficient to initiate bidirectional DNA replication [14–16] . Metazoan orthologues for each of the yeast core factors have previously been identified , except for Sld7 . Sld7-deleted cells are viable but have severe replication defects [17] . In vitro replication initiates in the absence of Sld7 [6] . Whether the dispensability of Sld7 in vitro reflects an indirect role of Sld7 in origin firing in vivo or whether it is a consequence of the special biochemical in vitro conditions is unclear . Sld7 forms a stable complex with Sld3 in vivo and in vitro [17 , 18] . Yeast cells require the interaction of Sld7 with Sld3 to replicate efficiently , because Sld7 binding mutants of Sld3 are sensitive to hydroxyurea , just like sld7Δ cells [17] . Conservation of most , but potentially not all , origin firing factors between yeast and metazoa suggested that although many fundamental initiation processes are conserved , processes that are specific for higher eukaryotes might have evolved . All metazoan orthologues of core yeast firing factors are essential for replication , as studies in Xenopus egg extracts and human cells have shown , but their molecular functions are largely unexplored [19–24] . The Sld3 orthologue , Treslin/TICRR ( Topoisomerase II binding protein 1 ( TopBP1 ) –interacting replication stimulating protein/TopBP1-interacting checkpoint and replication regulator ) [25] , utilises conserved domains for S-CDK–dependent interaction with the Dpb11 orthologue TopBP1 [24 , 26 , 27] . The Mdm2 binding protein ( MTBP ) protein was the last metazoan firing factor identified and described to be required for firing in human cells [28] . It did not fit a universal model of eukaryotic replication because , despite our extensive efforts , no homology with yeast initiation proteins was detected . MTBP is reminiscent of Sld7 in its binding to Treslin/TICRR/Sld3 . This binding appears essential for replication as MTBP nonbinding Treslin/TICRR mutants did not facilitate replication . These functional similarities of MTBP and Sld7 , and similarities in protein sequence and structure of the C termini [29] led to the hypothesis that MTBP performs Sld7-like functions in metazoans . However , no statistically significant evidence for orthology between MTBP and Sld7 has been provided . We here employed various approaches to search for remote homologies in the MTBP and Sld7 proteins . These revealed MTBP to possess two Sld7-homologous regions in its N and C termini , and a metazoa-specific region separating these two homology domains . We show that the Sld7-homologous domains are required for proper replication origin firing in human cells . We thus incontrovertibly demonstrate orthology between MTBP and Sld7 . This fills the last gap in the list of metazoan core origin firing factors , establishing a universal framework of eukaryotic replication initiation . Despite this conservation , metazoa have also evolved specific initiation processes , because the metazoa-specific middle domain of MTBP proved to be required for proper DNA replication . This domain apparently harbours more than one activity important for replication . Cyclin-dependent kinase 8/19–cyclinC ( Cdk8/19-cyclin C ) , a protein that was not previously implicated in DNA replication , with roles in controlling transcription [30] , binds the metazoa-specific MTBP domain . This interaction was required for complete genome replication and , consequently , for normal chromosome segregation . We hypothesise that the metazoa-specific binding of Cdk8/19-cyclin C to MTBP helps integrate the conserved initiation principles into the special requirements of the more complex metazoan cells to achieve well-regulated origin firing to guarantee genome stability .
Human MTBP ( hMTBP ) is surprisingly devoid of known domain homologues . To identify its domain architecture , we initiated an exhaustive computational sequence analysis . We identified three domains that are conserved in MTBP orthologues across most of the animal kingdom . Two of these domains proved conserved in yeast Sld7 ( Fig 1A ) . For this we employed iterative profile-based sequence similarity searches [31] of the UniRef50 database [32] . Focusing first on the most C-terminal of these regions , we found that its sequences are statistically significantly similar to the C terminus of Saccharomyces cerevisiae Sld7 of known tertiary structure ( protein data bank [PDB] identifier , 3×38 ) [18] ( E-value = 0 . 012; probability = 94 . 1% ) ( Fig 1B ) . This step employed HHpred searches against the PDB70 profile database [33] and a C-terminal conserved region , amino acids 824–896 of hMTBP , as input . Consistent with these C-terminal domains of MTBP and Sld7 being homologous , secondary structure predictions of a three-helix bundle for the former agree with the published Sld7 crystal structure [18 , 34] . Thus , the C-terminal domains of MTBP and Sld7 have evolved from a common ancestral sequence . We call them the Sld7-MTBP C-terminal domain ( S7M-C ) . Our findings provide the statistical evidence required to support a recent proposal of MTBP-Sld7 C-terminal domain homology [29] . Two complementary approaches then provided evidence for MTBP and Sld7 sharing an N-terminal homologous domain . The first approach queried the protein homology/analogy recognition engine V 2 . 0 ( phyre2 ) server with MTBP sequences and returned low confidence alignments with yeast Sld7 ( S1A Fig ) [18] . We term these amino acids the MTBP-phyre2 and Sld7-phyre2 region . Six amino acids in the Sld7-phyre2 region directly contact Sld3 to form Sld7-Sld3 dimers ( L45 , V48 , 52-KLPL-55 of Zygosaccharomyces rouxii Sld7; S1A Fig , blue asterisks; S2 Fig ) [18] , and four of them are conserved in MTBP ( V306 , I309 , L314 , P315 ) with respect to their chemical properties . These MTBP amino acids are among the most highly conserved residues in this region across animals ( S1B Fig ) . We tested next if these amino acids in the MTBP-phyre2 region are important for binding to Treslin/TICRR . We deleted the phyre2 region ( amino acids V295-T329 ) of hMTBP ( MTBP-Δphyr2 ) and tested its interaction with endogenous Treslin/TICRR in cell lysates after transient transfection of MTBP-Flag into 293T cells . Flag immunoprecipitation ( IP ) ( see Table 1 for all antibodies used ) of wild-type ( WT ) MTBP-Flag ( MTBP-WT ) , but not MTBP-Δphyr2 , co-purified Treslin/TICRR ( Fig 2A , lanes 1 and 2 ) . A quintuple point mutant ( MTBP-5m ) exchanging the MTBP-phyre2 region amino acids V306 , I309 , D313 , L314 , and P315 against alanine ( D313 ) or aspartate ( all others ) also showed no detectable binding to Treslin/TICRR ( lane 3 ) . These five residues map to Sld3-contacting amino acids in Sld7 ( Figs 1C and S2 ) . MTBP-Δphyr2 and MTBP-5m were specifically defective in binding to Treslin/TICRR but bound Cdk8 , a new MTBP interactor , whose function in replication we discuss below , as well as MTBP-WT , suggesting that the mutants are not misfolded . To assess further the folding quality of the MTBP-5m protein , we tested its migration behaviour in gel filtrations . We found that MTBP-WT and MTBP-5m eluted indistinguishably from each other as sharp peaks and did not form aggregates ( S3 Fig ) . MTBP-5m localises predominantly to the nucleus , like MTBP-WT ( S4 Fig ) . We realised that in some gels Treslin/TICRR in samples from cell lysates formed multiple bands representing different phosphorylation forms , whereas MTBP-bound Treslin/TICRR usually showed only one band ( Fig 2B left panel ) . To test if MTBP-bound Treslin/TICRR is phosphorylated we ran lambda phosphatase ( PPase ) -treated and untreated lysates side by side with MTBP immunoprecipitates on an SDS polyacrylamide gel to compare gel mobility of Treslin/TICRR . MTBP-bound Treslin/TICRR showed a PPase-dependent mobility shift ( S5 Fig ) . To fine-map the interaction between MTBP and Treslin/TICRR , we then mutated amino acids V306 to S316 in the MTBP-phyre2 region individually . MTBP-Flag IP showed that four of the five amino acids aligning with Sld3-interacting amino acids , V306 , I309 , L314 , and P315 ( S2 Fig ) , contribute to Treslin/TICRR binding ( Fig 2A , lanes 4 , 7 , 12 , 13 ) . Mutating other amino acids that were not aligned with Sld3-interacting Sld7 amino acids had variable effects . Mutating the most conserved , L311 ( lane 9 ) , showed a considerable reduction of Treslin/TICRR binding . Mutating the moderately conserved M308 ( lane 6 ) had mild effects , whereas mutating the moderately conserved Q307 ( lane 5 ) and the weakly conserved K310 and S312 ( lanes 8 and 10 ) did not alter Treslin/TICRR interaction . All mutants bound Cdk8 normally . These data show that the phyre2 regions of Sld7 and MTBP have a conserved function , the binding to Sld3/Treslin/TICRR . To identify homology between MTBP and Sld7 N termini , we then used the newly identified MTBP-phyre2 region as an anchor point for similarity searches based on amino acid sequences and secondary structures . This revealed extensive sequence similarities between amino acids V295 and G426 in the N terminus of MTBP and the full length of the N-terminal domain of Sld7 ( Figs 1C and S2A ) . These similarities are sufficient to predict homology because their alignment score greatly exceeds those expected from alignments with random sequences ( E = 3 . 0 × 10−5 ) . We term these domains in Sld7 and MTBP Sld7-MTBP N-terminal ( S7M-N ) domains . Independent evolution of shared domain architectures is rare [35] and highly unlikely for functionally similar proteins . Our evidence thus strongly indicates that metazoan MTBP and fungal Sld7 are orthologues . Regions outside MTBP-phyre2 are also important for interaction with Treslin/TICRR . Itou and colleagues showed that the S7M-N domain of Sld7 contains a second cluster of Sld3-contacting amino acids ( Fig 1C ) [18] . The corresponding positions in the S7M-N domain of hMTBP are amino acids C413–G426 , and were also required for interaction with Treslin/TICRR ( Fig 2B lanes 5–11 ) . MTBP-Q423-G426 ( mutated in MTBP-m3–6 ) and C413 ( MTBP-m1 ) map to positions in Sld7 that directly contact Sld3 ( S2 Fig ) . A415 ( MTBP-m2 ) probably affects the stability of its β-strand ( S2 Fig ) . Finally , deleting amino acids D2 to S55 , D2 to S31 or D2 to V8 from the N terminus also resulted in MTBP mutants that failed to bind Treslin/TICRR ( Fig 2B lanes 1–5 ) . MTBP-ΔD2-V8 and mutants of the region between amino acids C413 and G426 do not show signs of protein misfolding . Their Cdk8 binding ability ( Fig 2B ) and migration behaviour in gel filtrations were indistinguishable from MTBP-WT ( S3 Fig ) . MTBPm1–6 localises predominantly to the nucleus , like MTBP-WT ( S4 Fig ) . We conclude that various regions in the N terminus of MTBP contribute to Treslin/TICRR binding , and termed it Treslin/TICRR binding domain ( TresBD ) . We next investigated if the two Sld7-like domains of MTBP , S7M-N and S7M-C , are important for DNA replication . We used an established RNA interference ( RNAi ) system [28] ( see Table 2 for all small interfering RNAs [siRNAs] used ) to replace MTBP with siRNA-resistant MTBP-WT or mutant MTBP transgenes ( Fig 3A ) . Relative replication speed was then determined by 5-bromodeoxyuridine ( BrdU ) labelling and flow cytometry ( Fig 3B and 3C , gating strategy for all flow cytometry experiments , including raw data for quantification in S1 Data ) . In control cells expressing no transgene MTBP-RNAi ( siMTBP ) reduced the BrdU signal 2 . 9-fold compared with control RNAi ( siCtr ) ( Fig 3C and S1 Data ) , indicating effective suppression of replication . siRNA-resistant MTBP-WT rescued replication in siMTBP-treated cells almost completely ( 1 . 2× reduction ) . In contrast , Treslin/TICRR nonbinding S7M-N mutants , MTBP-5m ( phyre2 region ) and MTBP-m1–6 ( amino acid C413–G426 region ) , did not rescue replication . These mutants’ BrdU–propidium iodide ( PI ) profiles were indistinguishable from siMTBP-treated control cells . BrdU signals were reduced 3 . 0-fold ( MTBP-5m ) and 2 . 6-fold ( MTBP-m1–6 ) . We sought to use minimally invasive single point mutations to confirm these conclusions that were drawn from the MTBP-5m and m1–6 combination mutants . The three non-Treslin/TICRR-binding MTBP mutants I309D , L314D ( both mutated in 5m ) and A415Q ( mutated in m1–6 ) did not support replication over background , whereas cells expressing the Treslin/TICRR-binding proficient D313A replicated normally ( S6 Fig and S1 Data ) . To test which step of DNA replication was defective in Treslin/TICRR nonbinding MTBP mutants , we tested if origin licensing and origin firing occurred in cells expressing these mutants . Immunoblotting of chromatin fractions indicated normal licensing , as judged by Mcm5 ( pre-RC subunit ) signals ( Figs 3D and S7 and S1 Data ) . In contrast , Cdc45 and Sld5 ( GINS subunit ) antibodies indicated that CMGs did not form in the presence of the MTBP mutants , unlike with MTBP-WT . Together , we conclude that the Treslin/TICRR-binding S7M-N region is essential for MTBP function in replication at the origin firing step . These conclusions are consistent with earlier results showing that Treslin/TICRR mutants that do not bind MTBP are replication incompetent [28] . We reported before that binding of Treslin/TICRR to MTBP is required for normal cellular levels of MTBP [28] ( S8E and S8F Fig and S1 Data ) . Consistently , the Treslin/TICRR binding-deficient MTBP-5m and m1–6 transgenes , but not other mutants that can bind Treslin/TICRR , expressed to lower levels than MTBP-WT ( Figs 3A and S8A and S6A and S1 Data ) . Cycloheximide shut-off experiments with MTBP-5m and m1–6 confirmed that these degrade faster than MTBP-WT ( S8B Fig ) . Lower MTBP levels cannot be the sole reason for the lack of replication in MTBP-5m and m1–6 cells because partial siMTBP-mediated knock-down experiments showed that strong replication defects required suppression of MTBP signals to less than approximately 15% , whereas the 5m and m1–6 mutants have approximately 40% and 35% of MTBP signals left compared with endogenous and transgenic MTBP-WT ( S8A , S8C and S8D Fig and S1 Data ) . We then tested if elevating the expression levels of the MTBP-5m and m1–6 to MTBP-WT levels rescues their replication deficiency . To this end we selected stable Hela-Kyoto cell clones generated by random transgene integration into the genome that expressed MTBP-5m , m1–6 , or WT at comparable levels . Replication analysis showed very low if any replication activity of the MTBP mutants ( S9 Fig and S1 Data ) , indicating Treslin/TICRR binding to MTBP has essential roles in replication distinct from or in addition to stabilising MTBP levels . The S7M-C domain is also important for replication . Two MTBP mutants lacking the C-terminal 150 ( MTBP-ΔC150 ) or 81 amino acids ( MTBP-ΔS7M-C ) showed defects in rescuing replication . BrdU signals were reduced by 1 . 7-fold in MTBP-ΔC150 and 1 . 8-fold in MTBP-ΔS7M–C ( Fig 3F and S1 Data ) . Because these replication defects were milder compared with the Treslin/TICRR-binding mutants , we quantified replication of these mutants more thoroughly . Averaging multiple independent experiments confirmed that both mutant cell lines replicated significantly less than MTBP-WT–expressing lines ( 39% and 27% of WT ) ( Fig 3G and S1 Data ) . Both mutant proteins bound Treslin/TICRR normally ( S10A Fig ) . Moreover , chromatin immunoblots showed that , whereas licensing occurred normally in cells expressing MTBP-ΔS7M-C , origin firing ( CMG formation ) was impaired , albeit not as much as in Treslin/TICRR-nonbinding MTBP mutants ( Figs 3D and S7 Fig and S1 Data ) . Together , we conclude that the S7M-C domain is required for MTBP function in replication origin firing . The S7M-C function is distinct from Treslin/TICRR binding , consistent with a recent report [29] . The Sld7-Sld3 crystal structure suggested that the S7M-C domain of Sld7 mediates Sld7 dimerisation [18] . The relevance of Sld7 dimerisation for replication in yeast cells was not tested . If the S7M-C domain of MTBP mediates replication by homodimerisation , fusing a homodimerisation domain to MTBP that lacks the S7M-C domain should rescue the capability of this MTBP mutant to induce replication . We attached MTBP-ΔC150 and MTBP-ΔS7M-C with their C termini to the homodimerising glutathione S transferase ( GST ) tag . Cells expressing MTBP-ΔC150-GST and MTBP-ΔS7M-C-GST transgenes replicated to near WT levels , 80% and 99% , respectively , whereas fusing the nondimerising green fluorescent protein ( GFP ) tag had little effect ( Figs 3G and S10B and S1 Data ) . The slightly better replication by the GFP fusion ( 49% ) over MTBP-ΔC150 ( 39% ) might stem from its higher expression level . The most parsimonious explanation of the replication rescue by fusing GST is that GST replaces a dimerisation activity of the C terminus , and there is no other essential replication function of the C terminus . However , alternative , more complicated scenarios cannot be excluded . We hypothesised that the Sld7-homologous domains of MTBP confer the core replication activities common across eukaryotes , whereas the metazoa-specific MTBP middle region may mediate metazoa-specific roles in replication , or may have nonreplicative functions . To test the relevance of the MTBP middle domain for replication , we sought to delete the whole middle domain but retain the activities of the S7M termini . The smallest TresBD fragment of MTBP that was biochemically well behaved ( was soluble and showed no unspecific binding to control pull-down beads ) and retained full Treslin/TICRR binding activity contained amino acids M1 to Y513 ( S11A Fig ) . We fused to it amino acids L705 to K904 , Y770 to K904 , or Q810 to K904 ( Fig 4A ) . We call these mutants metazoan Sld7s ( mSld7s ) because they contain little more than the Sld7-equivalent MTBP termini . All three mSld7 versions supported replication poorly compared with MTBP-WT ( Fig 4B and S1 Data ) but better than cells carrying no transgene , cells expressing the TresBD ( amino acids M1–D515 ) ( Fig 4C and 4D and S1 Data ) , and cells that lack TresBD activity ( Fig 3A–3C and S1 Data ) . This shows that the Sld7-equivalent termini cooperate in mSld7s to mediate replication , albeit to sub-WT levels . Because MTBP-ΔA516-H769 and ΔA516-G809 lack an nuclear localisation sequence ( NLS ) around amino acid D740 , a simian virus 40 ( SV40 ) –NLS had been fused to the C terminus of the C-terminal GFP tag . All mSld7 versions bound Treslin/TICRR . ΔA516–P704 and ΔA516–H769 , but not ΔA516–G809 , showed some reduction of Treslin/TICRR association compared to WT ( S11B Fig ) . Because the C-terminal 388 amino acids of MTBP ( amino acids A516–K904 ) are dispensable for Treslin/TICRR binding ( S11A Fig ) , this indicates that the artificial domain boundaries in these two mutants may affect folding of the N-terminal TresBD to some degree . The relatively weak Treslin/TICRR binding of MTBP-ΔA516–P704 is probably the reason for the slightly lower expression of this transgene . We do not think , however , that lower Treslin/TICRR binding capacity and lower MTBP protein levels are critical for the low replication in these mutant cells , because the Treslin/TICRR binding capacities of the mSld7 versions do not correlate with their replication-promoting activities ( Figs 4B and S11B ) . We conclude that the metazoa-specific middle domain of MTBP and the Sld7-homologous domains cooperate to support DNA replication in human cells . We found , using purification of MTBP-Flag-GFP and mass spectrometry , that the Cdk8 and Cdk19 kinases and their activating subunit cyclin C co-purify with MTBP . Cdk8 and Cdk19 show extremely high sequence identity . In our experiments they behaved identically , and we do not functionally distinguish between them . Endogenous MTBP co-purified Cdk8 , cyclin C , and Cdk19 , and anti-cyclin C antibodies precipitated Cdk8 , Cdk19 , and MTBP ( Fig 5A ) . The interaction occurred in lysates from asynchronous cells and from cells synchronised in G1 , S , or mitosis ( Fig 5B ) . The slightly lower MTBP amounts in G1 cell lysates suggested that the interaction with Cdk8/19-cyclin C might require cell cycle CDK kinase activity , but adding Cdk2-Cyclin A to G1 lysates did not promote the interaction . Treslin/TICRR also co-purified with anti-Cdk8 antibodies from all cell cycle stages ( Fig 5B ) . Moreover , transiently transfected Treslin/TICRR-WT and mutants of the Treslin/TICRR termini , the Sld3/Treslin domain ( STD ) , and the TopBP1 interaction domain coimmunoprecipitated Cdk8/19-cyclin C ( Fig 5C ) . In contrast , two deletion mutants of the Treslin/TICRR M domain that are partially ( ΔM1 ) or strongly ( ΔM2 ) deficient in MTBP interaction showed proportional defects in Cdk8 binding . These interaction studies suggested that MTBP , CDK8/19-cyclin C , and Treslin/TICRR form a protein complex , with MTBP bridging Treslin/TICRR and the Cdk8 kinase ( Fig 5H ) . Corroborating this interpretation , a C-terminal MTBP-A516–K904 fragment , but not the TresBD fragment , interacted with Cdk8/19-cyclin C ( S12B ( i ) Fig ) , placing the binding site for the kinase between MTBP amino acids A516 and K904 . TopBP1 is also part of the Treslin/TICRR-MTBP-Cdk8/19-cyclin C complex . IP of endogenous TopBP1 from cell lysates co-purified Treslin/TICRR , MTBP , and a faint but specific signal for Cdk8 ( Fig 5D ) . Adding Cdk2-cyclin A to the lysate to enhance the binding of Treslin/TICRR to the N-terminal triple breast cancer type 1 susceptibility protein C terminal repeat ( BRCT ) repeat domain of TopBP1 [26] increased the signals for Treslin/TICRR , MTBP , Cdk8 , Cdk19 , and cyclin C , indicating that TopBP1 indirectly binds MTBP-Cdk8/19-cyclin C via Treslin/TICRR ( Fig 5H ) . Cdk8-cyclin C forms the kinase module of the mediator of transcription together with mediator of transcription subunits 12 and 13 ( Med12 and Med13 ) . However , Med12 and -13 do not associate with the MTBP-bound fraction of Cdk8/19-cyclin C: whereas anti-Cdk8 antibodies co-purified MTBP , Med12 , and Med13 ( Fig 5E ) , the same amount of Cdk8 purified with anti-MTBP antibodies did not contain Med12 and Med13 detectably . We conclude that Cdk8/19-cyclin C forms distinct protein complexes with the mediator of transcription and the origin firing regulator Treslin/TICRR-MTBP-TopBP1 . We then mapped the interaction site for Cdk8/19-cyclin C in MTBP-A516–K904 . Binding studies using truncated MTBP fragments in cell lysates showed that amino acids R595 to P704 of MTBP are required and sufficient for binding the Cdk8 kinase ( Figs 5F and S12A and S12B ) . The R595–P704 region contains a particularly well-conserved metazoa-specific domain ( Fig 5H; blue oval; amino acids 636–693 , S12D Fig ) . Deleting only this domain ( MTBP-ΔT635–P704 ) abrogated Cdk8 binding , as did a mutant lacking amino acids R595 to L634 ( Fig 5F ) . A series of MTBP point mutants subsequently showed that amino acids between L601 and E605 as well as L620 and T635 were important for MTBP binding to Cdk8/19-cyclin C ( Fig 5G and S12C ) . In order to create a maximally Cdk8/19-cyclin C binding-deficient MTBP mutant for subsequent functional analyses , we combined seven amino acid exchanges between amino acids L620 and T635 to generate MTBP–Cdk8 binding mutant ( Cdk8bm ) ( Fig 5G ) . MTBP-Cdk8bm bound Treslin/TICRR normally and behaved like MTBP-WT in gel filtrations ( S3 Fig ) and nuclear localisation ( S4 Fig ) . Next , we asked whether Cdk8/19-cyclin C , which is not amongst the yeast core firing factors , is a novel replication initiation factor in vertebrate cells . Because Cdk8 and Cdk19 may act redundantly , we knocked down cyclin C ( siCycC; cyclin C siRNA ) to test an involvement in DNA replication . BrdU–flow cytometry analysis revealed a mild reduction of replication speed in siCycC-treated HeLa-Kyoto and also in U2OS cells ( S13A–S13C Fig and S1 Data ) . Because Cdk8/19-cyclin C regulates the transcription of many genes as part of the mediator complex , eliminating Cdk8/19-cyclin C from cells will generate complex phenotypes , leading to difficulties in differentiating primary and secondary defects . Therefore , we sought to separate mediator-dependent from -independent replication functions of Cdk8/19-cyclin C . To this end we tested if the binding of the Cdk8/19 kinase to MTBP is required for DNA replication . Because MTBP does not interact with mediator subunits ( Fig 5E ) , Cdk8 kinase nonbinding MTBP mutants should reflect mediator-independent functions of the kinase . BrdU-incorporation analysis of siMTBP-treated cells expressing the Cdk8/19-cyclin C nonbinding MTBP-Cdk8bm and -ΔR595–P704 mutants showed a reduction of replication rescue of around 23% to 77% and 17% to 83% , respectively , compared with MTBP-WT ( Fig 6A and 6B and S1 Data ) . This decrease of replication was more moderate than in siCycC cells ( S13B and S13C Fig and S1 Data ) . We sought to confirm this mild replication reduction by a more sensitive method . Compromised replication typically increases the frequency of fragile metaphase chromosome sites ( FS ) . FS are under-replicated chromosome regions . A control experiment confirmed that FS frequency can be used to monitor replication defects by lower MTBP activity . HeLa cells treated with siMTBP , which reduces replication severely , displayed 5 . 6-fold more FS than siCtr-treated cells ( S13D Fig and S1 Data ) . siMTBP-treated cells expressing the Cdk8/19-cyclin C binding-deficient MTBP mutants showed about two times more FS than siCtr-treated cells and siMTBP-treated cells expressing MTBP-WT ( Fig 6C and S13E and S13F and S1 Data ) . We conclude that cells expressing Cdk8/19-cyclin C binding-deficient MTBP enter mitosis with incompletely replicated chromosomes . Such premature mitotic entry should increase aberrant mitotic figures . We fixed and stained siCtr- or siMTBP-treated MTBP-WT and Cdk8 binding-deficient cell lines with anti-tubulin antibodies and double stranded DNA ( dsDNA ) stain to visualise mitotic spindles and chromosomes . siMTBP-treated MTBP-Cdk8bm and -ΔR595–P704 cells showed more anaphases with DNA that was not pulled to the cell poles together with the remaining chromatin mass ( Fig 6D ( i ) and S1 Data ) . Subsequent cytokinetic abscission was delayed , as judged by more cells that had already formed interphase nuclei but that were still connected by a thin connection containing the spindle midbody ( ii ) . Delayed abscission upon incomplete replication has been described [36] . Occasionally , these connections contained Hoechst-positive structures . In line with chromosome segregation errors , we also found a higher frequency of micronuclei in MTBP-Cdk8bm and -ΔR595–P704 cells ( iii ) . MTBP-ΔR595–P704 appears to have slight dominant effects , as suggested by the mutant cell lines showing a minor increase of aberrant anaphases upon siCtr treatment . We did not observe a compromised spindle assembly checkpoint response of MTBP-Cdk8bm and -ΔR595−P704 cells , as was reported for MTBP knock-down cells [37] . Nevertheless , we cannot exclude that a direct role of MTBP in mitosis contributes to the observed phenotypes . That the observed DNA replication defect of cells expressing Cdk8/19-cyclin C nonbinding MTBP was mild was confirmed by the similar cell cycle distributions of siMTBP-treated MTBP-Cdk8bm and MTBP-WT cells ( Fig 6E and S1 Data ) . Also , the temporal replication program was not grossly affected in MTBP-Cdk8bm cells . Classification of replication patterns using GFP–proliferating cell nuclear antigen ( PCNA ) in live cells expressing MTBP-WT and -Cdk8bm showed a normal appearance and order of early , mid– , and late–S phase replication patterns ( Fig 6F and S1 Data ) . Also , in-detail inspection of mid–and late–S phase patterns revealed no differences ( Fig 6G and S1 Data ) . We conclude from Fig 6 that cells expressing Cdk8/19-cyclin C nonbinding MTBP mutants have moderate defects in genome replication and enter mitosis prematurely with under-replicated chromosomes . The metazoa-specific MTBP middle region appears to harbour at least one more activity in addition to Cdk8/19-cyclin C binding that is required for replication . This activity is situated outside amino acids R595 to P704 . MTBP-Cdk8bm and -ΔR595–P704 , which are both maximally Cdk8/19-cyclin C binding deficient , show less severe defects in replication ( around 20% reduction ) than the mSld7 mutants ( around 70% ) that have larger deletions in the middle domain .
The orthology between MTBP and Sld7 described here is the final proof that all core factors for origin firing are conserved across fungi and animals ( S14 Fig ) . The set of eukaryotic factors sufficient for conversion of pre-RCs into bidirectional replisomes with leading and lagging strand synthesis was defined by biochemical reconstitution with yeast proteins [16 , 38 , 39] . Whether these factors are also sufficient in higher eukaryotes now requires in vitro reconstitution of metazoan initiation from purified components . Despite conservation of the core origin firing factors in eukaryotes , diversions of origin firing process occur in some species , indicating some evolutionary plasticity . For example , in fission yeast and Caenorhabditis elegans , Sld7/MTBP has not been found . Furthermore , the TopBP1 equivalent of the BRCT3/4 repeat domain of yeast Dpb11 that facilitates origin firing by interacting with CDK-phosphorylated Sld2 does not have an essential replication role in Xenopus egg extracts [24] . Instead , TopBP1-BRCT3 is essential [24] . Its role is unknown but it probably does not bind CDK-phosphorylated recombination helicase Q like 4 ( RecQ4 ) , metazoan Sld2 , because TopBP1-BRCT3 lacks key amino acids for phosphopeptide binding [40] . Metazoan cells are more complex than yeasts and will probably require more sophisticated regulation of replication initiation . The metazoa-specific domains of MTBP and of other firing factors like Treslin/TICRR [25] may help integrate the conserved fundamental initiation processes into metazoan cells . Metazoa have less precisely defined origins than yeast , but they face the same problems as any cell with multiple replication start sites . Genome duplication needs to be complete before mitosis to minimise genome instability . This necessitates that no inter-origin distance is bigger than the amount of DNA that can be replicated in S phase . To avoid large random gaps , appropriately located origins must fire with the right timing . The organisation of metazoan replication in replication factories , in which neighbouring origins fire with near synchrony , may help prevent large random gaps [41] . The factory organisation may also help facilitate differential regulation of origin firing between different factories upon DNA damage . In these circumstances so-called dormant origins ( that are inactive in normal growth conditions ) fire in actively replicating factories after DNA damage , whereas origin firing becomes inhibited in nonreplicating factories . Both regulations help prevent genetic instability upon replication stress [42] . MTBP is a promising candidate for mediating some of these regulations , as it forms a main regulation platform of origin firing in eukaryotes together with Treslin/TICRR/Sld3 and TopBP1/Dpb11 . All complex members are subject to regulations by various pathways to mediate appropriately timed and localised replication initiation . These regulations mediate ( 1 ) S phase specificity of replication via CDK and DDK [5 , 8 , 9] , ( 2 ) replication timing of centromeres and other genome regions [43–45] , ( 3 ) inhibition of origin firing upon DNA damage by S phase checkpoint kinases [46–48] , and ( 4 ) firing inhibition in response to low cellular glucose levels involving the silent information regulator ( SIRT1 ) deacetylase [49 , 50] . These regulations all involve Sld3/Treslin/TICRR , Dpb11/TopBP1 , or Sld2/RecQ4 , and at least some are conserved between yeast and man [26 , 27 , 51] . How MTBP activity is controlled is unknown . Several posttranslational modification sites in MTBP were identified by high-throughput approaches but have not been functionally investigated , among them potential ataxia teleangiectasia mutated related/mutated ( ATR/M ) and CDK sites . These are good candidates to mediate cell cycle and checkpoint regulations . Cdk8/19-cyclin C binding to MTBP could also be involved in mediating firing control ( see below ) . We show that the S7M-N domain of MTBP mediates Treslin/TICRR binding . As with yeast Sld7 and Sld3 , complex formation of MTBP with Treslin/TICRR is important for MTBP’s replication function , as several nonoverlapping point and deletion mutants of MTBP that are deficient in Treslin/TICRR binding failed to support replication in HeLa cells . The same was true for MTBP interaction-deficient mutants of Treslin/TICRR [28] . Treslin/TICRR binding-deficient MTBP mutants , for example MTBP-5m and m1–6 , showed reduced MTBP levels in cell lysates and had lower levels of endogenous Treslin/TICRR . The lower levels of the MTBP mutants were at least partly due to their higher degradation rate compared with MTBP-WT ( S8A and S8B Fig ) . This corroborates earlier RNAi experiments suggesting that Treslin/TICRR and MTBP stabilise each other’s cellular levels [28] ( S8C and S8F Fig and S1 Data ) : Treslin/TICRR RNAi reduced MTBP levels and vice versa , and this was rescued by RNAi-resistant proteins . Mutual stabilisation is also seen with Sld3 and Sld7 , and it is important for replication in yeast [17] . Lower Treslin/TICRR levels may make a small contribution to the replication defects in human cells expressing Treslin/TICRR nonbinding MTBP mutants . However , Treslin/TICRR stabilisation cannot be the only function of MTBP in replication , because ( 1 ) deleting the MTBP middle and the S7M-C domains results in a fully Treslin/TICRR–binding-proficient MTBP1-515 protein that is , however , incapable of inducing replication over background , ( 2 ) elevating the expression levels of the non-Treslin/TICRR–binding MTBP mutants to MTBP-WT did not rescue replication in human cells , ( 3 ) restoring Treslin/TICRR levels in the absence of MTBP binding was insufficient to restore replication in Treslin siRNA ( siTreslin ) –treated cells [28] and in MTBP-immunodepleted Xenopus egg extracts [29] , and ( 4 ) our partial RNAi-depletion experiments with Treslin/TICRR showed that strong reduction of replication—as observed in MTBP-5m and m1–6 cell lines—is only achieved by depleting Treslin/TICRR to levels well below the levels in these MTBP mutant lines ( S8A , S8E and S8F Fig and S1 Data ) . The S7M-C domain of MTBP—albeit not essential for replication—is required for normal replication levels in HeLa cells . Its function is distinct from Treslin/TICRR binding because MTBP-ΔS7M-C bound Treslin/TICRR as well as MTBP-WT . In yeast , the S7M-C domain mediated Sld7 dimerisation in the crystal structure to form dimers of Sld3-Sld7 heterodimers [18] . How important Sld7 dimerisation is for replication in yeasts was not addressed . We found that attaching an artificial dimerisation domain in the form of GST to S7M-C deletion mutants of MTBP rescues the ability of these mutants to induce replication in Hela cells . Sld3/Treslin-Sld7/MTBP dimerisation could help ensure that both MCM helicases in pre-RCs become activated , but never only one , which could generate initiation intermediates that jeopardise genetic stability . We cannot formally exclude that GST rescues replication in cells expressing S7M-C MTBP deletion mutants by virtues other than homodimerisation that it does not share with the GFP control tag . We found no biochemical evidence of S7M-C–dependent MTBP dimerisation by binding studies using pulldowns with MTBP fragments expressed in 293T cells . This indicates that S7M-C–mediated MTBP dimerisation may be weak or spatiotemporally regulated , e . g . , by posttranslational modification . Three lines of evidence support the conclusion that Cdk8/19-cyclin C has a role in genome replication that requires its binding to the metazoa-specific middle domain of MTBP . ( 1 ) Cdk8/19-cyclin C forms a protein complex with the MTBP-Treslin/TICRR-TopBP1 replication initiation regulator . ( 2 ) Cells expressing deletion or point mutants of MTBP that are deficient in binding to Cdk8/19-cyclin C show compromised replication . ( 3 ) RNAi depletion of cyclin C also compromises replication . We believe that the function of Cdk8/19-cyclin C in replication is distinct from its classic role as the kinase module of the mediator of transcription because Cdk8/19-cyclin C forms distinct complexes with MTBP-Treslin/TICRR-TopBP1 , or the kinase module of the mediator subunits Med12 and Med13 . Moreover , in contrast to RNAi depletion of cyclin C knock-down of Med12 , which shares many mediator-related functions with Cdk8/19-cyclin C [52 , 53] , did not lead to replication defects ( S13G and S13H Fig and S1 Data ) . The exact role in the replication of Cdk8/19-cyclin C in complex with MTBP-Treslin/TICRR-TopBP1 is unclear . Cdk8/19-cyclin C nonbinding MTBP mutants replicate slightly slower than MTBP-WT cells . Apparently , the replication problems in the mutant cell lines are severe enough to cause unreplicated gaps in the genome that become evident as fragile metaphase chromosomes and aberrant mitotic figures . Compromised replication was previously shown to induce a checkpoint that delays cytokinetic abscission [36] . This checkpoint is thought to provide cells in which abscission is not possible normally , with an opportunity to nevertheless achieve division and avoid tetraploidy [54 , 55] . Consistently , we find delayed abscission in our mutant MTBP cell lines . Due to these phenotypic analyses , we believe that insufficient replication forks are generated in cells with Cdk8/19-cyclin C binding-deficient MTBP , resulting in replication gaps . One model is that Cdk8/19-cyclin C binding to MTBP is generally required for maximal activity of MTBP to promote origin firing . Alternatively , a specific subset of firing events could be controlled by Cdk8/19-cyclin C . This could be a positive control , so that inefficient firing of CDK8-activated origins causes replication gaps , or a negative control , so that ectopic firing of CDK8-inhibited origins causes replication gaps indirectly , for example , by promoting replication-transcription collisions or by depleting certain genomic regions of dormant origins . We cannot exclude that Cdk8/19-cyclin C binding to MTBP has roles in transcription or other processes that contribute to the observed phenotype .
All cell lines were maintained in DMEM ( Life Technologies ) supplemented with 10% FBS ( Life Technologies ) and 1% penicillin/streptomycin antibiotics ( Life Technologies ) in 5% CO2 conditions at 37°C . Stable MTBP transgene–expressing HeLa Flip-In T-Rex cell lines [56] were generated according to the manufacturer’s instructions ( Invitrogen ) using pcDNA5-FRT-TO-Flag3-Tev2-GFP or pcDNA5-FRT-TO-Flag3-Tev2 . Cell pools containing at least 50 individual clones were used to average out clonal variation . Stable MTBP transgene-expressing HeLa Kyoto cell lines were generated using pIRESpuro3-Flag-AcGFP . For cell synchronisations , HeLa-Flp-In T-Rex or U2OS cells were treated with 2 mM thymidine for 24 h . For S phase populations , double thymidine-arrested cells were harvested in thymidine or released for 3 h before harvesting . G2/M phase populations were released from thymidine arrest for 10 h and for G1 phase populations for 14 h . For release , cells were washed twice in PBS before incubation in normal medium . For mitotic blocks , unsynchronised cells or cells released from thymidine were treated with 0 . 1 μg/mL nocodazole overnight before harvesting by mitotic shake-off . For nocodazole release into G1 phase , mitotically arrested cells were washed twice in PBS before incubation in medium lacking nocodazole . Only attached cells were harvested by trypsinisation for G1-synchronised populations . Cells ( 293T ) were transfected using standard calcium phosphate transfection . In brief , an approximately 60%–70% confluent 10-cm dish of 293T cells was transfected with 13 μg plasmid DNA using 83 μL 2 M CaCl2 , 566 μL sterile H2O , and 667 μL 2×HBS ( 50 mM HEPES , 10 mM KCl , 2 mM dextrose , 280 mM NaCl , 1 . 5 mM Na2HPO4 ) . The medium was changed 24 h after transfection , and cells were cultivated for another 48 h before harvesting . HeLa Flip-In T-Rex cell lines carrying transgenes or control cells were treated with 1 μg/mL doxycycline roughly 2 h prior to transfection to induce expression . siRNA transfections were carried out using RNAiMAX ( Life Technologies ) according to the manufacturer’s guidelines . In brief , 1 . 5 × 105 cell per 6-cm tissue culture dish were transfected with 20 nM control siRNA ( GL2 ) or siRNA against MTBP ( 1:1 mix of MTBP1 and MTBP2 ) using 10 μL RNAiMAX in 5 mL volume . For MTBP and Treslin/TICRR immunoblots following RNAi , whole cells were boiled in 3× Laemmli sample buffer , separated by SDS-PAGE , and blotted on nitrocellulose membranes before detection with anti-MTBP and anti-Treslin/TICRR antibodies . Seventy-two hours after RNAi transfection , cells were pulse labelled using 10 μM BrdU for 30min . Cells were then harvested and fixed with −20°C cold methanol for 24 h . The fixed cells were treated with 2 M HCl/0 . 5% Triton X-100 for 30 min before washing extensively with PBS/0 . 5% Triton . Incorporated BrdU was detected using FITC-coupled mouse anti-BrdU in 1% BSA/PBS/0 . 01% Triton X-100 for 1 h according to the manufacturer’s instructions . After washing once with PBS/0 . 01% Triton X-100 , the DNA was stained with 25 μg/mL PI in the presence of 100 μg/mL RNaseA . Flow cytometry analysis was performed using a FACSCalibur flow cytometer ( BD Biosciences ) with a logarithmic setting of the FL1-H channel for FITC detection and a linear setting of the FL2-H channel for PI detection . Data analysis was performed using the Kaluza Analysis 1 . 3 software ( Beckman Coulter ) . Cell aggregates were excluded based on their high FL2-W signal . BrdU-PI profiles were generated as density plots . Reduction of BrdU incorporation upon siRNA treatment was analysed by gating S phase cells in BrdU/PI dot plots and visualised as BrdU histogram overlays . For quantifications of the efficiency of replication rescue by MTBP transgenes , the BrdU signal of BrdU-positive S phase cells was background-subtracted using the signal intensity of BrdU-negative G1 and G2/M phase cells to calculate the replication-dependent BrdU signal . This value was divided by the BrdU replication signal of control siRNA–treated cells . Visualisation as bar diagrams and statistics was done using GraphPad Prism 5 ( GraphPad Software ) . To assess the frequency of fragile chromosomes , cells were treated for 2 h with 0 . 08 μg/mL colcemide . After mitotic shake-off , cells were treated with 1% sodium citrate for 10 min and fixed by washing three times with methanol/acetic acid ( 3:1 ) . Droplets of cells in fixation solution were dropped on glass slides from a height of about 150 cm . After drying at RT for 24 h , cells were Giemsa stained using 1:20 Giemsa ( Merck 109204 ) in 4 mM NaHPO4 , 5 mM KHPO4 . Fragile chromosomes were imaged using conventional light microscopy ( 40× oil lens ) using the Zeiss ZEN 2 . 3 software . The fractions of fragile chromosomes were assessed manually . For quantification , around 8 , 000 chromosomes per sample were analysed . HeLa Flip-In T-Rex cell lines stably expressing AcGFP-PCNA were seeded 24 h after siRNA treatment into 4-well glass bottom plates ( ibidi 80427 ) . Imaging was done using an Andor/Nikon spinning disk confocal microscope at 37°C and 5% CO2 in imaging medium ( DMEM , Life Technologies ) supplemented with 10% FBS , 1% penicillin/streptomycin antibiotics ( Life Technologies ) . Excitation was at 475 nm laser wavelength . A frame rate of one image per 30 min was used for each 48-h imaging session . Three Z sections per frame were taken . For image processing , maximal intensity projections were made , and Fiji software was used to generate movies . Movies were analysed visually to assess replication patterns . To analyse the temporal replication program , S phase patterns were classified into early , mid– , and late–S phase patterns and quantified . Early S phases were from the first appearance of replication foci to the first frame , with mid–S phase patterns defined as nuclear rim and/or perinucleolar replication . Late patterns were all frames with distinctive large bright replication foci . Quantifications represent the fraction of early , mid , and late patterns of the total number of S phase frames . For detailed analysis of mid/late replication patterns , three classes of patterns were defined: ( A ) fragmented or complete nuclear rim staining , ( B ) small intranuclear foci , and ( C ) bright large foci in the nuclear interior or at the rim . Frames showing a single pattern or combinations of patterns were quantified and normalised to the total frame number . Cells on glass coverslips were fixed with 4% paraformaldehyde for 10 min at RT . After permeabilisation and blocking in PBS , 0 . 1% BSA , 0 . 1% Triton X-100 for 1 h , anti-tubulin and anti-mouse Alexa 488 antibodies were used at dilutions of 1:2 , 000 and 1:100 , respectively , in blocking buffer ( PBS , 0 . 1% BSA ) in distinct 1-h incubation steps to stain tubulin . A total of 1 μg/mL dsDNA stain ( Hoechst 33258; Sigma , B2883 ) in blocking buffer was used to stain DNA . After mounting in FluorSave Reagent ( Calbiochem , 345789 ) , cells were imaged using a Zeiss Axio Observer7 microscope . Eleven Z sections were taken and analysed visually using the Fiji software . Anaphases were classified as ‘aberrant’ if they showed a DNA signal between the separating DNA masses . For quantification , aberrant anaphases were normalised to the total number of anaphases . Cells with cytokinesis delays showed a thin connection containing a spindle midbody between two daughter cells that had already formed clear interphase nuclei with decondensed chromatin . Micronuclei were defined as DNA masses in interphase cells that were clearly distinct from the main nucleus . For quantification of cytokinesis delays and micronuclei , their number was normalised to the total number of cells . Hela Flip-In cells were grown on glass coverslips . Transient transfection was performed using Lipofectamine 2000 ( Life Technologies ) according to the manufacturer’s instructions . Cells were fixed 48 h after transfection using 4% paraformaldehyde for 15 min at RT . After three washing steps with PBS , cells were permeabilised and blocked with PBS , 0 . 1% BSA , 0 . 1% Triton X-100 for 1 h at RT . Subsequently , cells were incubated for 1 h at RT with the primary antibodies ( anti-FLAG M2 1:1 , 000 or anti-GFP 1:250 ) diluted in blocking buffer . After three washing steps using PBS with 0 . 1% BSA , cells were incubated with anti-mouse Alexa 488 ( 1:250 ) and 1 μg/mL dsDNA stain ( Hoechst 33258; Sigma , B2883 ) for 1 h at RT . After three washing steps and mounting using Roti-Mount FluorCare ( Carl Roth , HP19 . 1 ) , image acquisition was performed using the Zeiss Axio Observer 7 . For generation of the monoclonal antibody against human Treslin/TICRR ( clone 30E7: N-His-Treslin [amino acids 260–791] ) , MTBP ( clone 12H7: N-His-MTBP [amino acids M1–I284] ) , and clone 4H9 ( N-His-MTBP [amino acids F102-Y513] ) , approximately 50 μg of His-tagged fusion protein dissolved in PBS was emulsified in an equal volume of incomplete Freund’s adjuvant , and Wistar rats ( MTBP antigens ) or C57BL/6J mice ( Treslin antigen ) were immunised subcutaneously ( s . c . ) and intraperitoneally ( i . p . ) . Six weeks after immunisation , a 50-μg boost injection was applied i . p . and s . c . 3 d before fusion . Fusion of the splenic B cells and the myeloma cell line P3X63Ag8 . 653 was performed using polyethylene glycol 1500 according to standard protocols [57] . Hybridoma supernatants were tested by solid-phase ELISA using the His-fusion proteins and verified by western blotting . Monoclonal hybridoma cell lines of the Treslin/TICRR- and MTBP-reactive supernatants were cloned twice by limiting dilution . The IgG subclasses were determined with ELISA assay as MTBP clones 12H7 and 4H9: rat IgG2b and Treslin/TICRR clone 30E7 . The following nomenclature was used for Treslin/TICRR domain mutants: ΔCIT , the N-terminal 264 amino acids of Treslin/TICRR were deleted; ΔM1 , amino acids 265–408 were deleted; ΔM2 , amino acids 409–593 were deleted; ΔSTD , amino acids 717–792 were deleted; 2PM , the two essential CDK sites ( T969A; S1001A ) were mutated to alanine; ΔC-terminal , The C-terminus was deleted from amino acid 1057 . Overexpressed Flag-purified MTBP-3Flag was loaded onto a Superdex 200 column ( 2 . 4 mL 3 . 2/300 ) . FPLC was performed at +4°C at a 30 μL/min flow rate with a 100-μL fraction size in running buffer ( 20 mM HEPES , pH 8 . 0 , 200 mM NaCl , 0 . 5 mM TCEP , 10 mM NaF , 0 . 1% TritonX-100 ) . A total of 1/15 of each fraction was analyzed via western blot . Gel filtration standard ( 1511901 , BioRad ) was used to determine apparent molecular weights . Chromatin-enriched fractions were purified from Hela Flp-In cells via lysis in 10 mM HEPES , pH 7 . 0 , 100 mM NaCl , 300 mM sucrose , 3 mM MgCl2 , 0 . 5% Triton , 5mM ß-mercaptoethanol , complete EDTA-free protease inhibitor cocktail , and 5 μg/mL cytochalasin D . Chromatin was harvested by centrifugation at 3 , 200 rpm for 4 min at 4°C in table-top centrifuge and washing the chromatin pellet in lysis buffer three times , involving 5 min of incubation in washing buffer . | Efficient and well-regulated DNA replication origin firing is central to ensure complete and accurate genome duplication before cell division . We here use bioinformatics and cultured human cells to understand the role of the essential origin firing factor Mdm2 binding protein ( MTBP ) . We prove that MTBP is orthologous to yeast Sld7 . Their homologous N-terminal domains bind the orthologous firing factors Treslin/TICRR and Sld3 , respectively . The homologous C termini may constitute MTBP/Sld7 homodimerisation domains . Because Sld7 was the only yeast core origin firing factor for which no metazoan orthologue had been found , our work proves that the complement of core firing factors is conserved from yeast to man . We also find that MTBP contains a central region that is not found in yeast , and that is important for replication in human cells , showing that—despite fundamental conservation—higher eukaryotes have also evolved specific origin firing processes . The Cdk8/19-cyclin C transcription kinase that has not previously been implicated in DNA replication binds the central MTBP region , and this is required for complete DNA replication and proper chromosome segregation in human cells . Our work suggests that MTBP helps integrate universal conserved eukaryotic origin firing processes into the complex metazoan cells . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"cell",
"binding",
"cell",
"physiology",
"molecular",
"probe",
"techniques",
"gene",
"regulation",
"cell",
"cycle",
"and",
"cell",
"division",
"cell",
"processes",
"immunoblotting",
"green",
"fluorescent",
"protein",
"dna-binding",
"proteins",
"luminescent",
"proteins",
"dna",
"replication",
"immunoprecipitation",
"dna",
"molecular",
"biology",
"techniques",
"research",
"and",
"analysis",
"methods",
"small",
"interfering",
"rnas",
"proteins",
"gene",
"expression",
"molecular",
"biology",
"precipitation",
"techniques",
"biochemistry",
"rna",
"cell",
"biology",
"nucleic",
"acids",
"genetics",
"biology",
"and",
"life",
"sciences",
"non-coding",
"rna"
] | 2019 | The Cdk8/19-cyclin C transcription regulator functions in genome replication through metazoan Sld7 |
Ergosterol is an important constituent of fungal membranes . Azoles inhibit ergosterol biosynthesis , although the cellular basis for their antifungal activity is not understood . We used multiple approaches to demonstrate a critical requirement for ergosterol in vacuolar H+-ATPase function , which is known to be essential for fungal virulence . Ergosterol biosynthesis mutants of S . cerevisiae failed to acidify the vacuole and exhibited multiple vma− phenotypes . Extraction of ergosterol from vacuolar membranes also inactivated V-ATPase without disrupting membrane association of its subdomains . In both S . cerevisiae and the fungal pathogen C . albicans , fluconazole impaired vacuolar acidification , whereas concomitant ergosterol feeding restored V-ATPase function and cell growth . Furthermore , fluconazole exacerbated cytosolic Ca2+ and H+ surges triggered by the antimicrobial agent amiodarone , and impaired Ca2+ sequestration in purified vacuolar vesicles . These findings provide a mechanistic basis for the synergy between azoles and amiodarone observed in vitro . Moreover , we show the clinical potential of this synergy in treatment of systemic fungal infections using a murine model of Candidiasis . In summary , we demonstrate a new regulatory component in fungal V-ATPase function , a novel role for ergosterol in vacuolar ion homeostasis , a plausible cellular mechanism for azole toxicity in fungi , and preliminary in vivo evidence for synergism between two antifungal agents . New insights into the cellular basis of azole toxicity in fungi may broaden therapeutic regimens for patient populations afflicted with systemic fungal infections .
Pathogenic fungal species , including Aspergillus , Candida , Histoplasma and Cryptococcus among others , cause infections ranging from mucocutaneous disorders to life-threatening invasive diseases that can involve any organ . In the past two decades , expanding populations of immunocompromised patients and increased use of invasive devices and implants have led to an increase in the incidence of fungal infections [1] , [2] . Currently , four major categories of antifungal therapeutics are available to treat invasive fungal infections: polyenes , azoles , echinocandins and flucytosine [3] . Azole drugs are the most widely deployed in clinics , and inhibit the biosynthesis of ergosterol , the fungal-specific sterol . The primary molecular target of azole drugs is Erg11p ( Entrez GeneID: 856398 ) , a P450 cytochrome that catalyzes 14α-demethylation of lanosterol in the ergosterol biosynthesis pathway [4] . Besides azoles , a number of other drugs such as allylamines and morpholines used in medicine and agriculture also inhibit ergosterol biosynthesis [5] , [6] . Ergosterol is an important constituent of membrane lipids , similar to vertebrate cholesterol , and modulates the fluidity , permeability and thickness of the membrane . These sterols preferentially associate with sphingolipids in microdomains that have been postulated to have important roles in membrane organization and function [7] , [8] . Ergosterol is most abundant in the plasma membrane and has been implicated in several cellular processes including sporulation , pheromone signaling and plasma membrane fusion during mating and endocytosis [9] , [10] . Discernable amounts of ergosterol have also been found in membranes of intracellular organelles including peroxisomes , mitochondria , vacuoles and ER [11] . Some studies have ascribed a regulatory role at these intracellular compartments , including homotypic vacuole fusion [12] , mitochondrial biogenesis and inheritance , and protein sorting along exocytosis and endocytosis pathways [13] , [14] . The absence of ergosterol in mammals and suppression of fungal proliferation by a battery of ergosterol biosynthesis inhibitors emphasize the importance and utility of ergosterol as an effective target in antifungal chemotherapy . Yet , despite nearly two decades of use and the general recognition of the importance of ergosterol to fungal cells our understanding of the specific cellular processes disrupted by ergosterol deprivation following azole therapy remains minimal . The limited categories of antifungal agents and emergence of resistance to existing antimycotics have prompted a search for compounds with alternative modes of action . The anti-arrhythmia drug , amiodarone , was recently documented to exhibit fungicidal activity [15] , [16] . This cationic amphipathic compound inserts into the lipid bilayer where it elicits membrane hyperpolarization , and influx of H+ and Ca2+ into the cytoplasm [15] , [17] . Within minutes , amiodarone also elicits a transcriptional response to starvation and blocks cell cycle progression [18] . A screen of the yeast haploid deletion library for amiodarone hypersensitivity revealed multiple vma genes encoding subunits of the vacuolar membrane H+-ATPase [15] . The V-ATPase is critical for generation of a pH gradient that drives secondary transporters to maintain cellular ion homeostasis . Since the fungicidal activity of amiodarone appears to be tightly coupled to ion stress [19] , hypersensitivity of vma mutants was ascribed to defects in ion homeostasis . Notably , deletion mutants of several erg genes in the ergosterol biosynthesis pathway were also identified in the screen [15] , although the underlying mechanism for their amiodarone hypersensitive phenotype was unclear . Meanwhile , a screen of the yeast haploid deletion mutant library for strains with alterations in vacuolar pH revealed that erg mutants , like vma mutants , had severely alkaline vacuoles ( Brett , C . L . , Rao . R . et al . , unpublished data ) . A separate study showed that sphingolipid , the other major membrane lipid component found associated with ergosterol in detergent-resistant microdomains , was required for the structural integrity of V-ATPase domains [20] . In light of these observations , we investigated a potential link between ergosterol and V-ATPase function , which led to a mechanistic basis for the antifungal activity of azole drugs . As a first step in exploiting our observations for improved management of invasive fungal diseases , we assessed the efficacy of combining fluconazole with ion homeostasis-disruptive agent amiodarone in a murine Candidiasis model .
A genome-wide screen of the S . cerevisiae haploid deletion collection for hypersensitivity to the antifungal agent amiodarone revealed multiple vma and erg mutants [15] . Given our previous observation that amiodarone triggered Ca2+ and H+ influx leading to fungal death from ion stress [15] , [19] and the importance of V-ATPase in ion homeostasis , we considered the possibility that ergosterol may be important for V-ATPase function . A systematic examination of viable erg null mutants ( erg2Δ [Entrez GeneID: 855242] , erg3Δ [Entrez GeneID: 850745] , erg6Δ [Entrez GeneID: 855003] and erg24Δ [Entrez GeneID: 855441] ) revealed multiple vma− phenotypes , with erg24Δ displaying the most severe defects . In addition to hypersensitivity to amiodarone ( Fig . 1A ) , erg24Δ was unable to grow at alkaline pH ( Fig . 1B ) , a defining phenotype of vma mutants indicative of the inability to acidify vacuoles . Furthermore , erg24Δ exhibited hypersensitivity to Zn2+ toxicity and to the calcineurin inhibitor FK506 , consistent with broad ion homeostasis defects characteristic of vma mutants ( Fig . 1C–D ) . Yeast strains defective in trafficking of chitin synthase , including vma mutants , are more sensitive to toxicity from calcofluor white , an antimicrobial agent that binds to cell wall chitin . We showed that erg24Δ shared calcofluor white hypersensitivity with vma2Δ ( Fig . 1E ) . Poor growth of vma mutants on high concentrations of non-fermentable carbon sources has been ascribed to oxidative stress from respiration . Although erg24Δ was able to grow on non-fermentable carbon sources , growth was significantly impaired ( Fig . 1F ) . Overall , the novel observation that ergosterol biogenesis mutants largely phenocopy vma mutants suggests that cellular ergosterol content may be important for the function of V-ATPase . V-ATPase hydrolyzes ATP and acidifies vacuolar compartments . To assess a possible requirement for ergosterol in V-ATPase function , we first measured the vacuolar pH in erg mutants using the pH-sensitive fluorescent dye BCECF . The acetoxy methyl ester of BCECF is taken up by cells and de-esterified in the vacuole where it accumulates [21] . While the vacuolar pH of wild-type cells was 6 . 0 , vacuoles of vma2Δ ( Entrez GeneID: 852424 ) cells were significantly more alkaline , around pH 7 , as would be expected for loss of proton pump capacity ( Fig . 2A ) . Vacuolar pH of all viable erg mutants closely resembled that of the vma mutant , as shown for erg24Δ ( Fig . 2A ) . Next , we purified intact vacuolar vesicles from wild type , erg24Δ and vma2Δ strains , and compared V-ATPase function , including rates of proton pumping and ATP hydrolysis . There was no V-ATPase activity detectable in vma2Δ vacuoles as expected , whereas in vitro ATPase and H+ pumping activity were both diminished to about 40% of wild type levels in erg24Δ ( Fig . 2B ) . Examination of sterol profiles in purified vacuoles confirmed the presence of ergosterol in wild type vacuoles and its absence in erg24Δ ( not shown ) . Taken together , these results provide evidence for a role for ergosterol in V-ATPase activity . The P-type H+-ATPase Pma1 ( Entrez GeneID: 852876 ) has been documented to associate with ergosterol enriched domains [22] , [23] . Upon glucose activation , it pumps protons out of cells to acidify the extracellular medium . To assess the effect of ergosterol depletion on Pma1 function , we examined extracellular acidification upon glucose activation in erg24Δ cells . As shown in Fig . 2C , the kinetics of medium acidification was substantially similar between the wild type and erg24Δ , while a previously characterized PMA1 mutant , pma1-105 , reported to have a 65% reduction in activity , exhibited slower acidification rate as expected [24] . These data suggest that ergosterol is not required for Pma1 function and support the specificity of the ergosterol effect on V-ATPase . To investigate the mechanism underlying the requirement of ergosterol for optimal V-ATPase function , we first asked if V-ATPase localization was altered in the erg24Δ mutant . The V-ATPase is made up of 14 subunits organized into the Vo sector , integral to the membrane , and the cytoplasmic V1 sector that reversibly dissociates from the membrane [25] . Figure 3A & 3B show that representative subunits from the Vo sector ( Vph1-GFP ) and the V1 sector ( Vma5-GFP ) colocalized with the vacuolar membrane stain ( FM4-64 ) in erg24Δ cells , similar to the isogenic wild type . It was possible that the ergosterol biogenesis defect significantly decreased V-ATPase expression or caused the dissociation of V1 from Vo domain . However , analyses of representative V-ATPase subunits in vacuolar vesicles purified from the wild type and erg24Δ showed similar expression levels and V1/Vo ratios that were identical between the two strains ( Fig . 3C & 3D ) . The oligosaccharide Methyl-β-Cyclodextrin ( MβCD ) extracts sterols from cellular membranes . We observed a dose dependent loss of V-ATPase activity following treatment of purified vacuolar vesicles with MβCD , which could be blocked by preloading cholesterol into MβCD ( Fig . 4A ) . Both ATP hydrolysis rates and H+ pumping declined at similar rates , suggesting an inhibition of the intact V1Vo complex . This was verified by immunoblot analysis of vacuolar membranes collected by centrifugation after MβCD treatment ( Fig . 4B ) : we did not observe a loss of either V1 ( Vma2p ) or Vo ( Vph1p ) subunits , suggesting that the two sectors remain associated after ergosterol extraction . In contrast , a previous study pointed to a role for sphingolipids in maintaining structural integrity of the V-ATPase enzyme complex [20] . We conclude that ergosterol constitutes a critical component in the lipid membrane environment for V-ATPase function . Azole drugs exert their fungistatic effect by inhibiting ergosterol biosynthesis , specifically targeting lanosterol demethylase ( Erg11p ) , which is the enzymatic step immediately upstream from Erg24p . Despite the wide spread use of azole antifungals , our understanding of the specific cellular pathways disrupted by azoles is limited . In addition to ergosterol depletion , fluconazole treatment results in accumulation of lanosterol ( substrate of Erg11p ) and its derivative 14-methyl-3 , 6 diol [26] . Based on analysis of sterol profiles in fluconazole susceptible and resistant strains , it was concluded that 14-methyl-3 , 6 diol toxicity [26] , [27] was responsible for azole-mediated growth arrest . To specifically assess the functional effect of ergosterol depletion following azole treatment , we used the S . cerevisiae strain WPY361 with a gain-of-function mutation in UPC2 ( Entrez GeneID: 851799 ) , upc2-1 , that allows overexpression of ATP-binding cassette transporters required for uptake of exogenously added sterol under aerobic condition [28] , [29] . Table 1 shows that growth inhibition caused by fluconazole treatment in WPY361 can be reversed by exogenous supply of ergosterol . To examine whether ergosterol feeding represses endogenous sterol metabolism and reduces accumulation of intermediates and derivatives , we analyzed sterol profiles in upc2-1 cells after six hours of exposure to fluconazole and exogenous ergosterol . As expected , fluconazole alone caused reduction of ergosterol content ( nine-fold ) and accumulation of lanosterol and a major derivative , likely to be 14-methyl 3 , 6-diol ( Fig . S1 ) . Compared with cells treated with fluconazole alone , cells treated with fluconazole and ergosterol had three-fold higher ergosterol content , yet the level of lanosterol and its derivatives remained the same ( Fig . S1 ) . These data indicate that ergosterol feeding did not reduce accumulation of intermediates and derivatives . Thus , depletion of ergosterol , rather than the toxicity of intermediates and derivatives , is a plausible mechanism for the antifungal activity of fluconazole . Based on our evidence that ergosterol was required for optimal V-ATPase function , we predicted that azole treatment would similarly impair activity of the V-ATPase . Indeed , we show that fluconazole treatment resulted in a dose-dependent alkalinization of vacuolar pH , consistent with depletion of ergosterol from the vacuolar membrane ( Fig . 5A ) . Furthermore , vacuolar membrane vesicles purified from fluconazole treated cells showed significant reductions in both H+ pumping rates and ATPase activity ( Fig . 5B ) . Next , we evaluated the effect of ergosterol feeding on vacuolar pH in the upc2-1 mutant . Not only did exogenous ergosterol restore growth in fluconazole treated cells , vacuolar acidification closely resembled that of untreated cells ( Fig . 5C & 5D ) . This correlation strengthens the hypothesis that V-ATPase inhibition contributes to the cellular mechanism of azole activity . To extend these observations in the human pathogen Candida albicans , we monitored vacuolar uptake of the fluorescent weak base quinacrine . Previous studies have demonstrated pH-dependent vacuolar accumulation of quinacrine , which was abolished in the homozygous vma7−/− mutant [30] . We observed robust quinacrine fluorescence in C . albicans vacuoles , colocalizing with FM4-64 staining of vacuolar membranes . Fluconazole treatment drastically reduced vacuolar accumulation of quinacrine in most cells , indicative of impaired vacuolar acidification ( Fig . 5E ) . Additionally , trafficking of FM4-64 to the vacuolar membrane was impaired , consistent with endocytosis defects seen in vma mutants [30] . We note that following 6 h of fluconazole treatment , cells failed to divide but continue to increase in size , as previously reported [9] . These data suggest that the requirement of ergosterol for V-ATPase function is conserved in fungi . Given the importance of V-ATPase function and vacuolar acidification in diverse cellular processes , we conclude that disruption of V-ATPase function plays a critical role in antifungal activity of azole drugs . Consistent with this conclusion , both vma7−/− and erg24−/−mutants of C . albicans exhibit defective virulence in murine models of Candidiasis [30] , [31] . We showed previously that amiodarone triggered a cytosolic H+ and Ca2+ surge in the baker's yeast and that mutants defective in ion homeostasis were hypersensitive to amiodarone toxicity [15] . Given the pivotal role of the V-ATPase in maintaining intracellular cation homeostasis , we predicted that ergosterol depletion by azole treatment would impair V-ATPase function and exacerbate disruption of cation homeostasis by amiodarone . We first tested this hypothesis in the S . cerevisiae model . We used pH sensitive GFP ( pHluorin ) to monitor changes in cytosolic pH in wild type , vma2Δ and erg24Δ strains upon exposure to amiodarone [32] . Cytosolic acidification was most pronounced in the vma2Δ mutant consistent with a loss in the ability to transport H+ from the cytosol to the vacuole . Depletion of ergosterol in erg24Δ mutant or by fluconazole treatment also exacerbated cytosolic acidification , relative to wild type , upon amiodarone addition ( Fig . 6A ) . We have previously demonstrated defective clearance of cytosolic Ca2+ in vma mutants following exposure to amiodarone [15] . We now show that pretreatment with fluconazole exacerbates the cytosolic Ca2+ surge elicited by amiodarone , consistent with defective V-ATPase function ( Fig . 6B ) . The proton gradient established by the V-ATPase drives vacuolar sequestration of excessive cytosolic Ca2+ by H+/Ca2+ exchange mechanisms . The importance of the V-ATPase is demonstrated by the inability of purified vacuolar vesicles to sequester 45Ca2+ following treatment with the V-ATPase inhibitor concanamycin A ( Figure 6C ) . As predicted by our hypothesis , vacuolar vesicles purified from erg24Δ or fluconazole-treated wild-type S . cerevisiae both showed similar impairment in their ability to sequester Ca2+ ( Fig . 6C ) . Likewise , vacuolar vesicles purified from C . albicans cells treated with fluconazole were also impaired in sequestering Ca2+ ( Fig . 6D ) . Thus , our observations provide a mechanistic basis for previous reports of synergism between azoles and amiodarone against pathogenic fungi in vitro [15] . To investigate the potential clinical application of this synergism , we studied the effect of combining fluconazole and amiodarone in a murine candidiasis model . The microbial burden of Candida albicans in kidneys was assessed 3 days following intravenous infection of Balb/C mice . AMD treatment doses were presented at 5 . 0 and 25 mg/kg , while FLC doses were 0 . 5 and 1 mg/kg with treatments given once daily for three days after the first dose . In the absence of amiodarone , there was a significant dose-dependent effect of FLC on C . albicans ( ratio of geometric means of the cfu count per 1 mg/kg of FLC was 0 . 5914 , p = 0 . 01 ) . In the absence of FLC , amiodarone did not confer a significant antifungal activity at the concentrations tested ( ratio of geometric means of the cfu count per 1 mg/kg of amiodarone was 0 . 9996 , p = 0 . 95 ) . Yet in combination with fluconazole , the two doses of amiodarone significantly reduced C . albicans infection above and beyond the dose-dependent effect of fluconazole ( ratio of geometric means of CFU with amiodarone versus without , for the same dose FLC = 0 . 4969 , p<0 . 001 ) ( Fig . 6E ) . These data provide proof of principle that combining azole drugs with other antifungal compounds that disrupt intracellular cation homeostasis could be a promising therapeutic strategy to treat systemic fungal infections .
In fungal cells , V-ATPase acidifies intracellular compartments including the vacuole , endosomes , and late-Golgi . Mutants lacking V-ATPase exhibit characteristic phenotypes of growth sensitivity to alkaline pH , calcofluor white , Ca2+ and metal ion stress , and are unable to grow on high concentrations of non-fermentable carbon sources [25] . We show that mutants defective in ergosterol biosynthesis exhibit most of these characteristic vma− phenotypes . Furthermore , we showed a reduction of vacuolar acidification by ergosterol depletion , restoration of vacuolar acidity by ergosterol feeding , and used biochemical assays of H+ pumping with purified vacuolar vesicles to collectively demonstrate the requirement of ergosterol for optimal V-ATPase function . This functional link explains simultaneous identification of multiple ergosterol biosynthesis genes ( erg ) and V-ATPase subunit genes ( vma ) in a number of genome-wide screens , including sensitivity to low Ca2+ , alkaline stress , and acid stress [33]–[35] . V-ATPase mutants and ergosterol biosynthesis mutants also share defects in endocytosis [9] , [36] . Additionally , in pathogenic fungal species , both vma and erg mutants are avirulent [30] , [31] , [37] . Disruption of V-ATPase function by ergosterol deprivation provides a mechanistic basis for these similarities . To investigate the underlying basis for the requirement of ergosterol in V-ATPase function , we first checked the localization and abundance of V-ATPase in erg24Δ . Fluorescent microscopy and immunoblot analysis ruled out possible mislocalization and reduced abundance of V-ATPase in erg24Δ vacuolar membrane . In yeast , the V-ATPase complex undergoes rapid reversible dissociation into non-functional V1 and Vo sectors in response to glucose withdrawal [38] . A study with Baby Hamster Kidney cells suggested that increasing ratio of Intact V1/Vo along the endocytic pathway effectively increased acidity along the compartments in the pathway [39] . Analyses of immunoblot results in this study show that V1 and Vo domains are still associated , and V1/Vo ratio remains unchanged upon ergosterol deprivation by treating cells with fluconazole and treating vacuolar vesicles with MβCD . These data rule out dissociation of V1 from Vo domain as the cause of reduced V-ATPase function , and indicate that ergosterol directly modulates the activity of V-ATPase . Sphingolipid , another major component of lipid raft , was thought to affect V-ATPase function by maintaining the structural integrity of V-ATPase because V1 subunits ( Vma1p , Vma2p and Vma5p ) dissociate from Vo domain during Ficoll gradient procedure in the sphingolipid mutant sur4Δ . In contrast , Vma2p remained associated with Vph1p after Ficoll gradient procedure in erg24Δ mutant and after ergosterol extraction by MβCD in the wild type . Thus , these two key membrane lipid components play distinct roles in maintaining V-ATPase function . The precise molecular basis of V-ATPase regulation by ergosterol remains to be determined . In mammalian cells , V-ATPase has been shown to associate with cholesterol-rich microdomains , with loss of vesicular acidification reported upon treatment with β-methylcyclodextrin [40] . A number of studies showed that inhibitors could interact with lipid bilayer and affect V-ATPase function by restricting its structural flexibility [41] , [42] . Mechanisms proposed to explain the regulation of Ca2+-ATPase and Na+ , K+-ATPase by membrane lipids include lowering of free energy of activation and proper packing at protein-protein interfaces [43] , [44] . Altered sterol compositions are known to affect membrane packing and rigidity: fluorescence anisotropy probes have revealed increased membrane fluidity and permeability upon fluconazole treatment [45] , consistent with alterations in activity of membrane-localized pumps and transporters , and a critical role for ion homeostasis mechanisms in drug treated cells . It is also possible that ergosterol may affect V-ATPase function indirectly by modulating regulatory interactions with other proteins . The complex multi-subunit structure of V-ATPase and its intimate association with sterols and spingolipids indicate that sophisticated regulatory mechanisms must be in place to ensure proper assembly , configuration , and communication among these components . Ergosterol depletion may affect other membrane functions besides vacuolar acidification . The plasma membrane H+-ATPase , Pma1 , is a major efflux mechanism for protons , and is also found associated with ergosterol-rich microdomains [22] , [23] . However , we found that Pma1-mediated proton pumping function was not altered in erg24Δ , in contrast to the pronounced effect seen on the V-ATPase . This suggests that membrane proteins have specific lipid requirement for their functions . While the molecular target of azole drugs , Erg11p , has been extensively characterized , not much is known about the cellular basis of fungal growth inhibition . Inhibition of Erg11p by fluconazole results in accumulation of lanosterol and its derivative 14-methyl-3 , 6 diol . In azole resistant erg3 mutants , 14-methyl fecosterol accumulates upon treatment with fluconazole [26] , [27] . This has led to the notion that toxicity of sterol derivatives such as 14-methyl-3 , 6 diol mediates the action of fluconazole [45] , [46] . We exploited the ability of a recently described upc2-1 mutation that allows uptake of ergosterol under aerobic conditions to distinguish between the effects of byproduct accumulation and ergosterol depletion on cell growth . The ability of exogenous ergosterol to reverse growth inhibition by fluconazole supports a plausible alternative hypothesis that antifungal activity of azoles is due to ergosterol depletion . Although we ruled out a corresponding decrease in lanosterol and other derivatives in the ergosterol fed cells , we cannot exclude the possibility that a specific ratio of ergosterol to other sterols may counter potential toxic effects . This possibility could be tested by varying ratios of sterols in upc2-1 feeding experiments; however , 14-methyl-3 , 6 diol is not commercially available and its potential toxicity cannot be directly assessed at this time . Furthermore , our results warrant more careful examination of the role of 14-methyl fecosterol in the potential compensation of ergosterol function in membranes . Some reports suggest a role for ROS in the toxicity of azoles to fungal cells [47] . It is worth noting that cellular ROS level in these studies was measured with the fluorescent dye DCFH-DA ( 2′ , 7′-dichlorofluorescin-diacetate ) which has also been documented to respond to pH alterations [48] . Given the effect of azole drugs on pH homeostasis , the role of ROS in azole-induced growth inhibition may need to be re-evaluated . The far-reaching effect of V-ATPase is illustrated by diverse phenotypes exhibited by V-ATPase mutants . By disrupting the function of V-ATPase through ergosterol deprivation , azole drugs can affect a wide range of cellular processes , including cation homeostasis , protein sorting , processing and degradation . Importantly , V-ATPase function and vacuolar processing of virulence factors are required for pathogenesis [30] , [37] . Although we cannot preclude additional cellular targets of azole toxicity , disruption of V-ATPase function is sufficient to repress growth and attenuate fungal virulence and is likely to be a critical mechanism underlying antifungal activity of azole drugs . Amiodarone exhibits antimicrobial activity against a wide range of fungi and protozoa through disruption of H+ and Ca2+ homeostasis [15] , [16] , [49] . Additionally , in vitro studies showed azoles were synergistic with amiodarone against fungal pathogens , e . g . C . albicans and C . neoformans [15] . Interestingly , amiodarone interacts with fluconazole synergistically against fluconazole-resistant clinical isolates of A . fumigatus and C . albicans [50] , [51] . Moreover , the synergy of amiodarone and posaconazole has been shown on the protozoan T . cruzi both in vitro and in vivo [49] . Uncovering the mechanism underlying this synergism may provide insight guiding the design of more potent antifungal therapy . Data in this study show that ergosterol is required for the optimal function of V-ATPase , a central player in maintaining H+ and Ca2+ homeostasis . Therefore , depletion of ergosterol would be expected to exacerbate the disruption of H+ and Ca2+ homeostasis upon amiodarone treatment . Indeed , upon ergosterol depletion either by erg mutations or by azole treatment , 45Ca uptake by purified vacuolar vesicles was reduced while cytosolic H+ and Ca2+ surges increased upon exposure to amiodarone . Thus , we conclude that disruption of V-ATPase function in maintaining cation homeostasis by azoles contributes to the synergy between azoles and amiodarone . Recently , we demonstrated that a combination of amiodarone and fluconazole in C . albicans resulted in dampening of the transcriptional response to either drug alone [52] . This effect could potentially stunt cellular stress responses occurring downstream of drug toxicity and thereby contribute to synergistic effects of the drugs . These data reveal an additional mechanism contributing to the synergism that is distinct from the ion homeostasis defects presented here . We and others have argued that non-overlapping but complementary insights can be obtained from phenotype versus transcriptional profiling [18] . Thus , genes involved in membrane integrity and ion homeostasis such as VMA and ERG , determine phenotype of growth sensitivity to amiodarone . These genes tend to be constitutively expressed , have a non redundant function and represent pathways upstream of the transcriptional response . On the other hand , genes that are differentially regulated in response to drug appear to play a collective , rather than individual , response to adaptation to stress . Taken together , these mechanisms contribute to a more complete picture of the complex cellular response to drug toxicity . Finally , in this study , we evaluated the clinical potential of combining fluconazole and amiodarone in treating fungal infections in a murine Candidiasis model . The synergy demonstrated in this experiment is proof-of-principle that combining azoles with agents disruptive to cation homeostasis is a promising approach to better manage fungal infections .
S . cerevisiae erg− and vma− mutant strains are from MATα deletion library ( Invitrogen , Carlsbad , CA ) . WPY361 ( MATa upc2-1 ura3-1 his3-11 , -15 leu2-3 , -112 trp1-1 ) was kindly provided by Dr . Will Prinz ( NIDDK ) . Yeast cells were grown in standard SC ( synthetic complete ) medium or YPD ( yeast extract , peptone and dextrose ) medium at 30°C with shaking at 250 rpm unless specified otherwise . Media with non-fermentable carbon source contains 1% Bacto-yeast extract , 2% Bacto-peptone , 3% glycerol ( v/v ) plus 2% ethanol ( v/v ) , or 2% sodium lactate . Antibodies against Vph1p and Vma2p were purchased from Invitrogen ( Carlsbad , CA ) or provided by Dr . Patricia Kane ( Upstate Medical University , New York ) . Plasmid pZR4 . 1 with pHluorin gene under TEF1 promoter was constructed to measure yeast cytosolic pH . Briefly , TEF1 promoter sequence was amplified from BY4742 genomic DNA with primers XbaITEF1 and BamHITEF1 , which incorporate restriction sites of XbaI and BamHI . CYC1 terminator sequence was amplified with primers BamHICYC1 and EcoRICYC1 , which incorporate restriction sites of EcoRI and BamHI . pHluorin gene sequence was amplified from pCB190YpHc plasmid with primers BamHIPhluo and BamHIPhluoR , which incorporate BamHI restriction site . The amplicons were digested with corresponding restriction enzymes and ligated sequentially with the backbone of pYEplac181 , resulting in the gene for pHluorin being flanked by TEF1 promoter and CYC1 terminator . pZR4 . 1 was transformed to yeast strains to monitor cytosolic pH change upon exposure to amiodarone . Wild type ( mating type a ) Vph1-GFP and Vma5-GFP strains in which GFP was fused to the C-terminus of the two proteins were purchased from Invitrogen . The Vph1-GFP and Vma5-GFP fragments were amplified from these strains with primers Vph1L2 + Vph1R1 , and Vma5L2 + vma5R1 , respectively . The amplicons were transformed to erg24Δ to replace the endogenous VPH1 and VMA5 genes by homologous recombination . Primer sequences are available upon request . Ergosterol ( Sigma ) was dissolved in Tween 80∶ethanol ( 1∶1 ) as 5 mM stock . Stocks of ergosterol , fluconazole or their combination were added at the same time to WPY361 cells in YPD . For growth assay , stationary cultures were grown in 96-well plates for 30 hours at 30°C . For vacuolar pH measurement and sterol analysis , log phase cells were treated with ergosterol and fluconazole for 6 hours . Total sterols were extracted from ∼5×107 log-phase cells after washing twice with water and analyzed with an Agilent 6850 gas chromatograph with an HP-1 column and FID as described previously [53] . Retention times for cholesterol , ergosterol and lanosterol were determined using standards . Cholesterol was added to each sample to normalize extraction efficiency . Vacuolar vesicles were prepared as described previously except that 10% Ficoll , instead of 8% Ficoll , was used in the second ultracentrifugation step to facilitate purification of vesicles from erg24Δ and fluconazole treated cells [54] . For 45Ca uptake assay , vacuolar vesicles were incubated in reaction buffer containing 5 µM CaCl2 with tracer quantities of 45CaCl2 . After 5 min of incubation , vacuolar vesicles were filtered onto nitrocellulose membranes . The filters were washed and processed for liquid scintillation counting . Concanamycin A was added to 0 . 5 µM to assess the dependence of 45Ca uptake by vacuolar vesicles . For MβCD treatment , vacuolar vesicles were incubated with MβCD or MβCD preloaded with cholesterol ( cholesterol to MβCD ratio 1∶20 by weight , Sigma , C4951 ) for 30 min at 4°C with gentle shaking . Vesicles were spun down and suspended in buffer C with proteinase inhibitors ( aprotinin 2 µg/ml , leupeptin 1 µg/ml , pepstatin 1 µg/ml , chymostatin 2 µg/ml ) for V-ATPase function assays and SDS-PAGE [20] , [54] . ATP-dependent proton pumping was assayed by monitoring change of ΔA490–540 [36] . The initial H+ pumping rates were calculated from the absorbance change in the first 60 seconds . ATPase activity was assessed by an enzyme-coupled assay monitoring depletion of NADH through oxidation at 340 nm . ATPase activity was calculated based on the absorbance decrease between three and six minutes after initiating the reaction . Concanamycin A was added at a final concentration of 0 . 1 µM to assess the specificity of V-ATPase . Cytosolic pH was measured using pHluorin , a pH-sensitive GFP [32] , under TEF1 promoter in plasmid pZR4 . 1 . Early log phase cells harboring plasmid pZR4 . 1 were grown for 6 hours to mid log phase ( OD ∼1 ) in SC minus leucine medium with or without fluconazole . Cells were collected by centrifugation and suspended in SC to OD 3 in a 96-well plate , with 200 µl culture in each well . Amiodarone ( stock 300 µM in H2O ) was injected to give the specified final concentration . Fluorescence emission at 520 nm was measured in triplicate with excitation at 485 nm and 410 nm in a Fluostar Optima plate reader . Cytosolic pH was calculated based on the ratio of emission at 520 nm excited at 485 nm and 410 nm against a calibration curve covering pH from 4 to 8 established as described previously [32] . Vacuolar pH was measured with BCECF AM , a pH-sensitive fluorophore that accumulates in the yeast vacuole [32] . Strains were grown to mid log phase ( OD ∼1 ) in SC medium . Cells were collected by centrifugation and incubated in SC containing 50 mM BCECF AM for 25 min . The cells were washed twice , suspended in SC to OD 2 and transferred to a 96-well plate . Fluorescence emission at 520 nm was measured in triplicate with excitation at 485 nm and 450 nm in a Fluostar Optima plate reader . Vacuolar pH was calculated based on the ratio of emission at 520 nm at dual excitations of 485 nm and 410 nm against a calibration curve covering pH from 4 to 8 . 5 . Proton efflux was assayed in wild-type and erg24Δ cells as described previously [55] . Briefly , cells were grown to ∼OD 1 in YPD . 1 . 5×108 cells were pelleted by centrifugation and washed twice with distilled water . The washed cell pellet was resuspended in distilled water and placed on ice for 2–3 h . Just prior to use , the cells were centrifuged and resuspended in 3 ml of water . Once a stable pH base line was established , glucose was added to 2% to initiate proton efflux . Log phase cells of C . albicans ( SC5314 ) were diluted to OD 0 . 025 and grown in YPD with or without fluconazole ( 1 µg/ml ) for 5 hours . FM4-64 was added to 5 µM and the cultures were grown for another 30 min . Cells were washed twice in YPD , suspended in YPD with or without fluconazole , and shaken for 20 min . Quinacrine was then added to 100 µg/ml and the cultures were shaken for another 5 min . The cells were collected and washed twice in YPD and examined by fluorescence microscopy immediately . All animal experimentation was conducted following the United States Public Health Service guidelines for housing and care of laboratory animals and performed in accordance with Institutional regulations after pertinent review and approval by the Institutional Animal Care and Use Committee of the University of Medicine and Dentistry of New Jersey . Female BALB/c mice ( age , 6 to 8 weeks; weight , 17 to 20 g ) ( Charles River Laboratories , Wilmington , MA ) were used throughout the experiments . The mice were housed in micro-isolator cages with five animals per group and had access to food and water ad libitum . All animal experiments were conducted in the PHRI-UMDNJ Research Animal Facility . Disseminated infection with C . albicans ( ATCC 36082 ) was induced by injection of 5 . 0×105 blastoconidia ( ∼5 times the median lethal dose ) in 0 . 1 ml of sterile saline via the lateral tail vein of female BALB/c mice . Therapy was initiated 3 hours after challenge . Mice were treated with vehicle , Fluconazole ( LKT Laboratories Inc . , St . Paul , MN ) ( 0 . 1 to 60 mg/kg of body weight/dose ) , Amiodarone ( Henry Schein Animal Health , Melville , NY ) ( 5 . 0 or 25 mg/kg of body weight/dose ) or Fluconazole ( 0 . 1 to 60 mg/kg of body weight/dose ) + Amiodarone ( 5 . 0 or 25 mg/kg of body weight/dose ) administered intraperitoneally once a day for a total of 4 days . On day 4 post-infection , kidneys from euthanized mice ( 5 per group , unless specified ) were aseptically removed , weighed , and homogenized in sterile saline using an IKA Works ULTRA TURRAX Tube Disperser Workstation ( IKA Works Inc . , Wilmington , NC ) . Serial dilutions of homogenized kidneys were plated onto YPD agar plates containing chloramphenicol ( 70 µg/ml ) and ampicillin ( 50µg/ml ) to eliminate bacterial cross-contamination . Culture plates were incubated at 30°C for 48 h , after which the CFU were counted and the number of CFU per gram of tissue was calculated . The method was sensitive for detection of ≥10 CFU/g . The culture-negative plates were counted as having 0 CFU/g . The association of C . albicans proliferation with treatment group was assessed using multiple linear regression of log-transformed CFU counts . Preliminary model fit analysis was conducted to confirm that a parsimonious model which included dose-response slopes for AMD and FLZ , and a term for synergistic effect of dual drug treatment fit the data as well as the saturated model , where each of the nine treatment and dosage combinations were independent predictors ( likelihood ratio test p = 0 . 47 ) . The regression coefficients for drug dose are interpreted as slopes of the dose response curves . For the synergistic term , the exponentiated coefficient is the ratio of geometric means of the CFU for any given dose of each drug when the other drug is present versus when the other drug is absent . The significance of the association of CFU with predictors was assessed using Wald t-statistics for the regression coefficients . | Systemic fungal infections impose a significant threat to public health and therapeutic options to treat these diseases remain limited . Azoles represent the largest category of anti-fungal drugs and repress fungal growth by inhibiting biosynthesis of ergosterol , an important constituent of fungal membranes . Despite the wide use of azoles in the clinic for decades , the cellular basis for their antifungal mechanism remains elusive . In this study , we use a range of genetic , cellular and biochemical approaches to reveal a requirement for ergosterol in vacuolar H+-ATPase function . V-ATPase plays essential roles in diverse cellular processes , and is required for fungal virulence . Concomitant ergosterol feeding restores vacuolar acidification and growth in cells treated with fluconazole . These results suggest that the critical requirement for ergosterol in V-ATPase function may underlie the antifungal activity of azoles . Moreover , we show in a mouse Candidiasis model that combining an ion homeostasis-disruptive drug with azole is an effective approach to treat fungal infections . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"pharmacology",
"microbiology/medical",
"microbiology",
"biochemistry/membrane",
"proteins",
"and",
"energy",
"transduction",
"infectious",
"diseases/antimicrobials",
"and",
"drug",
"resistance"
] | 2010 | Requirement for Ergosterol in V-ATPase Function Underlies Antifungal Activity of Azole Drugs |
Group A Streptococcus ( GAS ) is deleterious pathogenic bacteria whose interaction with blood vessels leads to life-threatening bacteremia . Although xenophagy , a special form of autophagy , eliminates invading GAS in epithelial cells , we found that GAS could survive and multiply in endothelial cells . Endothelial cells were competent in starvation-induced autophagy , but failed to form double-membrane structures surrounding GAS , an essential step in xenophagy . This deficiency stemmed from reduced recruitment of ubiquitin and several core autophagy proteins in endothelial cells , as demonstrated by the fact that it could be rescued by exogenous coating of GAS with ubiquitin . The defect was associated with reduced NO-mediated ubiquitin signaling . Therefore , we propose that the lack of efficient clearance of GAS in endothelial cells is caused by their intrinsic inability to target GAS with ubiquitin to promote autophagosome biogenesis for xenophagy .
Streptococcus pyogenes , also known as group A Streptococcus ( GAS ) , is a common human pathogen that causes a variety of illnesses , ranging from mild self-limiting infections to severe invasive diseases . Pathogenesis involves various virulence factors for adhesion , invasion , colonization , and defense against the immune system [1–4] . Although GAS is defined as an extracellular bacterium that is recognized by phagocytes through PRRs ( pattern recognition receptors ) , triggering further immune responses [5] , it can also invade eukaryotic cells , allowing the bacteria to escape from immune cell clearance and antibiotic killing [6–8] . Nonetheless , the host can eliminate internalized bacteria via xenophagy [9–11] , a specialized form of the intracellular bulk degradation system known as autophagy . The xenophagy pathway is utilized not only in specialist immune cells , but also in other cell types such as epithelial cells . Autophagy digests intracellular components to obtain minimum energy and basic building blocks to ensure cellular survival during starvation conditions . Following non-selective engulfment of cytoplasmic components by a double-membrane structure called the autophagosome , these compartments fuse with lysosomes , where the cargos are degraded by hydrolases [12 , 13] . Autophagy can also specifically target intracellular components including damaged organelles , protein aggregates , and invading bacteria in order to maintain cellular and systemic homeostasis . As a form of selective autophagy , xenophagy requires a mechanism for cargo recognition . In addition , xenophagy is distinguished from other forms of autophagy by its requirement for engulfment of bacteria . The GAS-containing autophagosome-like vacuoles ( GcAVs ) , which trap invading GAS inside host cells , are often nearly 10 μm in diameter , and are thus capable of sequestering multiple bacteria; by contrast , regular autophagosomes range between 0 . 5 and 1 . 5 μm [11 , 14] . To build up this large structure , xenophagy requires Rab7 , Rab9A , and Rab23 GTPases in addition to the core sets of autophagy genes [15 , 16] . The most significant difference between xenophagy and other forms of selective autophagy is that it targets intruders from outside the body . Thus , in order to understand the significance and mechanism of xenophagy , we must consider the penetration path and cell-type specificity of the invading pathogens . Most research in xenophagy has been focused on epithelial cells in the intestinal and respiratory tracts , because these cells are the primary targets of bacterial invasion into organisms . However , given that intrusion of bacteria into the cardiovascular system causes their dissemination throughout the body , leading to potentially fatal consequences , endothelial cells should also be investigated as important targets . Consistent with this idea , GAS can enter human umbilical vein endothelial cells ( HUVEC ) [17] , and invasive GAS can survive in endothelial cells [18 , 19] . However , the details of bacterial fate after engulfment into endothelial cells remain poorly defined . In this study , we found that GAS could survive and multiply in endothelial cells , whereas epithelial cells efficiently removed them via functional xenophagy . Invading GAS is decorated with the autophagosome marker LC3 in endothelial cells , but they could not be surrounded by the autophagic double-membrane structure , leading to fatal consequences . This intrinsic defect in endothelial xenophagy most likely stems from insufficient ubiquitination of invading GAS mediated by the nitric oxide ( NO ) pathway . Our findings regarding the cell-type specificity of xenophagy provide essential insights into the mechanisms of cellular defenses against bacterial invasion in vivo .
To determine the fate of GAS following internalization into endothelial cells , we compared the human microvascular endothelial cell line-1 ( HMEC-1 ) with lung epithelial A549 cells . The efficiency of GAS ( NZ131 , type M49 ) invasion was 4–5 times higher in endothelial cells ( 10 . 6±2 . 031 x 104 cfu/ml ) than epithelial cells ( 2 . 433±0 . 339 x104 cfu/ml ) , so we adjusted the ratio of the MOI between endothelial and epithelial cells to 1:5 in order to monitor GAS survival following engulfment of equal numbers of GAS into both cell types . Strikingly , a time-course growth analysis of intracellular GAS revealed that GAS multiplied in HMEC-1 cells , but remained suppressed in A549 cells ( Fig 1A ) . GAS , detected as DAPI-stained dots in the cytoplasm , increased in abundance in a time-dependent manner ( Fig 1A ) . Majority of endothelial cells with GAS was still viable at 6 hours post infection , but eventually resulted in necrotic cell death at 24 hours ( S1A–S1D Fig ) . To confirm that internalized GAS is capable of replication , we recovered bacteria inside the cells and performed colony-formation assays . The results revealed an increase in the number of colonies formed by GAS recovered from endothelial cells 6 h post-infection , whereas those from epithelial cells exhibited reduced viability ( Fig 1B ) . This expansion was not specific to HMEC-1 cells , as we observed a similar phenotype even in untransformed , primary human umbilical vein endothelial cells ( HUVEC ) ( Fig 1B ) , suggesting that the insufficient GAS clearance is not the artificial event specific to immortalized cells . Furthermore , in addition to A549 cells , GAS did not grow in two other epithelial cell types ( HeLa and NRK ) . This endothelial defect in GAS growth suppression is not specific to the M49 serotype , because HMEC1 cells , but not control A549 cells , failed to suppress expansions M1 and M6 GAS strains as well ( Fig 1C ) . We assumed that bacterial virulence factors might be involved in this event , and thus examined the growth of GAS harboring a mutation in emm1 gene encoding a M protein , or hasA gene which is required for formation of hyaluronic acid capsule . We observed the expansion of GAS with either mutation in endothelial cells ( Fig 1D and 1E ) , suggesting that GAS expansion in endothelial cells is not due to a lack of endothelial defense mechanism against those virulence factors . Moreover , we determined specie specificity for this defect . To this end , we infected cells with Salmonella and also S . aureus . Both two bacterial strains showed striking growth in endothelial cells but not in control epithelial cells , suggesting that this is not an event specific to GAS but rather a phenomenon generally observed ( Fig 1F ) . Taken together , we concluded that endothelial cells are deficient in suppression of intracellular growth of GAS . To obtain mechanistic insights into the failure of endothelial cells to suppress GAS growth , we studied the role of autophagy in GAS clearance in light of its significant role in GAS elimination in epithelial cells [11 , 15 , 16 , 20–22] . This idea is consistent with our observation that GAS with mutation in slo encoding Streptolysin O ( SLO ) did not expanded even in endothelial cells ( Fig 1D ) , as GAS mutated in slo is degraded by a manner independent from autophagy [20] . To compare GAS autophagy between the two cell types , we first sought to exclude the possibility that endothelial cells have a defect in canonical autophagy induced by serum starvation . In response to starvation , endothelial cells are able to form LC3 puncta and exhibited autophagic flux at levels comparable to those in epithelial cells ( S1E–S1G Fig ) , indicating that endothelial cells retain normal activity of canonical autophagy . Next , we focused on selective autophagy in response to GAS infection . We observed that GAS induced LC3 lipidation in a time- and infectious dose–dependent manner , as well as LC3 puncta formation in endothelial cells ( S2A–S2C Fig ) , implying that autophagy could be induced by GAS invasion even in endothelial cells . However , using conventional EM , we rarely observed GAS surrounded by double-membrane structures in endothelial cells , whereas GAS in epithelial cells were positive for the structure ( Fig 2A ) . To rule out the possibility that formation of the structure was simply delayed , we performed EM at later time points after infection . At these later times , we still found no isolation membrane surrounding the bacteria , and GAS continued growing in the cytoplasm of endothelial cells ( S2D Fig ) . These results strongly suggest that endothelial cells are defective in sequestration of cytoplasmic GAS with autophagosome , and that the elevation in LC3 lipidation induced by GAS is unlikely to be related to the ability to suppress bacterial growth . To investigate the molecular underpinnings of the less efficient xenophagy in endothelial cells , we asked whether endothelial cells were capable of detecting invading GAS . It has been suggested that invading bacteria can be recognized as invaders by the ruptured endosome surrounding them: luminal exposure of ruptured endosomes facilitates further recruitment of ubiquitin , which plays an essential role in xenophagy [23–25] . To explore this possibility , we measured GAS with Gal-3 , a marker for ruptured endosome [24] . However , we observed no significant difference in the proportion of GAS with Gal-3 signals in endothelial versus epithelial cells ( Fig 2B ) , suggesting that initial cue for xenophagy is triggered by GAS , even in endothelial cells . Moreover , we could observe overlap between the LC3 and Gal-3 signals on invading GAS , even in endothelial cells ( Fig 2C ) . However , correlative light electron microscopy ( CLEM ) analysis clearly showed that those LC3 signals were not correlated with autophagosomal double membranes ( Fig 2D and S3A–S3D Fig ) , again indicating that endothelial cells were defective in formation of autophagosomes surrounding invading GAS . Consistent with this , the number of LC3-positive GAS kept increasing over time in HMEC-1 cells , whereas their multiplication was suppressed in epithelial cells ( S4A and S4B Fig ) , suggesting that the LC3 signals on GAS in endothelial cells did not indicate functional autophagosomes . A lack of autophagosome formation surrounding GAS is observed in primary HUVEC as well . Consistently , in HUVEC , GAS is able to be positive to both LC3 and Gal-3 signal but totally negative to autophagosomal double membrane structures ( S3E Fig ) . Taken together , these findings indicate that endothelial cells are incompetent to suppress GAS growth due to impairment in formation of autophagosome membranes surrounding GAS with ruptured endosomes . To determine the mechanism that gives rise to defective formation of autophagosomal membrane surrounding GAS , we focused on ubiquitination , which is an essential step for initiation of selective autophagy [26–28] . Consistent with this , deregulation of ubiquitination impairs recruitments of downstream components of xenophagy [29 , 30] . Strikingly , we found that recruitment of ubiquitin to GAS was reduced in endothelial cells in comparison with epithelial cells ( Fig 3A ) , correlated with defective recruitment of downstream ATG proteins . The recruitments of ULK1 , ATG14 , and ATG9 , all involved in formation of isolation membrane and xenophagy [31 , 32] , were significantly reduced in endothelial cells ( Fig 3B and S5A Fig ) . Not only for GAS , we confirmed that invading Salmonella is targeted by ATG9 at a lower level in endothelial cells than that in epithelial cells , together indicating that endothelial cells are compromised in early step of xenophagy ( S5B Fig ) . As expected , GAS with LC3 signals were observed regardless of ubiquitination in endothelial cells , in contrast to the case of epithelial cells , where most LC3-positive GAS were associated with ubiquitin ( Fig 3C and 3D ) . Accordingly , we hypothesized that ineffective ubiquitination on bacteria could be the major cause of the defect in xenophagy in endothelial cells . To directly test this hypothesis , we pre-coated GAS with ubiquitin and asked whether they would be eliminated even in endothelial cells . To generate a physiologically relevant coat of ubiquitin , we used epithelial cells as ‘craftsmen’ . Notably , we used ATG9-knockout ( KO ) HeLa epithelial cells ( S6A–S6C Fig ) for this purpose in order to avoid degradation of the product by the functional xenophagy system in wild-type cells . Using this system , we could successfully recover GAS with a ubiquitin coat from ATG9-KO HeLa cells , whose integrity of ubiquitination was confirmed by immunostaining ( Fig 4A and 4B ) . To recover intact GAS with the coat , we treated infected HeLa cells with distilled water without any detergents so that the osmotic imbalance would disrupt the plasma membrane of the host cells , concurrently enabling us to exclude the possibility that harvested GAS were surrounded by intact endosome structure that might burst under such harsh condition . When we infected endothelial cells with naïve GAS or ubiquitin-coated GAS , we found that , even in endothelial cells , GAS were engulfed by autophagosomal double membrane , and efficiently eliminated if they were coated with ubiquitin ( Fig 4C and 4D ) . This observation strongly supports the hypothesis that the defect in xenophagy in endothelial cells results from reduced ubiquitination activity on invading GAS , rather than a deficiency in autophagosome biogenesis itself . This is consistent with our observation showing that endothelial cells retained canonical autophagic activity , comparable to the level in epithelial cells ( S1A–S1C and S2A–S2C Figs ) . The observed clearance of ubiquitin-coated bacteria in endothelial cells requires the intact autophagic machinery , because ATG9-KO endothelial cells ( S6D–S6F Fig ) failed to suppress the replication of GAS with a ubiquitin coat ( Fig 4D ) . In addition , under naïve GAS infection , bacterial growth was slightly exacerbated in ATG9-KO endothelial cells in comparison with wild-type cells ( Fig 4D and S6F Fig ) , suggesting that autophagic machinery does indeed help endothelial cells to defend against bacterial infection , even though the resultant suppression of GAS growth is far from sufficient . To further explore these findings , we sought to determine whether the defect in ubiquitination was functionally associated with the nitric oxide ( NO ) -mediated signaling cascade . Recent work showed that this pathway facilitates ubiquitination and selective clearance of intracellular GAS in macrophages [33] . GAS invasion induces NO synthase ( NOS ) -mediated NO production , resulting in generation of 8-nitroguanosine 3′ , 5′-cyclic monophosphate ( 8-nitro-cGMP , an endogenous derivative of cGMP ) . It is followed by S-guanylation on cysteine residues of targets at the bacterial surface . S-guanylation promotes ubiquitination of the bacteria ( summarized in S7A Fig ) . In this cascade , L-NMMA , a NOS inhibitor , or NaHS , an 8-nitro-cGMP eliminator , can inhibit ubiquitination and degradation of bacteria [33] . To determine whether the NOS-ubiquitin pathway is intact in endothelial cells , we measured endogenous 8-nitro-cGMP levels in these cells , and found that they had an intrinsically lower level of 8-nitro-cGMP than epithelial cells ( Fig 5A ) . Additionally , GAS infection–induced 8-nitro-cGMP failed to compensate for this defect , whereas exogenous introduction of 8-nitro-cGMP increased its level inside the cell , even in endothelial cells . This observation indicates that endothelial cells are intrinsically deficient in both production and induction of 8-nitro-cGMP . This conclusion was further confirmed by treating cells with two inhibitors , L-NMMA and NaHS . These experiments showed that the poor ubiquitination of GAS in endothelial cells was not altered by the inhibitors , whereas ubiquitination in epithelial cells was significantly reduced ( Fig 5B ) . Notably , this is not merely due to low levels of 8-nitro-cGMP in endothelial cells , because introduction of exogenous 8-nitro-cGMP failed to restore GAS ubiquitination in these cells , implying that , in addition to an intrinsically low level of 8-nitro-cGMP , endothelial cells lack a mechanism to facilitate ubiquitination of GAS mediated by this compound ( Fig 5B ) . Consistent with this , endothelial cells did not exhibit any detectable increase in GAS growth upon treatment with L-NMMA or NaHS ( Fig 5C and 5D , S7B–S7D Fig ) , whereas GAS replication was elevated in epithelial cells treated with those inhibitors or harboring silencing/knockout of autophagy genes , supporting the idea that endothelial cells are deficient in bacterial clearance via NOS–ubiquitin pathway . Taken together , our data suggests that deficiency in the NOS–ubiquitin pathway in endothelial cells is at least partially responsible for their diminished ability to defend themselves against GAS . Additionally , our results also imply that endothelial cells not only fail to form GcAVs , but also cannot induce formation of regular autophagosomes in response to 8-nitro-cGMP . Formation of LC3 puncta formation could be induced by 8-nitro-cGMP in epithelial cells , but not in endothelial cells ( Fig 5E ) , indicating that the lack of the NOS-dependent induction of autophagy in endothelial cells is linked to a wide range of physiological outcomes related to autophagy .
In this report , we showed that xenophagy is not equally efficient among different cell types , raising the question of why endothelial cells do not retain this vital ability to defend themselves against bacteria . We found that GAS efficiently replicates in endothelial cells at a rate comparable to that of in vitro GAS growth [34 , 35] , indicating that the bacteria are essentially growing freely . We speculate that this situation evolved because robust activation of the NO pathway in endothelial cells could cause a lethal side effect on cardiovascular systems that would outweigh the beneficial effects of activating xenophagy . Essentially , constitutive releases of NO from endothelial cells via endothelial NOS ( eNOS ) serves a protective role in cardiovascular homeostasis by relaxing blood vessel pressure [36] . However , excess induction of NO , which activates the 8-nitro-cGMP pathway and xenophagy , could deregulate blood vessel pressure . In fact , robust NO induction by bacterial infection or LPS causes hypotension and sepsis , a phenomenon mediated by inducible NOS ( iNOS ) [37] . We found that endothelial cells have low levels of 8-nitro-cGMP and do not efficiently ubiquitinate invading GAS in response to 8-nitro-cGMP . This is consistent with the fact that NO production in endothelial cells is directly connected to that of NO in vascular muscle [36] , creating a situation more sensitive than that in epithelium . Therefore , we speculate that , given the need for tight regulation of this pathway , endothelial cells had to abandon NO-mediated induction of xenophagy . We found that exogenous coating of GAS with ubiquitin was sufficient to induce their clearance , even in endothelial cells , in a manner dependent on the autophagic machinery . This observation is in line with the recently established consensus that ubiquitination plays an essential role in cargo recognition in selective autophagy . Thus , endothelial cells lack a key ability to target invading bacteria . What , then , is the underlying mechanism explaining the failure of endothelial cells to perform this function ? We suspect that this deficiency is not merely due to their poor NO-pathway response . We showed that in epithelial cells , NO pathway inhibitors only partially increased bacterial growth , to levels lower than those observed in ATG7-KO cells ( Fig 5D ) . This was not simply because of an insufficient suppression of the NO-mediated pathway , as much higher doses of inhibitors did not further increase GAS growth ( S7E Fig ) , It suggests that inhibition of epithelial NO-mediated ubiquitination pathway only partially recapitulates the xenophagy defect in endothelial cells , which is supposed to be caused by insufficiency in multiple systems . We assume that it is mediated by ubiquitin targeting system involving specific E3 ligase ( s ) [38–40] . Identification of these pathways in endothelial cells is a high priority because it may facilitate development of a novel approach to fighting GAS , in particular to alleviate their expansion in cardiovascular system and prevent life-threatening bacteremia , which is often exacerbated by antibiotic resistance or delay in a treatment . In this study , we found that invading GAS is often positive for the autophagosome marker LC3 . However , we found that these LC3 signals are not related to functional autophagy . Considering the poor ability of endothelial cells to eliminate GAS , this previously unknown autophagy-independent coating of GAS with LC3 may play only a minor role in innate immunity , whereas ATG9 knockout results in slightly exacerbated GAS growth , even in endothelial cells . Because LC3 lipidation can be induced by GAS infection in endothelial cells ( S2A–S2C Fig ) , LC3 molecules on GAS are likely to be in the lipidated form . We speculate that the autophagic machinery facilitates LC3 lipidation for this event trying to limit their growth by autophagy-independent mechanism although it remains unclear how lipidated LC3 is recruited to invading GAS . It could be related to another autophagy-independent role for autophagy-related genes , as represented by LC3-associated phagocytosis ( LAP ) , although it is less effective than LAP . In summary , we showed that endothelial cells are not capable of carrying out xenophagy due to an intrinsic defect in ubiquitin-targeting system . Because ubiquitination is a universal key event in selective autophagy , further characterization of the mechanism may provide perspective and valuable insights into host defenses , as well as alleviate a wide range of diseases suppressed by selective autophagy .
Human microvascular endothelial cell line-1 ( HMEC-1 ) [41] ( obtained from the Centers for Disease Control and Prevention , USA ) was cultured in endothelial cell growth medium M200 ( Cascade Biologics ) supplemented with 10% fetal bovine serum ( FBS ) , 1 μg/ml hydrocortisone , 10 ng/ml epidermal growth factor , 3 ng/ml basic fibroblast growth factor , and 10 μg/ml heparin . Human umbilical vein endothelial cells ( HUVEC , gift from Dr . H . Y . Lie ) were maintained in M200 medium containing 10% FBS as HMEC-1 cells . A549 ( laboratory stock ) , HeLa ( laboratory stock ) , and NRK ( normal rat kidney epithelial , laboratory stock ) cells were maintained in Dulbecco’s modified Eagle’s medium ( DMEM ) supplemented with 10% FBS . Cells were cultured at 37°C in 5% CO2 . Streptococcus pyogenes strain NZ131 ( type M49 ) is a gift from Dr . D . R . Martin ( New Zealand Communicable Disease Center , Porirua ) . A20 ( type M1 ) , JRS4 ( type M6 ) wild type , emm1 mutant , slo mutant , hasA mutant GAS strains and standard strain of Staphylococcus aureus ( ATCC 25923 ) are provide from Dr . J . J . Wu . Salmonella enterica serovar Typhimurium SR-11 x3181 was used for infection . Bacteria were grown overnight at 37°C in tryptic soy broth with 0 . 5% yeast extract ( TSBY ) for GAS and S . aureus , and LB broth for Salmonella , and then transferred to fresh broth for 3 h . The culture was centrifuged and suspended in phosphate-buffered saline ( PBS ) , followed by measurement of cell concentration as ( 0 . 2 OD600 = 1 × 108 cfu/ml , confirmed by plating ) . This procedure for bacterial preparations was used for all infectious experiments in this study . For heat inactivation , the suspended bacteria were treated at 70°C for 30 min . A monolayer of cells was plated in 24-well or 6-well plates and incubated overnight . The prepared bacteria were directly added to the wells at a multiplicity of infection ( MOI ) of 1 , 5 , 10 , or 25 , and then centrifuged at 500 g for 5 min to ensure simultaneous infection of cells . In one set of experiments , the ratio of GAS and S . aureus infectious MOI between epithelial cells and endothelial cells was maintained at 5:1 , due to the 5-fold higher internalization efficiency in endothelial cells relative to epithelial cells . Salmonella was used at MOI of 100 for both cell type infections . The cell and bacteria mixture was incubated at 15 min for Salmonella and 30 min for GAS and S . aureus . After incubation , the cell culture was washed twice with PBS to remove unattached bacteria , and then fresh medium containing 100 μg/ml gentamicin was added to kill the remained extracellular bacteria . After various time periods , cells were collected for individual experiments . Cells seeded at 6 × 104/well in 24-well plates with cover glasses were cultured overnight and infected with GAS according to the infection protocol . Cover glasses were coated with cellular matrix ( Cellmatrix type I-C , 100 μg/ml , 37°C , 30 min ) in advance . At various time points post-infection , the cells were fixed with 4% paraformaldehyde ( PFA ) , permeabilized with 50 μg/ml digitonin , and stained with anti-GAS ( gift from Dr . J . J . Wu ) , anti-galectin-3 ( M3/38 , Santa Cruz Biotechnology ) , anti-LC3 ( PM036 , MBL ) , anti-FK2 ( BML-PW8810 , ENZO ) , anti-LAMP-1 ( H4A3 , Santa Cruz Biotechnology ) , or anti-8-nitro-cGMP ( 1G6 ) antibodies ( gift from Dr . T . Akaike ) , followed by staining with secondary antibodies conjugated with Alexa Fluor 488 or 568 and imaging on a confocal microscope ( FV1000; Olympus ) . Cells at 5 × 104/well in 24-well plates were incubated overnight and transfected for 4 h with 1 μg of pEGFP-LC3 or GFP-ATGs ( pEGFP-ULK-1 , pEGFP-ATG14 , pEGFP-ATG9 ) plasmid DNA using Lipofectamine 2000 reagent ( Invitrogen , Carlsbad , CA , USA ) in OPTI-MEM medium plus complete culture medium . After incubation at 37°C for 20 h , cells were prepared for infection . Control and ATG13 siRNA ( sense 5'- GAGUUUGGAUAUACCCUUUdTdT -3' and anti-sense 5'- AAAGGGUAUAUCCAAACUCdGdT -3' ) were purchased from Sigma-Aldrich . Cells for siRNA transfection were prepared at 1 × 104/well in 24-well plates and incubated overnight . The siRNAs ( 20 nM ) were mixed with Lipofectamine RNAiMAX reagent ( Invitrogen ) in OPTI-MEM medium and added to the cell culture medium . After 4 h incubation , the medium was replaced with fresh medium . After 20 h culture , a second transfection was performed using the same protocol . Forty-eight hours after the second transfection , the cells were infected with GAS or collected for western blot assays to confirm efficient knockdown of ATG13 and measure autophagic flux . To generate knock out cell lines of autophagy-related genes , we utilized the clustered regularly interspaced short palindromic repeats ( CRISPR ) -Cas9 system [42] . We designed RNA-guided Cas9 targeting the first exons of the human ATG7 and ATG9 genes . The specific recognition sequences of the 20 bp before the protospacer adjacent motif ( PAM ) in each construct were as follows: ATG7 , 5’- CACCGAAATAATGGCGGCAGCTACG-3’; ATG9 , 5’-CACCCCCTGGGGGTGAATCAC TAT-3’ . These guide oligos , with BbsI restriction sites at both ends , were annealed with their anti-sense oligos and inserted downstream of U6 promoter in vector pSpCas9 ( BB ) -2A-GFP ( pX458 ) [43] ( purchased from Addgene ) ; the resultant plasmids were used for transfection . Single-cell sorting was performed after a 48-h transfection . After 1-week culture of single cells with antibiotics , fresh medium was added to support growth for 1 more week to allow colony formation . Genomic DNA of each clone was extracted , and the target gene was confirmed by sequencing using the following primers; ATG7 , Fw 5’-GTCGACGTTCTGGAGATCTGTTTCACAACG-3’ and Re 5’-GAATTCTGGGATCAAAAAGTCAGGAAG-3’; ATG9 , Fw 5’-ATATGTCGACCAGGATGAGCTCCATTCCCGT-3’ and Re 5’-ATATGAATTCCAGCCCCCAACAAAGGGACAG-3’ . Cells were seeded at 8 × 104 cells/well in 24-well plates , incubated , overnight , and then infected with GAS . At various times post-infection , cells were lysed with lysis buffer containing protease inhibitor mixture , followed by denaturation at 95°C in sample buffer , SDS-PAGE , and immunoblotting using rabbit anti-LC3 ( PM036 , MBL ) , anti-p62 ( PM045 , MBL ) , anti-ATG13 ( SAB4200 , Sigma-Aldrich ) , anti-ATG7 ( 013–22831 , Wako ) , and mouse anti-GAPDH antibodies ( Millipore ) . Cells were infected with GAS at MOI of 25 and 5 for 1 h and Ub-coated GAS for 2 h . Cell pellets were gently collected by trypsinization and resuspended in phenol red–free medium containing 5% FBS and 40% Dextran T2000 . Samples were kept on ice prior to high-pressure fixation using a High Pressure Freezer ( Leica EM HPM100 ) . After fixation , the specimens underwent freeze-substitution under low temperature and embedding in plastic ( Epon812 , TAAB Laboratories Equipment , Aldermaston , UK ) . Ultrathin sections ( 70 nm thick ) were stained with saturated uranyl acetate and Reynolds lead citrate solution . Micrographs were acquired on a JEOL JEM-1011 transmission electron microscope ( JEM-1011 , JEOL , Tokyo , Japan ) . Cells stably expressing GFP-LC3 and mStrawberry-galectin3 were cultured on glass-bottom dishes with a grid pattern ( P35G-2-14-C-GRID; MatTek , Ashland , MA , USA ) and infected for 1 h with GAS at an MOI of 25 for A549 , and 5 for HMEC-1 and HUVEC cells . The cells were fixed with 4% formaldehyde in HEPES buffer ( 30 mM HEPES , 100 mM NaCl , 2 mM CaCl2 , pH 7 . 4 ) , and 1 μg/ml DAPI for 30 min at room temperature , washed in HEPES buffer , and observed using a confocal microscopy ( FV1000; Olympus ) . After marking the locations of the target cells , the same specimens were further incubated with 2% formaldehyde and 2 . 5% glutaraldehyde in HEPES buffer at 4°C overnight . After three washes , the samples were post-fixed with 1% osmium tetroxide and 0 . 5% potassium ferrocyanide in HEPES buffer for 1 h , washed three times in distilled water , dehydrated in ethanol , and embedded in Epon812 ( TAAB Laboratories Equipment , Aldermaston , UK ) . Ultrathin sections ( 70 nm thick ) were stained with saturated uranyl acetate and Reynolds lead citrate solution . Micrographs were acquired on a JEOL JEM-1011 transmission electron microscope . After overnight culture , ATG9-KO HeLa epithelial cells ( 4 × 105 cells/well in 6-well plates ) were infected with GAS for 30 min at an MOI of 25 , as described in the infection protocol . Gentamicin was added to the culture medium to kill extracellular bacteria . At 3 h post-infection , GAS replicated in the cytoplasm of ATG9-KO HeLa epithelial cells . To obtain intracellular GAS from infected host cells , the cells were treated with distilled water , causing an osmotic imbalance that disrupted the plasma membranes of host cells and the endosomal membranes surrounding the GAS . Bacterial number was determined before next infection ( average ~107 cfu/ml ) . No centrifugation was performed in order to avoid aggregation between cell debris and Ub-coated GAS . For ubiquitin staining of GAS , the bacteria were attached to poly-L-Lysine–coated cover glasses and fixed with 4% PFA , and then subjected to immunofluorescence staining with anti-FK2 and anti-GAS antibodies . For conventional EM sample , HMEC-1 cells were infected with Ub-coated GAS at an MOI of 5 ( 25 μl of the 1 ml lysate: expected bacterial number , 2 . 5x105 cfu ) for 2 h . Samples were fixed and prepared for convention EM observation . For the xenophagy rescue experiment , wild-type and ATG9-KO endothelial cells ( 1 × 105 cells/well , in 24-well dish ) were infected with Ub-coated GAS at an MOI of 1 ( 10 μl of the 1 ml lysate: expected bacterial number , 105 cfu ) . GAS prepared in regular culture were used as controls . Colony formation assays were performed at 1 and 6 h post-infection . The fold increase in GAS number , reflective of replication , was calculated as the bacterial number at 6 h normalized against the number at 1 h . Cells were infected with bacteria as described in the infection protocol . At the indicated time point post-infection , bacteria-infected cells were washed twice with PBS and lysed in sterile H2O , 1 ml/well ( 24-well plates ) . After serial dilution with PBS , the bacteria-containing PBS was plated on TSBY or LB agar plates . Colonies were counted after 24 or 48 h incubation at 37°C . Bacterial number determined 1 h post-infection was interpreted as the number of internalized GAS . The fold increase in bacterial number was calculated as the number at later time point normalized against the number of internalized bacteria . Each colony-forming assay was performed at least three times . | Autophagy is an intracellular bulk degradation system to survive within starved condition , which is one of the most common threats limiting organism’s expansion . By the system , cells digest their own cytoplasmic compartments that are sequestered by double membrane structure called autophagosome . It is also utilized for selective targeting of unwanted materials inside the cells including invading bacteria . This system targeting bacteria is called xenophagy , and provides “non-specialists” with innate immune system . Xenophagy is well characterized in epithelial cells since they are primary targets for invading bacteria . However , even though bacterial penetration into blood vessel could cause severe symptoms , it remains unknown weather endothelial cells retain functional xenophagy fighting against them . In this report , we showed that endothelial cells fail to suppress group A Streptococcus ( GAS ) growth due to defect in xenophagy . It is due to endothelial cell’s defect in ubiquitination of GAS , which plays a key role in target recognition during selective autophagy . Because we confirmed that endothelial cells are proficient in canonical , non-selective autophagy , our findings illustrate intrinsic defect in xenophagy as unique but general character of endothelial cells , shedding light onto cell type specificity in selective autophagy . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"cell",
"death",
"medicine",
"and",
"health",
"sciences",
"autophagic",
"cell",
"death",
"neurochemistry",
"pathology",
"and",
"laboratory",
"medicine",
"gene",
"regulation",
"endothelial",
"cells",
"pathogens",
"cell",
"processes",
"microbiology",
"neuroscience",
"salmonellosis",
"epithelial",
"cells",
"bacterial",
"diseases",
"enterobacteriaceae",
"bacteria",
"bacterial",
"pathogens",
"small",
"interfering",
"rnas",
"infectious",
"diseases",
"animal",
"cells",
"proteins",
"neurochemicals",
"nitric",
"oxide",
"medical",
"microbiology",
"gene",
"expression",
"microbial",
"pathogens",
"biological",
"tissue",
"ubiquitination",
"salmonella",
"biochemistry",
"rna",
"cell",
"biology",
"anatomy",
"post-translational",
"modification",
"nucleic",
"acids",
"genetics",
"epithelium",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"non-coding",
"rna",
"organisms"
] | 2017 | Endothelial cells are intrinsically defective in xenophagy of Streptococcus pyogenes |
The pyrabactin resistance 1 ( PYR1 ) /PYR1-like ( PYL ) /regulatory component of abscisic acid ( ABA ) response ( RCAR ) proteins comprise a well characterized family of ABA receptors . Recent investigations have revealed two subsets of these receptors that , in the absence of ABA , either form inactive homodimers ( PYR1 and PYLs 1–3 ) or mediate basal inhibition of downstream target type 2C protein phosphatases ( PP2Cs; PYLs 4–10 ) respectively in vitro . Addition of ABA has been shown to release the apo-homodimers yielding ABA-bound monomeric holo-receptors that can interact with PP2Cs; highlighting a competitive-interaction process . Interaction selectivity has been shown to be mediated by subtle structural variations of primary sequence and ligand binding effects . Now , the dynamical contributions of ligand binding on interaction selectivity are investigated through extensive molecular dynamics ( MD ) simulations of apo and holo-PYR1 in monomeric and dimeric form as well as in complex with a PP2C , homology to ABA insensitive 1 ( HAB1 ) . Robust comparative interpretations were enabled by a novel essential collective dynamics approach . In agreement with recent experimental findings , our analysis indicates that ABA-bound PYR1 should efficiently bind to HAB1 . However , both ABA-bound and ABA-extracted PYR1-HAB1 constructs have demonstrated notable similarities in their dynamics , suggesting that apo-PYR1 should also be able to make a substantial interaction with PP2Cs , albeit likely with slower complex formation kinetics . Further analysis indicates that both ABA-bound and ABA-free PYR1 in complex with HAB1 exhibit a higher intra-molecular structural stability and stronger inter-molecular dynamic correlations , in comparison with either holo- or apo-PYR1 dimers , supporting a model that includes apo-PYR1 in complex with HAB1 . This possibility of a conditional functional apo-PYR1-PP2C complex was validated in vitro . These findings are generally consistent with the competitive-interaction model for PYR1 but highlight dynamical contributions of the PYR1 structure in mediating interaction selectivity suggesting added degrees of complexity in the regulation of the competitive-inhibition .
The plant hormone abscisic acid ( ABA ) controls seed development , germination , dormancy and stress response [1] , [2] . Using combined molecular , biophysical and genetic techniques many details of ABA signal transduction have been elucidated [3]–[5] . In particular , under abiotic stress , such as drought and high salinity , ABA levels have been shown to increase in the plant , initiating adaptive responses involving inhibition of type 2C protein phosphatases ( PP2C ) , and stimulation of protein sucrose non-fermenting related kinases 2 ( SnRK2 ) . In 2009 , two research groups independently reported a family of at least 13 ABA-binding proteins in Arabidopsis thaliana , known as PYR1 ( pyrabactin resistance 1 ) and PYL ( PYR1-like ) or RCAR ( regulatory component of ABA response ) proteins , that play a central role in ABA signal transduction mediating PP2C inhibition [6]–[8] . The structure of a typical PYR/PYL receptor comprises seven β-strands and two α-helices [9]–[11] . The second α-helix , located in the C-terminal region of PYR1 , forms a helix-grip fold , which in turn provides a large hydrophobic cavity , the ligand pocket ( Figure 1 ) . Upon binding to the receptor , ABA makes direct stabilizing contacts with two flexible loops ( Lβ3β4 and Lβ5β6 named gate/proline cap and latch/leucine lock ) , which then act as a scaffold mediating an interaction with the PP2C [11] . In the course of PYR/PYL receptors docking to PP2Cs , some side chains of the gate ( e . g . conserved S85 of PYR1 ) interact with the phosphatase in such a way as to block access to the PP2C catalytic site [12] , [11] . The overall effect of PYR/PYL binding to PP2C is inhibition of the phosphatase which can then no longer inhibit SnRK2 activity . As a result , active SnRK2 phosphorylates and activates downstream transcription factors leading to well documented ABA-responsive gene induction events [13] . More recently it has been found that PYR1 and PYLs 1–3 generally do not show basal activity , whereas other family members PYLs 4–10 ( not including untested PYL7 ) are constitutively active ( CA ) in vitro [7] , [14] , [15] . Interestingly however , PYL4 was only active against HAB1 under the conditions tested [15] . Indeed , the demonstrated ability of PYL10 to strongly inhibit PP2C even in the absence of any ligand has been employed to engineer a gate-modified mutant of PYL2 ( V87L ) with improved basal activity [15] . Mechanistically , the ability of a native receptor to mediate basal activity has been correlated to its inability to homodimerize in solution [15] , [16] . According to published structural data , the receptor's binding surfaces involved in the formation of dimers and complexes with PP2C mainly overlap [8] , [10] , [17] suggesting a possibility of competition between dimerization and PP2C docking . In this context , it has been shown that mutation I88K in PYL2 both prevents dimer formation in solution and increases its constitutive activity [15] . PYR1 mutation H60P was also found to yield a mixture of monomeric and dimeric PYR1 forms , which showed weak basal activity [16] , [18] . While these mutations may explain why PYLs 8–10 tend to be monomeric , factors contributing to PYLs 4–6 being monomeric remain unknown . Together these reports show that for homodimeric receptors such as wild type PYR1 , ABA binding prevents dimerization , induces dimer dissociation and stimulates receptor-PP2C interactions . Toward characterizing molecular aspects regulating basal versus inducible receptor activity against PP2Cs , numerical modeling and simulations are required . Molecular dynamics ( MD ) simulations have been reported involving phosphatase proteins other than PP2C . For example , a system comprising the N-terminal part of phosphatase SHP-2 and a peptide ( 101 residues ) has been studied using MD simulations for 10 ns [19] , and a somewhat larger complex of protein-tyrosine phosphatase 1B ( 337 residues ) with phosphorylated peptide substrate ( 193 residues ) [20] has been simulated for 1 ns . At the time of writing , the only modeling work on PYR1 and PYL1 [21] is addressing the design of small ligands that could replace ABA , improving the ligand binding energy . However in order to elucidate mechanistic information about the receptors more broadly , the entire PYR/PYL-PP2C complex needs to be simulated . The complexity is that MD simulations of such a system , which consists of approximately 8 , 000 atoms , or 530 residues , embedded in an explicit water solvent of approximately 60 , 000 atoms , at physiologically relevant timescales ( ≥100 ns ) are prohibitively computationally expensive . Toward overcoming this problem , various dimensionality reduction or coarse grained approaches may be employed to draw predictions from available MD simulations [22]–[29] . In these approaches , collective dynamical descriptors are derived from MD data by such techniques as the principal component analysis ( PCA ) [23] , [24] , normal mode analysis [30] , [27] , or related methods , expecting that relevant structural properties , such as the peptide flexibility , could be extracted [28] , [25] , [20] . Toward enabling robust interpretations , a novel essential collective dynamics ( ECD ) modeling framework has been recently introduced , which allows probing of persistent dynamic correlations in proteins based on short ( a few hundreds of picoseconds ) all-atom MD trajectories [31]–[35] . Relying on a statistical-mechanical theory , the ECD framework provides a transparent physical interpretation of the dimensionality reduction analyses in terms of a proteins' structural properties such as the main chain flexibility or dynamic domains of correlated motion , as well as allows for a reasonable match of the corresponding predictions with NMR-based measurements representing significantly longer time regimes than the MD trajectories used in the analysis . In this work , extensive MD simulations for PYR1 bound to HAB1 phosphatase as well as PYR1 homodimer [10] , [17] are reported in the presence and absence of ligand . Employing the ECD framework the structural stability and dynamics of PYR1 complexes were investigated Employing the ECD framework the structural stability and dynamics of PYR1 complexes were investigated . The results are consistent with the ABA-dependent ‘competitive interaction’ model ( receptor homodimer versus receptor-phosphatase complexes [16] ) proposed for regulation of PYR1 , suggest a stronger potential role for ABA-independent , concentration-dependent regulation of basal signaling for all PYR1/PYL receptors and define the dynamical contributions that mediate the selective interactions of PYR1 .
Molecular models of PYR1 monomers and dimers as well as PYR1-PP2C complexes in water were constructed employing crystallographic models from the Protein Data Bank ( PDB ) [36] , see also Table 1 and Methods section . Molecular dynamics simulations were carried out for these constructs and the results analyzed by the ECD method [31]–[34] as described in Methods . Examples of intra-molecular correlation maps are shown in Figure 2 for the PYR1 open lid and PYR1 closed lid conformations from Figure 1 . The maps show ECD pair correlation descriptors computed by Eq . 1 , and represent inter-atomic correlations originating from direct binding , steric constraints , and water-mediated interactions in the constructs considered . From comparison of the correlation maps , a significant loss of correlations is evident for the gate ( V83-N90 ) and latch ( E114-T118 ) regions around residues C30 , H60 , L87 and M158-L166 in the open lid PYR1 construct . The correlations of the gate with loop Lβ7α5 and helix α5 which are observed in the closed lid conformation are not pronounced in the open lid construct . Examples of the resulting main chain ECD flexibility profiles of the various PYR1 constructs are shown in Figure 3 . In the flexibility profiles , high levels of the descriptor usually correspond to flexible loops , whereas most of the flexibility minima indicate α-helices and β-sheets , in accordance with other methods of flexibility assessment [25] , [37] . In particular , high flexibilities of the loop Lβ7α5 ( residues P148-D155 ) and the gate region ( residues 85–89 ) as well as around residues Q69 , I134 and Y23 are observed . The flexibilities of open and closed lid structures in the areas of these loops differ by up to 62% . Interestingly , the flexibility of loop Lβ7α5 , which is only indirectly involved in gate closure , is affected even more strongly than that of the gate upon ABA-binding , whereas the latch flexibility is similar in these constructs , reflecting the “recoil motif” interaction upon ABA-affected PYR1 enclosure [9] . Figure 4 compares our computed flexibility profiles of the PYR1 constructs with crystallographic B factors [9] . B-factors ( or Debye-Waller factors ) of atoms are derived from X-ray diffraction intensities and are indicative of the relative structural order in the crystal structure . Low crystallographic B-factors correspond to relatively well defined lattice positions , whereas higher B-factors can be interpreted as more flexible and less ordered regions in the crystal structure . While the physical meaning of the crystallographic B factors is somewhat different from the ECD descriptors , and the dynamics of the protein in crystal structure also differs from that in solution , validation of molecular dynamics analyses against B-factors is a popular choice [34] . The backbone B-factors for PYR1 closed lid , ABA-bound construct and PYR1 open lid , ABA-free construct ( chains B and A , respectively , from PDB entry 3K3K [9] ) are plotted by dashed lines in Figure 4 . The computed flexibilities in the figure ( solid lines ) represent the corresponding profiles from Figure 3 , normalized and over-imposed on the experimental profiles to facilitate the comparison . The comparison shows a good agreement in general . For the open-lid construct , the crystallographic B-factors are relatively lower than the ECD flexibility in the N-terminal area and in the gate , whereas for the closed-lid construct , the B-factors are higher than the ECD flexibilities in the latch and in the area of loop Lβ7α5 . The reason of the differences around the gate for the open-lid construct is evident and related with constrained motion of the gate in the crystallographic structure . For the closed-lid construct , the relatively high B-factors in the regions of the loops Lβ7α5 and the latch can be attributed to the luck of stabilizing water-mediated hydrogen bonds of loops' residues with the rest of the protein in the phosphatase-binding area of the crystallographic structure . One can conclude that , considering the different physical meaning of B-factors and the ECD flexibilities , the agreement is satisfactory . The differences observed highlight the importance of a thorough analysis of protein dynamics in solution . While the availability of crystallographic structures is major to understand the dynamics , the structures alone do not fully represent important aspects that are addressable by molecular dynamics studies . Figure 5 presents typical examples of dynamic domains of collective motion in PYR1-ABA-free , open lid construct and PYR1-ABA-bound , closed lid construct obtained by ECD analysis of the MD trajectories as described in Methods . It can be seen that the largest domain in PYR1 ABA-bound construct contains the gate , sheets β7 , β6 , β5 from the ligand pocket , loops Lβ2β3 and Lβ7α5 , and helices α2 and α5 . This includes the three main structural motifs of closed lid PYR1 [9] and also the ligand pocket elements . In the open lid construct , the gate and second helix are dynamically uncoupled , and only the central part of helix α5 is correlated with the ligand pocket area . Finally , simulations for the PYR1 closed lid , ABA-extracted construct were conducted over a range of temperatures including 281 K , 300 K , 310 K and 325 K and analyzed for impact of temperature on dynamics in order to identify a suitable temperature for PYR1 simulations ( Figure S1 ) . Interestingly , the closed lid ABA-extracted construct exhibits a stable gate-latch lock , which opened only after a 20 ns simulation at increased temperature ( 325 K ) , while the open lid construct showed switching of the gate from closed to open lock positions in the course of all simulations . At 281 K the largest domain of correlated motion , which indicates the most extensive correlations , includes most of the β-sheets , gate and latch , whereas helix α5 is decoupled . At temperatures 300 K and 310 K , the gate and latch are largely decoupled; however most of helix α5 is involved in the largest domain , which contains more than 10% of the receptor's atoms . At 325 K the size of the largest domain decreases and its structure is different from that observed at 300 K and 310 K , indicating as expected , that simulations at 325 K may not be representative of physiologically relevant conditions . Finally , in addition to analyzing the dynamical domains at various temperatures , we also inspected visually the evolution of the ABA-extracted , closed lid PYR1 construct during 50 ns simulations ( Figure S2 ) . At 281 K and 300 K temperatures , the conformation remained closed most of the time . In contrast , the contacts of gate and C-terminal helix with the rest of receptor were disrupted after 30 ns of simulations at 310 K and at 325 K . Based on these tests , 310 K has emerged as a condition which allows observation of changes of conformation such as the gate and latch decoupling upon removal of ABA , whereas the dynamics of the receptor are not altered significantly in comparison to 300 K . Therefore 310 K was adopted as a physiologically relevant temperature , yet high enough to allow observation of pertinent conformational evolution in silico . The simulations described below are performed at 310 K , unless indicated otherwise . The structure of ABA-bound PYR1 complexed with homology to ABA insensitive 1 ( HAB1 ) phosphatase ( PDB ID 3QN1 ) is presented in Figure 6 . The polar and non-polar interactions on the binding surface indicated in the figure were identified using Accelrys VS , and are consistent with the original structure report [17] . Polar interactions which involve hydrogen bonds correspond to a 4 Å cutoff , and other polar interactions correspond to a 3 . 5 Å cutoff . The non-polar interactions comprising van der Waals and hydrophobic interactions correspond to a 4 . 5 Å cutoff . MD simulations conducted for the PYR1-ABA-HAB1 complex over the course of 30 ns simulations showed considerable inter-molecular correlations in the PYR1 and HAB1 binding areas as one would expect ( Figure 7 ( A ) ) . Simulations for ABA extracted PYR1-HAB1 complex ( Figure 7 ( B ) ) demonstrated inter-molecular correlations in similar locations , but with some variability in relative extent . In the ABA extracted complex , N-terminal regions helix α1 and loop Lα1β1 of PYR1 showed a slightly stronger correlation with HAB1 in comparison to the ABA-bound complex , although the level of correlation is slightly lower for the rest of the receptor . Other , minor differences are observed in the areas of loop Lα3β2 , the gate , latch , and loop Lβ7α5 . In the PYR1-ABA-HAB1 construct these areas show a more pronounced correlation with HAB1 , whereas in the case of PYR1-HAB1 complex the loop Lα3β2 has a weaker correlation around residue K63 , the gate has reduced correlation around residue G86 , correlations for the entire latch are decreased , and loop Lβ7α5 loses the correlations around residue P148 . Interestingly , the loop Lα1β1 shows stronger correlations in the PYR1-HAB1 complex than in PYR1-ABA-HAB1 complex , which may be related to disulfide bond formation with C30 as previously reported for residues R157 and C30 [9] . As it has been shown elsewhere , HAB1 residue W385 forms a hydrogen bond with a water molecule , which in turn binds to P88 and R116 of PYR1 , and with the ketone group of ABA [17] . It is important that for both constructs A and B of Figure 7 , correlations between HAB1 residues around W385 with Lα3β2 , Lβ7α5 and the gate are very strong , indicating that tryptophan is inserted in the gate-latch gap independently of the presence or absence of ABA . As it can be seen from close up of PYR1-HAB1 binding area presented on , both ABA-bound and ABA-free closed-lid constructs have a water molecule mediating the interactions of W385 with R116 . Other water molecules mediate different important interactions inside of the ligand cavity of both constructs , however the absence of ABA makes these interactions weaker . In Figure 8 , PYR1 main chain flexibility profiles are presented for both PYR1-ABA-HAB1 and PYR1-HAB1 complexes . Notable is the absence of the flexibility maximum around the gate ( which is one of areas of the phosphatase binding ) in distinction to the profiles of HAB1-free constructs shown in Figure 3 . In agreement with the correlation maps , the flexibility of the latch is somewhat higher in the ABA-free construct than in the ABA-bound construct , indicating than these areas are less constrained in the former case . However , overall the locations of maxima as well as minima of the two flexibility profiles are found at largely similar locations , except for latch residue L117 , where a flexibility maxima is observed in the ABA-free construct , but not in the ABA-bound construct . L117 swings outward in the absence of ABA . In conclusion , ABA-bound and ABA-extracted PYR1-HAB1 constructs demonstrate notable similarities in their dynamics , indicating that similar binding mechanisms are likely involved in basal and ligand-induced interactions and suggest that apo-PYR1 should be able to interact with PP2Cs if it is ever free of the homodimer . We also compared the normalized ECD flexibility profiles of HAB1-bound PYR1-ABA construct with the corresponding experimental crystallographic B-factors of the backbone [17] ( Figure 9 ) . The two profiles show a reasonable agreement , both exhibiting absence of the flexibility maximum in the area of the gate , as well as a relative decrease of the flexibility in other areas of the phosphatase binding such as the latch and loop Lβ7α5 , in comparison to the corresponding profiles of HAB-free monomeric PYR1 in Figure 4 . Beyond the region of N-terminus , the most significant difference between the normalized ECD flexibility profile and B-factors is observed in the area of helix α3 and loop Lα3β2 . In this area , the values of B-factors show a maximum that is absent in the ECD flexibility , which may be explained by the formation of bonds K59-ABA and H60-S322 observed in the MD simulation . MD simulations were also performed for both ABA-bound and ABA-free PYR1-HAB1 complexes , in which dissociated closed-lid PYR1 was shifted back against HAB1 for 15 Å , after which 40 ns MD simulations were performed as described in Methods . Snapshots from two independent MD simulations of ABA-bound PYR1 and HAB1 systems are shown respectively in Figures S4 and S5 . In one of the MD trajectories ( Figure S4 ) , ABA-bound PYR1 has shifted toward the phosphatase during equilibration; the first bond was formed between the gate and residue W385 of HAB1 at 4 ns , and subsequently other connections developed . After approximately 20 ns the entire bond network was recovered . Figure 7 ( C ) demonstrates the correlations at the binding area after docking has occurred in the first simulation , which bears a significant resemblance with the correlation map for unperturbed PYR1-ABA-HAB1 complex ( Figure 7 ( A ) ) . In another trajectory ( Figure S5 ) the shift of ABA-bound PYR1 toward HAB1 occurred over the first 8 ns after equilibration , a bond was formed between the gate and W385 followed by development of other connections which also stabilized the folds of HAB1 . After approximately 30 ns , PYR1 and HAB1 adopted a similar docking pose as in the initial crystallographic model ( PDB ID 3QN1 [17] ) . With this control experiment in hand , next , we performed similar MD simulations with ABA-extracted ( closed-lid ) dissociated PYR1 shifted against HAB1 . Interestingly , two such simulations again demonstrated binding of apo-PYR1 to HAB1 . In one of the simulations , the initially distanced ABA-free PYR1 and HAB1 developed some binding after only 1 ns , however the PYR1-HAB1 interaction remained flexible allowing for a slight rotation and formation of stable bonds after 28 ns . Subsequently , the recovered complex construct remained stable ( Figure S6 ) , although the inter-molecular correlations were weaker than in the crystallographic model ( Figure 7 ( D ) ) . In another simulation , after approximately 20 ns PYR1 and HAB1 have slightly rotated against each other and formed some contacts ( Figure S7 ) . In both simulations , the binding surface was somewhat different from that of the complex PYR1-ABA-HAB1 . The PYR1 main chain flexibility profiles for the partially recovered ABA bound and ABA-free PYR1-HAB1 complexes after the initial 15 Å shift are shown in Figure 8 . Segments for the ECD analysis were taken from the last 20 ns of the production run , when a recovery of bonds between shifted PYR1 and HAB1 was observed ( see Figures S4 and S6 ) . The corresponding flexibility profiles represent constructs with significantly decreased PYR1-HAB1 distances compared to the starting point of the run , approximately from 10 Å to 3 . 3 Å . For this reason , the levels of PYR1 main chain flexibility ( especially in loops Lα1β1 , Lβ2β3 , Lβ3β4 , Lβ6β7 and Lβ7α5 on Figure 8 ) in the recovered and unperturbed PYR1-HAB1 complexes are significantly lower than in PYR1 receptor alone ( see Figure 3 ) . Notably , this decrease of loop flexibility is observed in both ABA-bound and ABA-free recovered complexes , further supporting the possibility that an affinity may be possible between closed-loop PYR1 and HAB1 even in the absence of ABA . If one compares the four flexibility profiles in Figure 8 , it is clear that the flexibility is overall lower in both unperturbed ( non shifted ) simulations , in particular in the areas of the loop Lα3β2 , the gate and helix α5 . Indeed the regions around the gate , the loop Lβ7α5 , the C-terminal helix α5 ( residues D146-L171 ) , as well as residues K54-S66 all show a lower flexibility in unperturbed PYR1-ABA-HAB1 complex than in both recovered constructs . This is further emphasized by the higher flexibility in the recovered ABA-free complex , in particular around residues I62 , G86 ( near the gate ) , H115 , R116 ( near the latch ) , and E149 , S152 , W156 , A160 ( in Lβ7α5 and helix α5 ) . Interestingly the unperturbed PYR1-ABA-HAB1 complex does have higher flexibility around residues F20 ( loop Lα1β1 ) and E132 ( loop Lβ6β7 ) , but these are both relatively distant from the binding area . The flexibility of recovered PYR1-ABA-HAB1 complex is lower than that of recovered ABA-free PYR1-HAB1 complex by 35% around residue M158 in loop Lβ7α5 and by 42% around G86 in the gate . This indicates that a significant constraining of the loop Lβ7α5 as well as the gate occurs in the presence of ABA in the recovered system . Complementary to the ECD analysis , nearest inter-molecular neighbors identified by Accelrys VS with a 5 Å cutoff have also been compared for the initial PYR1-ABA-HAB1 complex and the recovered PYR1-ABA-HAB and PYR1-HAB complexes ( see Table S1 , blocks I , II , and III , respectively ) . Nearest inter-molecular neighbors for the PYR1-ABA-HAB1 construct with mutation H60P ( PDB ID 3ZVU , [16] ) , PYL2-ABA-HAB1 complex ( PDB ID 3KB3 [11] ) , PYL3-ABA-HAB1 complex ( PDB ID 4DS8 [38] ) , and ABA-free PYL10-HAB1 complex ( PDB ID 3RT0 , [15] ) are also listed in Table S1 ( blocks IV-VII , respectively ) . It can be seen that the binding interfaces of receptors in constructs 3QN1 , 3ZVU , 3KBS , 4D58 , and 3RTO are highly conserved and include residues of Lα3β2 , the gate , the latch , Lβ7α5 and helix α5 ( F61 , S85-P88 , R116 , N151 , D155 , F159 , T162 , L166 ) . The recovered PYR1-ABA-HAB1 complex exhibits largely similar binding areas , in agreement with the described ECD analysis . In the recovered ABA-free PYR1-HAB1 complex , an extended Lβ3α2 area , some residues of the gate and helix α5 also developed bonds with the phosphatase . Further to this , PYR1 regions around the loop Lβ6β7 ( E132-R134 ) , C-terminal region of α5 ( R180 ) , and the N-terminal region of Lβ3α2 ( P55-K59 , not shown in Table S1 ) developed bonds with HAB1 only in the recovered ABA-free PYR1-HAB1 complex and not in the other constructs . It can also be seen that the regions of the phosphatase involved in the binding of the recovered ABA-free PYR1-HAB1 complex are somewhat different from those in the initial PYR1-ABA-HAB1 construct ( 3QN1 ) as well as in the other PYR and PYL constructs listed in Table S1 . Overall these results highlight some of the dynamical aspects of the PYR1 structure that are important in mediating PYR1-HAB1 complex formation; highlighting some ABA-dependent differences in the roles of loop Lα3β2 , the gate , the latch , loop Lβ7α5 , and helix α5 . These findings are consistent with previous crystallographic and mutagenic reports , and also support the possibility of the formation of functionally relevant ABA-free PYR1-PP2C interactions . On the other hand , MD simulations demonstrate that the absence of ABA decreases the aptitude of PYR1 to quickly develop proper contacts with the phosphatase . This supports the possibility of some low level basal PYR1 activity in association with these interaction kinetics . This computationally predicted dynamical data begs the question of whether binding of PYR1 to PP2Cs occurs only in the presence of ABA or whether select concentrations of receptors and phosphatases might enable an ABA-free PYR1-HAB1 interaction . While there is no substantial evidence for such an apo-PYR1-PP2C interaction in the literature to date , one fairly systematic analysis of the basal activity of the ABA receptors against four different PP2Cs did reveal up to ∼45% inhibition of HAB1 activity by apo-PYR1 at an RCAR∶PP2C ratio of 100∶1 at 0 . 27 µM HAB1 [15] . However a different study [39] working with a 4∶1 ratio was unable to detect any inhibition of PP2CA by PYR1 and only a weak inhibition by PYL8 was observed . Further detailed evaluation by in vitro experiments in our own lab , in which freshly purified recombinant PYR1 was titrated into the phosphatase ABI2 across a range of protein ratios showed a maximal basal effect at a 4∶1 receptor∶phosphatase ratio ( up to 80% inhibition of ABI2 activity ) at a constant concentration of 0 . 5 µM ABI2 ( Figure S8 ) . It is interesting that this interaction seems to be conditional , occurring over a relatively narrow range of RCAR∶PP2C ratios starting at around equimolar and peaking at 4∶1 and then decreasing at higher ratios . The decrease at higher ratios would be consistent with increased homo-dimerization of PYR1 , sequestering it away from the PP2C . The potency of the apo-PYR1-ABI2 interaction reported here , while at odds with those published previously [15] , [16] , may relate to subtle differences in the actual protein concentrations tested and the identity of the PP2C . Hao et al . , [15] only see potency of apo-PYR1 against HAB1 , and not ABI1 , HAB2 or PP2CA ( they do not provide data for ABI2 at all ) . As well the freshness of the protein preparation may have an impact , as freeze storage of at least one PYL has been shown to selectively abolish basal signaling functionality of the receptor , without affecting its ABA-induced activity ( Figure S9 ) . Overall , these relative activities for the basal PYR1-ABI2 versus PYR1-ABA-ABI2 interactions ( compared directly at a 1∶1 RCAR∶PP2C ratio ( Figure S8 ) as well as the apo-PYR1-HAB1 interaction described previously [15] , are consistent with the dynamic variations described above with increased correlations and more stable profiles detected for the PYR1-ABA-HAB1 complex in the areas of loop Lα3β2 , the gate , latch , and loop Lβ7α5 . That the low or complete lack of basal activity reported for dimer-forming receptors including PYR1/PYLs 1–4 , may result from a ‘competitive interaction’ process between homo-dimer complexes and receptor/phosphatase complexes has been put forward in several publications [7] , [15] , [16] . The results reported in the previous section support the possibility that PYR1-HAB1 binding is possible in the absence of ABA , lending further support to the ‘competitive interaction’ mechanism for PYR1 . Vice versa , our observation that repeated freezing abolishes the basal activity of a ‘monomeric’ receptor ( PYL5; see Figure S9 ) , but not its ABA-inducible activity , suggests that under certain conditions , it may even be possible for these monomeric receptors to form homo-dimers ( oligomers ) . To date , published experimental studies have shown that loops Lα3β2 , as well as the gate , the latch and the loop Lβ7α5 are involved in determining the outcome of the competition [9] , [16] , [17] . In particular the importance of loop Lα3β2 was demonstrated in experiments by Dupeux et al [16] , which revealed that residue H60 may determine the oligomeric state of PYR/PYL family members . In turn , conformations of the gate , the latch and the loop Lβ7α5 ( often denoted as Pro cap , Leu lock and partially “Recoil motif” subunits ) have been shown to be somewhat different for ABA-bound and ABA-free PYR1 asymmetric dimers [9] . A subtle difference of conformations around residue S85 and Lβ7α5 loop between ABA-bound dimer and PYR1-ABA-HAB1 also indicate that these regions indeed play a role in the binding of PYR1 and HAB1 [17] . However , the dynamical contributions of these regions to determining the selectivity of interactions remain to be explored by molecular simulations . Thus we extended our study to address PYR1 dimers in comparison with the PYR1-HAB1 complexes in the presence and absence of ABA . Studies have shown that PYR1 forms homodimers in the absence of ABA [15] , [16] , or possibly with ABA occupying one binding site between the two dimer partners [9] . On the assumption that the presence of two ABA molecules in a dimer leads to monomerization , simulations on a PYR1 dimer construct containing two ABA ligands , which is denoted as the 2ABA-bound dimer , were initiated . This construct has been prepared from PDB structure 3NJO , where the pyrabactin ( PYV ) and P2M ligands were replaced with ABA molecules . The structures of symmetric ligand-free dimers , and occasionally , the asymmetric 1ABA-bound dimer have also been used for comparison ( see Table 1 and Methods ) . Because the available PDB crystallographic structure of PYR1 dimer contains PYV/P2M ligands , we also investigated the binding of these ligands along with ABA . The structure of a 2ABA-bound dimer after equilibrations in water is shown in Figure 10 . The identified interactions are in agreement with published structural data [9] , [10] . As previously reported , the regions of PYR1 involved in binding to HAB1 ( Figure 6 ) and in the dimer ( Figure 10 ) overlap significantly [17] . The MD simulations for the dimers have been carried out for 30 ns after a 10 ns NPT equilibration . Figure S10 shows the main chain flexibility profiles for the PYV/P2M- and 2ABA- bound PYR1 dimers , as well as for ABA-free apo-dimer , whereas Figure 11 compares the flexibilities of PYR1 from ligand-free and 2ABA-bound PYR1 dimers with that from the complex PYR1-ABA-HAB1 . These comparisons demonstrate that in most regions , flexibility and its standard deviation are higher in ABA-bound PYR1 dimer ( both 2 ABA and 1 ABA forms ) than in any other complex . In particular , the flexibility in the latch region ( H115 ) of 2ABA-bound and ABA-free dimers is higher than in PYR1-ABA-HAB1 complex by 67% and 27% , respectively . We attribute this difference to the effect of bonds L117-W385 and R116-Q386 in the PYR1-ABA-HAB1 construct . In ABA-free dimer , residues H115 in both chains are asymmetrically involved in intra-receptor binding with D154 and A89 each , whereas in ABA-bound dimers residue P88 interacts only indirectly with ABA , making the gate and latch areas more flexible and also destabilizing helix α4 . Interestingly , the increase of flexibility around residue H115 is more pronounced in the 1ABA-bound-dimer than in either the 2ABA-bound dimer or the pyrabactin-bound dimer ( Figure S10 ) . In the PYV-bound dimer , H115 can interact with A89 because of the smaller size of PYV size and residue 88 tends to interact with the latch rather than with the ligand , stabilizing helix α4 . As it also can be seen from Figure S10 , binding of PYV/P2M or two ABA ligands decreases the flexibility of helix α5 in comparison with ligand free dimers . Another notable feature is observed in the loop Lβ7α5 and around residue A89 . In the first region , dimerized PYR1 develops bonds at residue E153 , whereas in the PYR1-HAB1 complex , the binding occurs at residues N151-Q384 and P148-W385 which causes a slight shift in the flexibility maximum in the region of this loop . These differences between dimerized and phosphatase-bound PYR1 constructs agree with the recent analysis of corresponding crystallographic data [17] . Finally , for all dimer models considered , a relatively high flexibility for loops Lα1β1 and Lβ2β3 is observed since these loops are dynamically uncoupled from the rest of the construct . Main chain correlation maps for the dimer constructs were subsequently analyzed . Intra-receptor correlations for atoms in ABA-free dimer and PYR1-HAB1 as well as 2ABA-bound dimer and PYR1-ABA-HAB1 complex are shown in Figure 12 . It is clear from comparison of Figure 12 panels ( A ) and ( B ) that intra-receptor correlations are overall weaker in the ABA-free PYR1 dimer compared to the ABA-free PYR1-HAB1 complex . However , in 2ABA-bound dimer intra-receptor correlations for the latch region and loop Lβ4β5 are even weaker still , supporting a model in which 2ABA-binding destabilizes apo-homodimer complexes . In contrast , the presence of ABA leads to stronger correlations in the ABA-bound PYR1-HAB1 ( Figure 12 D ) complex than in either dimer or the apo-PYR1-HAB1 complex . Especially evident is the relative lack of intra-molecular correlations in the area of the C-terminal helix α5 in both apo- and holo- dimers ( Figure 12 A and D ) . As well , helix α5 coupling to residues 30–62 , gate , latch and β7 is approximately three times weaker in the dimers than in the PYR1-ABA-HAB1 complex . Inter-molecular correlation maps of the binding areas of ligand-free and 2ABA-bound dimers shows a similar inter-correlation network in the two constructs ( Figure 13 ) . One difference is that the residues from the latch region in chain A of the 2ABA-bound dimer show less correlation with chain B generally . Also , the coupling of gate residues is slightly weaker with ABA in the dimer . Overall , the inter-molecular correlations are slightly less symmetrical for the 2ABA-bound dimer , arising from the observation that correlation of the chain A gate with the chain B helix α4 is not mirrored by a similar correlation between chain B gate and chain A helix α4 in agreement with published X-ray scattering experiments of assymetric dimeric units [9] . Comparison of Figures 13 and 7 ( A ) confirms that similar regions of PYR1 are involved in HAB1 and dimer interfaces . In particular , residues K54-S66 , R79-W93 , and D146-L171 are involved in both dimer association and HAB1 binding . However , in most of these areas PYR1 residues are less constrained in the dimer than in the PYR1-HAB1 complex . Thus , a relatively weak coupling of chains A and B of the dimer is observed in the region of residue H60 , the gate , N-terminal of helix α5 , and to a lesser extent in the latch and C30 region . These differences originate from the interactions of the latch region and T142-N151 region residues . While the latch region is strongly correlated with several large parts of the phosphatase , pronouncedly weaker correlations are observed in the dimer , particularly around residue 111 . A similar trend of decoupling is observed at residue 145 of the dimer . In an effort to elucidate details of the molecular mechanisms mediating PP2C inhibition by ABA receptors , we report extensive molecular dynamics simulations of apo and holo pyrabactin receptor PYR1 in complex with HAB1 as well as in dimeric form . We also report our comparative analysis of the dynamical stability of these complexes by novel ECD method , which we have validated against crystallographic B-factors . In agreement with recent experimental findings [8]–[16] , our MD simulations and the ECD analysis indicate that ABA-bound PYR1 should efficiently bind to HAB1 . In particular , the loop Lα3β2 , the gate , latch , loop Lβ7α5 , and helix α5 have been found to develop stronger dynamic correlations with HAB1 in presence of ABA in comparison to ABA-free constructs . However , ABA-bound and ABA-extracted PYR1-HAB1 constructs demonstrated notable similarities in their dynamics , suggesting that apo-PYR1 should be able to make a substantial interaction with PP2Cs . This possibility was validated by in vitro data that demonstrate a conditional functional interaction between apo-PYR1 and ABI2 in our hands . In the context of competing interactions , our dynamical analysis indicates that although similar regions of PYR1 are involved in dimer association and HAB1 binding both ABA-bound and ABA-free PYR1 in complex with HAB1 exhibit a lower flexibility , higher intra-molecular structural stability , and stronger inter-molecular dynamic correlation , in comparison with either holo- or apo- PYR1 in dimeric form . This may be interpreted as dimeric PYR1 being under less steric constraint in comparison with PYR1-HAB1 complex . Furthermore , comparison of 2ABA-bound and ABA-free dimers reveals a loss of intra-receptor correlations , in particular in the areas of the latch and loop Lβ4β5 , upon ABA binding . Inter-chain correlations in the area of the latch and the gate are also somewhat weakened in the presence of ABA . Together these results are consistent with ABA having an opposite effect on PYR1-HAB1 and PYR1-PYR1 complexes , constraining the former and destabilizing the latter , as expected . They also suggest that ABA-free PYR1 can bind to the phosphatase , and that such binding would be in competition with PYR1 dimerization , particularly in the absence of ABA . These findings , validated in vitro , suggest that the model of receptor regulation by ‘competing interactions’ may be more complex and at the same time more broadly applicable to all PYR1/PYL type ABA receptors both in the presence and absence of ABA . Finally , these findings raise the question of whether the dynamics of receptor-phosphatase interactions might be preferred over that of homo-dimer interaction regardless of the presence or absence of ABA in the PYR1 binding pocket , and whether the ‘competing interactions’ mechanisms could , to a significant extent , be regulated by kinetic factors such as the interaction reaction pathway or the availability of protein [39] .
Crystallographic coordinates of the pyrabactin receptor PYR1 , as well as the complexes PYR1-ABA-HAB1 and PYR1-pyrabactin-bound dimer ( resolution 1 . 70 Å , 1 . 80 Å and 2 . 47 Å respectively ) were taken from the Protein Data Bank [36] , entries 3K3K , 3QN1 , and 3NJO [9] , [10] , [17] , see Table 1 . Homology to ABA insensitive 1 ( HAB1 ) phosphatase model was chosen based on the availability of the 3D structure in complex with PYR1 . Initially , all water molecules were removed from the crystal structures . All ligands extractions/insertions were created in silico with Accelrys Discovery Studio software [40] . In structure 3QN1 , 26 missing residues , G222-L231 , D271-R282 , and P462-E465 and missing atoms in residues 214 , 281 , 233 , 406 , 422 , 468 , and 504 were reconstructed using a replacement structure 3RT0 . pdb , chain A of apo-PYL10-HAB1 complex [15] . Missing residue P229 was built with Accelrys DS . For the ABA molecule , Accelrys DS was used to add the hydrogens , and acpype [41] and mktop [42] scripts were employed to evaluate the charge distribution and generate topology files . All constructed regions were optimized . As a model of monomeric ABA-bound closed lid PYR1 construct , chain B of structure PDB ID 3K3K was used , whereas chain A of 3K3K was used to model open lid ABA-free PYR1 construct . To simulate ABA-free closed lid PYR1 construct , ABA was extracted from chain B of 3K3K ( see also Figures S1 and S2 ) . All constructs were minimized in vacuo , and solvated in water with counterions afterwards . The model of ABA-bound PYR1 in complex with HAB1 employed structure PDB ID 3QN1 , and the ABA-free PYR1-HAB1 structure was obtained by extraction of ABA from 3QN1 and further minimization and equilibration . The same crystallographic model 3QN1 was used to construct ABA-bound and ABA-extracted PYR1-HAB1 systems , in which PYR1 was initially shifted against HAB1 ( see Figures S4 , S5 , S6 , S7 ) . To prepare a separated ABA-bound construct , the centers of mass of PYR1 and HAB1 were oriented along the 0Z axis , and then the coordinates of ABA-bound PYR1 receptor were shifted against the phosphatase for 15 Å along the 0Z axis using Accelrys DS [40] . After in vacuo minimization of the ABA-extracted PYR1-HAB1 complex , similar to described above centers of mass alignment has been made , then the receptor was shifted against HAB1 for 15 Å along the 0Z axis . The PYR1 dimer constructs were built from PDB ID 1NJO structure . In this model , subunits of the P88S mutant dimer contain synthetic ligands C16H13BrN2O2S ( pyrabactin , PYV ) and C16H14N2O2S ( P2M ) , and the mutation P88S is introduced to improve binding of these ligands . To construct PYR1 models containing one and two ABA ligands , PYV/P2M ligands were extracted from 1NJO and replaced with one or two ABA molecules respectively , after which reverse mutation S88P has been made using Accelrys DS , and minimization was performed . In the paper , dimers containing one and two ABA molecules are denoted as 1ABA-bound and 2ABA-bound dimers , respectively . In the model of ligand-free dimer , PYV/P2M molecules were extracted from structure 3NJO , residue S88 replaced with P88 , and minimization performed . For simulations involving PYV-bound PYR1 monomers and PYV/P2M bound dimers , residue S88 was retained , structure minimized and solvated . After in-vacuo minimization , each system was solvated in a triclinic box with walls located at distances ≥15 Å from the protein . Simple Point Charge ( SPC ) water molecules were employed , and Na+ or Cl− counterions were added to make system net charge equal to zero . Minimizations , equilibrations and production MD simulations were carried out using Gromacs v4 . 0 . 7 [43] and AMBER v11 packages [44] with OPLS and AMBER03 force fields , respectively . The trajectories generated by Gromacs were used to analyze structural changes as well as for ECD analysis . In vacuo minimization of starting models described in the previous section comprised 10000 steps of steepest descent minimization . After the systems were solvated , solvent minimization was made using 500 steps of a steepest descent algorithm with strong positional restraints on all heavy protein atoms to prevent distortion of protein structure by non-equilibrated solvent . Next , seven steps of short steepest descent minimization were performed on each solvated system with decreasing position restraints on non-hydrogen protein atoms ( Kposre = 1×105 , 1×104 , 1000 , 100 , 10 and 0 kJ mol−1 nm−2 ) followed by system heating . The temperature of proteins and solvent was maintained at desired level ( 310 K in most cases ) by coupling the systems with Berendsen thermostats [45] . Seven NVT-like MD equilibration steps with decreasing non-hydrogen protein position restraints ( Kposre = 1×105 , 1×104 , 1000 , 100 , 10 and 0 kJ mol−1 nm−2 ) were then made , the last one with no restraints and stronger bath coupling . The last equilibration step and production simulations were conducted at the desired temperature of 310 K ( unless indicated otherwise ) and pressure at 1 atm with isotropic pressure coupling ( NPT ensemble ) , bond length restrained with the LINCS algorithm with a fourth order of expansion . The short-range electrostatic and van der Waals interactions cutoff radii were equal to 14 Å each . Long-range electrostatic interactions were treated with particle-mesh Ewald ( PME ) summation with grid spacing 0 . 135 nm for the fast Fourier transform and cubic interpolation . In order to validate our model building protocol , we have computed the root-mean-square deviations ( RMSD ) between main-chain atoms of asymmetrical one-ABA-bound dimer construct ( PDB ID 3K3K [9] ) and a similar model structure prepared from modified 3NJO [10] construct where S88P mutations were introduced in both chains , pyrabactin and P2M extracted , and one ABA ligand inserted . After minimization of the modified structure , 6 heating steps of 50 ps each were performed followed by a 450 ps NVT equilibration , and subsequently by a 200 ps NPT equilibration . The coordinates of the modified 3NJO construct were subsequently aligned with the reference 3K3K construct , and the corresponding RMSDs were computed using VMD software [46] . For the modified 3NJO structure , core domains ( 33–37 , 53–57 , 59–65 , 80–84 , 90–94 , 105–108 , 110–112 , 121–127 , 134–141 , 158–162 , 166–168 , 170–172 ) were employed to obtain the backbone ( C , Ca , N ) RMS deviations . Both the initial crystallographic 3K3K structure and the solvated and equilibrated 3K3K construct were employed as references . When using the original crystallographic structure 3K3K as a reference , the RMSD after NVT equilibration were between 0 . 32 and 1 . 32 Å . Figure S11 illustrates the evolution of the RMSD during last stages of NVT equilibration and also during NPT equilibration . Over the first 200 ps long process of NVT equilibration seen in the figure ( 400 ps–600 ps ) , the constraints applied kept the RMSD stable around 1 . 34 Å , whereas on the interval from 600 ps to 750 ps the constraints were released and backbone atoms have readjusted resulting in RMSD levels fluctuating between approximately 1 . 05 Å and 1 . 45 Å . Over the following 200 ps NPT equilibration was performed , during which RMSD first increased from approximately 1 Å to an average of about 1 . 25 Å and then stabilized . Figure S12 shows the secondary structure alignment of the modified 3NJO construct and the crystallographic 3K3K model after completion of the NPT equilibration right before the production run . When a solvated and equilibrated 3K3K model was used as a reference , the corresponding RMSD were between 0 . 26 and 1 . 23 Å . The production MD simulation runs were performed from 20 ns to 50 ns depending on the system with 1 fs time steps , and snapshots saved every 20 fs in order to analyze the essential collective dynamics . The ABA-bound PYR1 closed lid system was simulated for 40 ns at 300 K and for 30 ns at 325K , the ABA-free PYR1 closed lid systems were simulated for 20 ns at 281 K , for 50 ns at 300 K , for 20 ns at 310 K , and for 30 ns at 325K; and ABA-free open lid PYR1 construct was simulated for 30 ns at 281K and for 30 ns at 300 K , using the Gromacs MD simulation package . The ABA-bound and ABA-free PYR1-HAB1 complexes , and PYR1 dimer complexes were simulated for 30 ns at temperature of 310 K . For ABA-bound and ABA-free complexes with PYR1 shifted against HAH1 , two independent molecular dynamics ( MD ) simulations of 40 ns each have been carried out . The ECD method relies upon a recently developed statistical-mechanical framework [31]–[34] , according to which a macromolecule can be described by a set of generalized Langevin equations ( GLE ) with essential collective coordinates , which can be deduced by applying PCA on MD trajectories . The latter procedure provides a set of principal eigenvectors of the covariance matrix , . Here represent the direction cosines of the eigenvectors , where is the number of atoms in the system , and is the number of eigenvectors which sample a sufficient percentage of the total displacement , which are often referred to as essential collective coordinates . Usually , 10–30 essential coordinates are sufficient to sample approximately 90% of the displacement for a typical MD trajectory of a protein . In the ECD method , an all-atom projected image of the protein is constructed in the dimensional space of essential collective coordinates such that the position of each atom is characterized by , where . The theory shows that such an image represents the degree of dynamic correlation ( coupling ) between the protein's atoms: points ( images of atoms ) that are located close to each other correspond to atoms whose motions are strongly correlated regardless of their proximity in secondary or tertiary structure of the protein , and more distant points correspond to a relatively independent motion [31] , [33] . A suite of simple structural descriptors , such as the main chain flexibility and domains of correlated motion , have been derived within the ECD framework and successfully employed to analyze dynamics of proteins [32] , [35] , [47] , [34] . It has been both proven theoretically [33] and confirmed by comparing numerical predictions with NMR experiments representing microsecond time regimes [31] , [34] . In this work , the dynamics of PYR1 constructs were characterized primarily employing ECD derived correlation maps and flexibility profiles . The ECD correlation maps are distances between images of atoms in the dimensional space of essential collective coordinates . These distances are dimensionless quantities represented by [34] ( 1 ) with lower values of representing stronger correlations . In this work , Equation 1 has been employed to visualize correlations between atoms both within PYR1 molecules and across molecules in PYR1 dimers as well as complexes with HAB1 atoms in order to examine the corresponding intermolecular and intramolecular dynamics . While ECD derived flexibility profiles allows characterizing the flexibility with atomic-level precision , a per-residue flexibility assessment is sufficient in many cases [32] , [34] . Here , the flexibility for atoms in the main chain of the various PYR1 constructs was analyzed . The ECD flexibility descriptor for a atom in residue , ( 2 ) defined by the distance in the dimensional space of essential collective coordinates between the image point representing the atom and the centroid calculated over the images of all atoms , ( 3 ) In Equation 3 , is the total number of atoms in the molecule . By definition , the ECD flexibility descriptor represents the level of dynamic coupling of the motion of individual atoms with the entire molecule , which in turn is represented by the centroid of all atoms in the space of essential collective coordinates . For each construct considered , we employed the ECD analysis with , on 100 segments , each of 20 fs , from the last 20 ns of MD trajectories . Finally , the ECD framework also allows identifying dynamic domains of correlated motion , which represent relatively rigid parts of the protein composed of atoms moving coherently . Such domains can be identified through a simple nearest-neighbor clustering of the protein's image in the space of essential collective coordinates , as described in detail elsewhere [31] , [32] , [35] , [34] . When employing the nearest-neighbor clustering technique to identify the dynamic domains , an interdomain distance must be selected which represents the minimum degree of correlation for two atoms to belong to the same domain . In this paper , has been adopted as an optimum value , which maximizes both the domains number and the difference of total number of atoms in all domains and number of atoms in the largest domain , as well as includes reasonable ( more than 10% ) average number of atoms in the largest domain . A further discussion of the choice of the interdomain distance can be found elsewhere [31] , [32] , [34] . An extensive series of test MD simulations for PYR1 , both closed lid and open lid constructs , as well as ABA and pyrabactin bound and ABA-free , at various temperatures for 20 to 50 ns was carried out . | Protein pyrabactin resistance 1 ( PYR1 ) belongs to a group of PYR1-like ( PYL ) proteins that regulate plant development and responses to conditions of drought and salinity . Recent studies have reported characterization of their molecular structures as well as elucidation of important aspects of their function; highlighting their roles as receptors for the stress responsive phytohormone , abscisic acid ( ABA ) . However details of the molecular mechanisms regulating their receptor signalling remain enigmatic . In this work , we use molecular dynamics simulations complemented by a sophisticated statistical-mechanical analysis to investigate structural and dynamical properties of PYR1 protein and how its interaction with ABA modifies receptor-protein complex formation . Our results provide detailed insight into how the PYR1-mediated inactivation of its downstream phosphatase target is regulated by homodimer formation and yield new hypotheses , supported by in vitro experiments , for further investigation . Ultimately , this knowledge provides insight into how plants respond to stress , with potential applications in the development of crops with improved growth characteristics and higher stress tolerance . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Methods"
] | [
"biomacromolecule-ligand",
"interactions",
"physics",
"statistical",
"mechanics",
"protein",
"folding",
"macromolecular",
"assemblies",
"biophysics",
"theory",
"chemistry",
"biology",
"computational",
"chemistry",
"biophysics",
"simulations",
"biophysics"
] | 2013 | Molecular Mechanisms in the Activation of Abscisic Acid Receptor PYR1 |
ARFI elastrography has been used as a noninvasive method to assess the severity of liver fibrosis in viral hepatitis , although with few studies in schistosomiasis mansoni . We aimed to evaluate the performance of point shear wave elastography ( pSWE ) for predicting significant periportal fibrosis ( PPF ) in schistosomotic patients and to determine its best cutoff point . This cross-sectional study included 358 adult schistosomotic patients subjected to US and pSWE on the right lobe . Two hundred two patients ( 62 . 0% ) were women , with a median age of 54 ( ranging 18–92 ) years . The pSWE measurements were compared to the US patterns of PPF , as gold standard , according to the Niamey classification . The performance of pSWE was calculated as the area under the ROC curve ( AUC ) . Patients were further classified into two groups: 86 patients with mild PPF and 272 patients with significant PPF . The median pSWE of the significant fibrosis group was higher ( 1 . 40 m/s ) than that of mild fibrosis group ( 1 . 14 m/s , p<0 . 001 ) . AUC was 0 . 719 with ≤1 . 11 m/s as the best cutoff value for excluding significant PPF . Sensitivity and negative predictive values were 80 . 5% and 40 . 5% , respectively . Whereas , for confirming significant PPF , the best cutoff value was >1 . 39 m/s , with specificity of 86 . 1% and positive predictive value of 92 . 0% . pSWE was able to differentiate significant from mild PPF , with better performance to predict significant PPF .
Schistosomiasis is considered a public health problem in many parts of the world and one of the most frequent causes of liver fibrosis worldwide [1] . In Brazil , schistosomiasis mansoni infection affects 1 . 5 millions of people , particularly in the Northeastern region [2] . The state of Pernambuco is considered an endemic area with 4207 cases in 2016 , and a mortality average of 173 deaths per year between 2005 and 2015 [3] . Hepatic schistosomiasis represents the best known form of chronic disease with a wide range of clinical manifestations [4] . Accordingly , periportal fibrosis ( PPF ) or Symmers fibrosis is a schistosoma-induced liver disease , and the progression of PPF from hepatointestinal ( HIS ) to hepatosplenic ( HSS ) form is associated with the development of portal hypertension and its complications , such as esophageal varices and hypertensive gastropathy . Moreover , the patients with significant PPF consequently presents more portal hypertension and more risk of upper gastrointestinal bleeding [5] . The gold standard method in evaluating schistosomal hepatic fibrosis was wedge liver biopsy , a safe procedure performed during abdominal surgery , but not justified in non-surgical patients . An alternative procedure , percutaneous liver biopsy , has low sensitivity because it retrieves small and fragmented samples with few portal tracts [6] . Furthermore , liver biopsy is invasive , usually unsuitable for long-term follow-up and not applicable in the field , indicating that there is no gold standard for schistosomal portal fibrosis available [7] . Therefore , imaging techniques , such as ultrasound ( US ) scan , have been proven valuable for liver morbidity assessment and its complications in endemic areas [1 , 8 , 9] . Hence , US is considered a complementary tool often used to assist the diagnosis of the S . mansoni infection , mainly on the study of the liver damage caused by the chronic infection [10] . US also assess liver , spleen and portal vein size , and it is important to exclude or confirm other intra-abdominal diseases [11] . For the assessment of PPF , and consequently schistosomiasis-related morbidity , US is one of the most important means for this purpose , especially with the formulation of a standardized WHO Niamey-Belo Horizonte protocol . [1 , 12] . The WHO Niamey-Belo Horizonte protocol shows practical usefulness it is considered satisfactory in terms of reproducibility , assessment of evolution of pathology , and comparability between different settings , even though physician expertise is required [7 , 13] . Nevertheless , over the last few years , Niamey-protocol has been increasingly used in the quantification of PPF [7] Recently , elastography methods , such as transient elastography ( TE ) and point shear wave elastography ( pSWE ) using acoustic radiation force imaging ( ARFI ) , have been applied to assess liver fibrosis in some diseases , like chronic hepatitis C using US grading as gold standard [14–17] . These methods have also been considered promising for liver fibrosis assessment in patients with hepatitis B , non-alcoholic fatty liver disease , and post-transplant patients [15 , 18 , 19] . In addition , TE was recently evaluated to assess PPF induced by S . mansoni using US grading as gold standard [20 , 21] . As a friendly operator method , pSWE could be used as a complementary approach to assess PPF in schistosomotic patients . However , there is no systematic report of the literature on the use of pSWE for evaluation of PPF due to schistosomiasis mansoni . Therefore , we aimed to evaluate the performance of pSWE as a tool for predicting significant PPF in patients with schistosomiasis mansoni and to determine its best cutoff point , using US grading as gold standard .
This cross-sectional study was conducted on adult patients with schistosomiasis mansoni , evaluated in the Gastroenterology outpatient clinic , Hospital das Clínicas of Universidade Federal de Pernambuco ( HC-UFPE ) , from 2016 to 2017 . Three hundred and ninety-six patients were eligible during the recruitment period . A total of 38 patients were excluded due to non-reliable pSWE results , characterized by an interquartile range of more than 30% of the median elasticity and a success rate of less than 60% . Among the 358 patients included , 222 ( 62 . 0% ) were women , with a median age of 54 ( ranging 18–92 ) years . The diagnosis of schistosomiasis mansoni infection was based on clinical history of contact with fresh water sources from endemic areas , reports of prior treatment with praziquantel , along with US findings of PPF . Patients were classified into two groups according to the PPF patterns using Niamey’s classification: 86 patients with mild PPF ( pattern C ) and 272 patients with significant PPF ( patterns D , E and F ) . Patients with chronic viral hepatitis , fat liver disease , history of alcohol abuse , and other causes of liver disease , such as autoimmune hepatitis , drug-induced liver injury and haemochromatosis , were not included in this study . All included patients were subjected to ultrasonography after to overnight fasting of about 8 hours , and were examined by the same experienced operator applying the Niamey classification , using a Siemens Acuson S2000 ultrasound system ( Siemens Medical Solutions , Mountain View , CA , USA ) with a 6C1 MHz transducer for both US and pSWE . Patterns of periportal fibrosis was registered according to the Niamey classification: pattern C ( peripheral ) , D ( central ) , E ( advanced ) and F ( very advanced ) [12 , 22] . After the US scan , the patients were submitted to pSWE , using the same machine . pSWE and measurements were acquired using Virtual Touch Quantification ( VTQ ) under standardized conditions on the right lobe ( RL ) of the liver , with the patient lying in a dorsal decubitus position , and the probe aligned along the intercostal space . The region of interest ( ROI ) was positioned into the hepatic parenchyma avoiding fibrosis tracts . The detection pulses measured the shear wave velocity , which was considered to directly relate to tissue stiffness . For each patient , 10 valid shear wave speed ( SWS ) measurements were performed on the RL , and the results were expressed as the median value of the total measurements in meters per second ( m/s ) . Reliable results were defined as those with an interquartile range ( IQR ) of less than 30% of the median value and a success rate of at least 60% [23] . This study was approved by the Ethics Committee on Research Involving Human Subjects of the Health Sciences Center–Universidade Federal de Pernambuco ( CCS-UFPE ) , Recife , Brazil ( Approval no . 1 . 782 . 771/2016 ) and all patients were adults and signed terms of informed consent . The statistical analysis was carried out using the GraphPad Prism 6 software ( GraphPad Software , Inc . , La Jolla , CA ) and MedCalc 15 . 2 software ( MedCalc Software , Mariakerke , Belgium ) . A descriptive analysis including frequency , distribution , median and interquartile range was performed in both groups , and the D'Agostino & Pearson test was used to verify the normality of data distribution . A univariate analysis with χ2-test was used to compare categorical variables and Mann-Whitney test was used to compare two independent groups of continuous variables . The performance of pSWE was assessed with receiver operating characteristic ( ROC ) curves . Optimal cut-off values for predicting significant PPF with best sensitivity and specificity were determined and positive ( PPV ) and negative predictive values ( NPV ) were calculated for these cut-off values . The area under the curve ( AUC ) was used to represent the accuracy of the predictions . A p value less than 0 . 05 was considered to be statistically significant .
The main characteristics of the patients included in the study are summarized in Table 1 . Measurements of SWS liver stiffness on the RL ranged from 0 . 66 to 4 . 27 m/s . The median values of liver stiffness assessed by pSWE , according to the PPF patterns , were: mild/pattern C of 1 . 14 m/s ( IQR: 0 . 97–1 . 32 ) ; moderate/pattern D of 1 . 33 m/s ( IQR: 1 . 03–1 . 69 ) ; advanced/pattern E of 1 . 40 m/s ( IQR: 1 . 19–1 . 90 ) ; very advanced/pattern F of 1 . 53 m/s ( IQR: 1 . 29–1 . 88 ) . Some of the median liver stiffness values for PPF patterns showed substantial differences when compared: C vs D ( p = 0 . 01 ) , C vs E ( p < 0 . 001 ) , C vs F ( p < 0 . 001 ) , and D vs F ( p = 0 . 014 ) . Among significant PPF patients , a higher medium age was identified ( 56 years , IQR: 45–65 ) and 64% were women . Comparing the clinical forms of schistosomiasis mansoni , higher frequencies of severe forms ( hepatosplenic ) were observed in significant PPF patients when compared with those of mild PPF , that showed higher frequencies of hepatointestinal forms , as shown in Table 1 . Results of liver stiffness measurement , in median values , by pSWE elastography revealed a substantial difference between patients with mild ( 1 . 14 m/s; IQR: 0 . 97–1 . 32 ) and significant PPF ( 1 . 40 m/s; IQR: 1 . 17–1 . 88 ) , with a p value < 0 . 001 , as shown in Table 1 . The best cutoff value of liver stiffness measurement for excluding significant PPF was equal or less than 1 . 11 m/s , with AUC of 0 . 719 and a sensitivity , NPV , and an overall accuracy of 80 . 5% , 40 . 5% and 71 . 2% respectively . Whereas , for confirming significant PPF the best cutoff value was above 1 . 39 m/s , with AUC of 0 . 719 and a specificity , PPV , and an overall accuracy of 86 . 1% , 92 . 0% and 59 . 2% respectively , as shown in Table 2 and Fig 1 . Applying the cut-off values , 24 . 9% patients showed SWS values below or equal the lower cut-off point ( ≤ 1 . 11 , n = 89 ) and among these patients , 40 . 5% were correctly classified , using US as gold standard . For 150 patients with SWS values above the upper cut-off point ( > 1 . 39 ) , 92 . 0% were correctly classified . In addition , 119 patients ( 33 . 2% ) had indeterminate test values ( between the lower and upper cutoff point: > 1 . 11 and ≤ 1 . 39 ) and among these , 68 . 1% had significant PPF and 31 . 9% had mild PPF .
According to liver stiffness measurement , our study showed that pSWE was a useful technique for confirming significant PPF . In addition , there were differences between the median liver stiffness values of mild pattern when compared to moderate , advanced and very advanced patterns , and also between moderate and very advanced patterns . Furthermore , the liver stiffness measurement was higher ( p < 0 . 001 ) in the group with significant fibrosis when compared with the mild fibrosis group . Our results revealed that the best cutoff value for excluding and confirming significant PPF was equal or less than 1 . 11 m/s and above 1 . 39 m/s , with a AUC of 0 . 719 with sensitivity of 80 . 5% for the lower cutoff and specificity of 86 . 1% for the higher cutoff . Accordingly , the method showed better performance to confirm significant PPF , since more than 90% of patients were correctly classified . Similarly , Veiga et al . , in a multicentric study that included patients from Rio de Janeiro , Brazil , compared liver stiffness among hepatosplenic schistosomiasis , hepatitis C virus-cirrhotic , and control patients using TE . They concluded that liver stiffness evaluation may be a useful tool to differentiate portal hypertension related to cirrhosis from that of hepatosplenic schistosomiasis [21] . In this study , however , no association was detected between Niamey PPF patterns and TE evaluation . The lack of association between these parameters in this study might have been caused by the small sample size , predominant inclusion of patients with mild PPF patterns ( patterns B and C ) and lack of patients with more advanced patterns ( patterns E and F ) . On the other hand , Shiha et al . reported that TE would not be useful to diagnose liver fibrosis and esophageal varices in patients with pure schistosomiasis [20] . Nevertheless , these authors , did not applied the Niamey classification , and their paper aimed to compare patients with and without esophageal varices and splenomegaly with TE results . Moreover , since schistosomiasis PPF is not diffuse by the liver parenchyma , TE evaluation does not allow ROI to be placed in a suitable area , which is between the periportal fibrosis septa . On the other hand , pSWE enable a completely view of the liver parenchyma and ROI positioning to improve liver stiffness evaluation in this area . US scan has become the most consolidated tool for evaluation of PPF in endemic areas because it shows practical usefulness , satisfactory interobserver agreement , and enables data comparison on international scale [6–8 , 21] . Despite the widespread use of US for diagnosing and monitoring changes caused by PPF , its use has some limitations , since it is a subjective method and requires physician expertise to apply the WHO Niamey-Belo Horizonte protocol . In our region , Northeast Brazil , which is endemic for schistosomiasis mansoni , the US scan is widely used in clinical practice , mainly in the rural zone and for scientific research in the field . Consequently , with the recent advent and widespread use of liver elastography , we aimed to evaluate the performance of pSWE as complementary tool for assessment and staging of PPF in schistosomotic patients . This method was chosen since it is more easily applied than Niamey US protocol , and , could enhance the diagnosis of those patients with advanced PPF , who need greater care and supervision . Moreover , it would provide a foundation for further studies to assess the presence and severity of portal hypertension and prediction of esophageal varices using pSWE and TE [21 , 24–26] . pSWE is a promising US-based method and it has been currently used to assess the severity of liver fibrosis in patients with chronic viral hepatitis , showing accuracy similar to TE [15 , 16] . In addition , pSWE has also showed promising results in patients with non-alcoholic fatty liver disease and non-alcoholic steatohepatitis [27] . As advantages pSWE is a technique that can be integrated into routine ultrasound protocols , it can be used in patients with ascites , require less sophisticated skills from the operator , and it is also possible to determine optimal region of interest ( ROI ) placement , avoiding structures such as focal lesions , large blood vessels and heterogeneous areas [15 , 28] . Nevertheless , there are only few studies evaluating elastography in schistosomotic patients , and schistosomiasis mansoni monoinfection has not been evaluated by pSWE so far . Accordingly , in studies evaluating coinfected patients with PPF and chronic hepatitis B [29] or C [25 , 26 , 30] , the liver stiffness measurement using pSWE could be overestimated . Moreover , patients with schistosomiasis mansoni are usually excluded from elastography analyzes to assess the severity of liver fibrosis in patients with viral hepatitis [31 , 32] . Overall , this study showed some limitations considering that US and pSWE were performed by the same operator . Nonetheless , the pSWE assessment was performed after the US scan , avoiding or reducing operator subjectivity interference . Additionally , further studies comparing US and pSWE results of different operators and analyzing inter and intraobsever variability , would provide better evidence about pSWE reproducibility for predicting PPF . In the future , new portable US devices capable of performing pSWE can be developed , allowing this technique to be applied to schistosomotic patients in endemic areas , especially in the field . Therefore , it will be possible to improve these patients’ diagnosis and prognosis without the need of a trained examiner . In conclusion , pSWE elastography was able to differentiate significant from mild PPF , with better performance to predict significant PPF . Therefore , pSWE is a potential tool for noninvasive assessment of disease severity in patients with schistosomiasis mansoni . | In the developing world , over 207 million people are infected with parasitic Schistosoma worms . Among the species of Schistosoma that infect humans Schistosoma mansoni is one of the most common causes of illness . Here , we investigated the performance of point shear wave elastography ( pSWE ) for predicting significant periportal fibrosis ( PPF ) in schistosomotic patients and to determine its best cutoff point . We examined 358 people from northeast of Brazil for Schistosoma infections . The present study showed that pSWE was able to differentiate significant from mild PPF , with better performance to predict significant PPF . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] | [
"schistosoma",
"invertebrates",
"schistosoma",
"mansoni",
"medicine",
"and",
"health",
"sciences",
"diagnostic",
"radiology",
"ultrasound",
"imaging",
"helminths",
"tropical",
"diseases",
"fibrosis",
"parasitic",
"diseases",
"animals",
"liver",
"diseases",
"developmental",
"biology",
"hepatitis",
"gastroenterology",
"and",
"hepatology",
"neglected",
"tropical",
"diseases",
"portal",
"hypertension",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"imaging",
"techniques",
"liver",
"fibrosis",
"helminth",
"infections",
"schistosomiasis",
"radiology",
"and",
"imaging",
"eukaryota",
"diagnostic",
"medicine",
"biology",
"and",
"life",
"sciences",
"viral",
"diseases",
"organisms"
] | 2018 | Liver ultrasound elastography for the evaluation of periportal fibrosis in schistosomiasis mansoni: A cross-sectional study |
It is widely believed that the modular organization of cellular function is reflected in a modular structure of molecular networks . A common view is that a “module” in a network is a cohesively linked group of nodes , densely connected internally and sparsely interacting with the rest of the network . Many algorithms try to identify functional modules in protein-interaction networks ( PIN ) by searching for such cohesive groups of proteins . Here , we present an alternative approach independent of any prior definition of what actually constitutes a “module” . In a self-consistent manner , proteins are grouped into “functional roles” if they interact in similar ways with other proteins according to their functional roles . Such grouping may well result in cohesive modules again , but only if the network structure actually supports this . We applied our method to the PIN from the Human Protein Reference Database ( HPRD ) and found that a representation of the network in terms of cohesive modules , at least on a global scale , does not optimally represent the network's structure because it focuses on finding independent groups of proteins . In contrast , a decomposition into functional roles is able to depict the structure much better as it also takes into account the interdependencies between roles and even allows groupings based on the absence of interactions between proteins in the same functional role . This , for example , is the case for transmembrane proteins , which could never be recognized as a cohesive group of nodes in a PIN . When mapping experimental methods onto the groups , we identified profound differences in the coverage suggesting that our method is able to capture experimental bias in the data , too . For example yeast-two-hybrid data were highly overrepresented in one particular group . Thus , there is more structure in protein-interaction networks than cohesive modules alone and we believe this finding can significantly improve automated function prediction algorithms .
Using a suited algorithm , any network can be separated into cohesive groups of nodes with more internal than external connections . Accordingly , protein-protein interaction networks can also be divided into relatively independent units as putative functional modules [4] . Do these modules really reflect a typical characteristic of the cellular network ? Here we used an alternative approach for the clustering of protein interactions . We grouped proteins of a similar functional role together . The functional role was defined by the interactions with proteins of other groups . In contrast to cohesive modules , which are more or less independent , groups which specifically linked other groups of proteins could be identified . Thus , an interconnectivity of biological units , as in the case of shared components in protein complexes , can also be observed at the cellular level . Using a GeneOntology based classification of all proteins within the modules , we found that these roles are mainly determined by cellular localization but also by function . Although possibly not too surprising to the biologist , this result underlines that the classes we identified by automatic clustering do represent a biological signal . Using HPRD as a data source , a large-scale set of interactions with , on average , eight connections per protein could be analyzed . As HPRD contains manually curated data , their quality should be high enough to extend the results to higher coverage . The analysis of interactions derived by different experimental methods revealed a bias in the coverage , especially for yeast-two-hybrid data . The great difference of the protein interactions verified only by Y2H to the other methods reminds us to pay attention to the careful weighting of quality and quantity . As large scale binary interaction analyses were mainly based on Y2H , using high coverage data such as that from Saccharomyces cerevisiae or Drosophila melanogaster might even blur the signal . Another drawback was the small number of interactions per protein , around three or four for the yeast , fly and nematode sets analyzed in the study by Wang and Zhang [17] . Still , it would be interesting to compare networks between different organisms to see whether there are changes in the clusters correlated , for example , with the emergence of multicellularity . Contrasting to previous approaches , which compared networks either globally [43] , [44] or locally [45] , [46] , comparing the image graphs allows detection of changes in the overall layout of the protein interaction network . But , reliable results can only be obtained when analyzing data sets of comparable quality and size [47] . In summary , our analysis showed that protein interaction networks are more than sparsely interacting cohesive modules . Rather , groups of proteins are connected by distinct sets of other proteins . These may be highly connected internally , but do not have to be . Therefore , functional roles and corresponding image graphs provide better descriptors for the characteristics of a protein interaction network than cohesive modules alone . They can help to further improve protein function prediction based on protein-interaction networks . | Cellular function is widely believed to be organized in a modular fashion . On all scales and at all levels of complexity , relatively independent sub-units perform relatively independent sub-tasks . This functional modularity must be reflected in the topology of molecular networks . But how a functional module should be represented in an interaction network is an open question . On a small scale , one can identify a protein-complex as a module in protein-interaction networks ( PIN ) , i . e . , modules are understood as densely linked ( interacting ) groups of proteins , that are only sparsely interacting with the rest of the network . In this contribution , we show that extrapolating this concept of cohesively linked clusters of proteins as modules to the scale of the entire PIN inevitably misses important and functionally relevant structure inherent in the network . As an alternative , we introduce a novel way of decomposing a network into functional roles and show that this represents network structure and function more efficiently . This finding should have a profound impact on all module assisted methods of protein function prediction and should shed new light on how functional modules can be represented in molecular interaction networks in general . | [
"Abstract",
"Discussion"
] | [
"computational",
"biology/systems",
"biology"
] | 2010 | Protein Interaction Networks—More Than Mere Modules |
The cytotoxin-associated gene ( Cag ) pathogenicity island is a strain-specific constituent of Helicobacter pylori ( H . pylori ) that augments cancer risk . CagA translocates into the cytoplasm where it stimulates cell signaling through the interaction with tyrosine kinase c-Met receptor , leading cellular proliferation . Identified as a potential gastric stem cell marker , cluster-of-differentiation ( CD ) CD44 also acts as a co-receptor for c-Met , but whether it plays a functional role in H . pylori-induced epithelial proliferation is unknown . We tested the hypothesis that CD44 plays a functional role in H . pylori-induced epithelial cell proliferation . To assay changes in gastric epithelial cell proliferation in relation to the direct interaction with H . pylori , human- and mouse-derived gastric organoids were infected with the G27 H . pylori strain or a mutant G27 strain bearing cagA deletion ( ∆CagA::cat ) . Epithelial proliferation was quantified by EdU immunostaining . Phosphorylation of c-Met was analyzed by immunoprecipitation followed by Western blot analysis for expression of CD44 and CagA . H . pylori infection of both mouse- and human-derived gastric organoids induced epithelial proliferation that correlated with c-Met phosphorylation . CagA and CD44 co-immunoprecipitated with phosphorylated c-Met . The formation of this complex did not occur in organoids infected with ∆CagA::cat . Epithelial proliferation in response to H . pylori infection was lost in infected organoids derived from CD44-deficient mouse stomachs . Human-derived fundic gastric organoids exhibited an induction in proliferation when infected with H . pylorithat was not seen in organoids pre-treated with a peptide inhibitor specific to CD44 . In the well-established Mongolian gerbil model of gastric cancer , animals treated with CD44 peptide inhibitor Pep1 , resulted in the inhibition of H . pylori-induced proliferation and associated atrophic gastritis . The current study reports a unique approach to study H . pylori interaction with the human gastric epithelium . Here , we show that CD44 plays a functional role in H . pylori-induced epithelial cell proliferation .
The major cause of chronic inflammation in the stomach is Helicobacter pylori ( H . pylori ) [1] , and it is widely accepted that chronic inflammation is a trigger for the development of gastric cancer [2] . The severity and localization of the inflammation that results from H . pylori infection is believed to dictate the pathological consequence of disease . Individuals most at risk of developing gastric cancer are those in whom the bacteria colonize the corpus ( or fundus ) of the stomach , when acid secretion is impaired . The subsequent development of severe inflammation in the gastric fundus leads to atrophy of the acid-secreting parietal cells and subsequently further hypochlorhydria , metaplasia and carcinoma [3 , 4 , 5] . Given that individuals most at risk of developing gastric cancer are those in whom the bacteria colonize the corpus [3 , 4 , 5] , the current research is focused on the use of human- and mouse-derived fundic gastric epithelium , cultured as 3-dimensional structures called gastrointestinal organoids , for the study of H . pylori pathogenesis . The cytotoxin-associated gene ( cag ) pathogenicity island is a strain-specific constituent of H . pylori that augments cancer risk [6] . The cag pathogenicity island encodes a type IV secretion system that is a multimolecular complex that mediates the translocation of bacterial factors into the host cell [6 , 7] . Upon delivery into the host cells by the type IV cag secretion system , CagA translocates into the host cell cytoplasm where it can stimulate cell signaling through interaction with several host proteins [6 , 8 , 9] including the tyrosine kinase c-Met receptor [10 , 11 , 12] . CagA exerts effects within host cells that mediate carcinogenesis , including aberrant activation of phosphatidylinositol 3-phosphate kinase ( PI3K ) and β catenin , disruption of apical-junctional complexes , and loss of cellular polarity [13 , 14 , 15] . Another host molecule that may influence carcinogenesis in conjunction with H . pylori and CagA is the cluster-of-differentiation ( CD ) CD44 cell surface receptor for hyaluronate [16] . CD44 is a cell surface adhesion molecule , expressed on a variety of cells including gastric epithelial cells , that has recently been identified as a gastric cancer stem cell marker , whereby cells expressing CD44 have been shown to possess the properties of gastric cancer stem cells [17] . CD44 variant isoforms , in particular CD44v6 , was identified as a marker for invasive intramucosal carcinoma and premalignant lesions [18] . Suzuki et al . [19] demonstrated that CagA CM motifs interact with Met leading to sustained PI3K-AKT signaling in response to H . pylori resulting in cellular proliferation . Notably , the isoform containing exon v6 ( CD44v6 ) acts as the coreceptor for c-Met , most probably , through binding of c-Met ligand hepatocyte growth factor ( HGF ) [20 , 21] . The coreceptor function of CD44v6 for c-Met is of particular interest given that studies pinpoint CD44v6 as a marker of early invasive intramucosal gastric carcinoma [18] . Whether CD44v6 acts as a coreceptor for the function of c-Met in response to H . pylori infection is unknown . Our current knowledge of H . pylori pathogenesis is largely based on data generated from gastric cancer cell lines or in vivo animal models of inflammation . Thus , despite extensive evidence demonstrating that H . pylori induces gastric epithelial changes , the direct impact of the bacterium on the normal epithelium is unclear . Culture of primary human- and mouse-derived gastric stem cells as 3-dimensional structures called gastrointestinal organoids are a rapidly emerging approach to study gastrointestinal development , physiology , stem cell biology and disease [22 , 23 , 24 , 25 , 26 , 27 , 28 , 29] . Troy-positive cells are expressed at the corpus gland base in a subset of differentiated chief cells [23] . Stange et al . [23] demonstrate that Troy-positive chief cells may be used to generate long-lived gastric organoids , but in vitro these cultures are differentiated toward the mucus-producing cell lineages of the neck and pit regions . The Troy-derived organoids are distinct from the cultures that we derive from whole dissociated glands reported here such that we have devised a method to maintain all the major cell lineages of the fundus [22 , 28] . In this investigation , we used our method of mouse-derived gastric organoid cultures as an approach to assay changes in gastric epithelial cell proliferation in relation to the direct interaction with H . pylori [22 , 24 , 29] . To study the functional role of CD44 in the context of human epithelial tissue , we developed a protocol for culturing human-derived gastric organoids . We developed cultures of human-derived fundic gastric organoids independent of the recent report by the Clevers group demonstrating the establishment of a similar culture model for the study of H . pylori pathogenesis [25] . Despite the extensive use of these culture systems for the study of stem cell biology and gastrointestinal development [22 , 23 , 24 , 25 , 26 , 27 , 29] , the degree to which these cultures reflect the physiology of native tissue has been reported by our laboratory alone [28] . Here we extend our current knowledge of H . pylori pathogenesis by identifying the signaling mechanism by which bacterial infection induces proliferation in the gastric epithelium . While it is known that c-Met is an important CD44 partner in proliferation , this is the first report that this association occurs in response to H . pylori infection . We find that CD44 plays a functional role in H . pylori-induced proliferation both in vitro and in vivo .
To determine if CD44 plays a functional role in H . pylori-induced proliferation , C57BL/6 ( BL/6 ) control and CD44 deficient mice were infected with mouse adapted LSH100 H . pylori strain for 4 weeks . The LSH100 mouse-adapted strain [30] is a descendant of the clinical isolate G27 [31] . We chose the LSH100 H . pylori to study the mechanism of bacterial-induced proliferation in vivo because this particular strain efficiently expresses virulence factor CagA [14 , 31 , 32 , 33] . The proliferating cells were measured by BrdU incorporation ( Fig . 1 ) . There was a significant increase of BrdU positive cells per gland in H . pylori LSH100 strain infected mice ( 6 . 23 + 0 . 53 BrdU+ cells/gland , Fig . 1B , E ) compared to Brucella broth control mice ( 3 . 64 + 0 . 04 BrdU+ cells/gland , Fig . 1A , E ) . Infection of BL/6 mice with LSH100 strain bearing a CagA deletion ( ∆CagA ) lacked the significant increase in proliferation ( 3 . 68 + 0 . 46 BrdU+ cells/gland , Fig . 1C , E ) . Importantly , CD44 deficient mice infected with H . pylori did not exhibit an increase in proliferation ( 2 . 09 + 0 . 11 BrdU+ cells/gland , Fig . 1D , E ) when compared to the Brucella broth uninfected control CD44KO mouse group ( 2 . 10 + 0 . 29 BrdU+ cells/gland , Fig . 1E ) . These data show that CD44 mediates CagA dependent H . pylori-induced proliferation . To identify the direct impact of H . pylori on the host gastric epithelium , we employed the use of a mouse-derived gastric organoid culture system . We have previously described a system for culturing gastric organoids derived from mouse fundic tissue , in which the fundic organoids are embedded in Matrigel , provided gastric organoid growth media , and co-cultured in transwell plates with immortalized stomach mesenchymal cells ( ISMCs ) [22 , 28] . Organoids were microinjected with H . pylori strain G27 . H . pylori strain G27 , originally isolated from an endoscopy patient from Grosseto Hospital ( Tuscany , Italy ) [31] , is readily transformable and therefore amenable to gene disruption [34] . Of relevance to the current study , strain G27 efficiently delivers the translocated virulence factor CagA to cells in culture [14 , 31 , 32 , 33] . Therefore , we chose G27 to study the mechanism of H . pylori-induced proliferation using a strain that efficiently expresses virulence factor CagA . Organoids were microinjected with H . pylori G27 strain ( Fig . 2A ) , and bacterial adhesion was confirmed by Warthin-Starry stain ( Fig . 2C , D ) and culture ( Fig . 2E ) . Quantitative cultures showed a significant increase in the bacteria cultured from organoids infected for 7 days compared to 24 hours and thus confirming bacterial viability within the cultures ( Fig . 2E ) . Therefore , the fundic organoids provided a method by which bacterial-host cell interactions may be studied in the context of an intact normal gastric epithelium in vitro . When the mFGOs were infected with H . pylori G27 strain , we observed a significant increase in epithelial cell proliferation in response to bacterial infection compared to the uninfected controls ( Fig . 3A , B ) . The proliferative response to H . pylori was significantly blocked when organoids were pretreated with the c-Met inhibitor ( c-MetI PF04217903 mesylate ) ( Fig . 3A , B ) . Organoids infected with the G27 H . pylori strain that expressed a deletion of CagA ( ∆CagA ) did not differ from the controls with regards to proliferation ( Fig . 3A , B ) . These data show that H . pylori-induced proliferation is mediated by activation of c-Met signaling as previously reported [19] . Therefore , to advance this current knowledge , we examined whether c-Met was associated with CD44 . Lysates were prepared from uninfected organoids and organoids infected with either H . pylori ( G27 strain ) or ∆CagA and immunoprecipitated using an anti-c-Met antibody . Immunoprecipitates analyzed by Western blot using an anti-phosphotyrosine antibody showed an increase in phosphorylated c-Met in response to H . pylori ( Fig . 3C ) . Consistent with published studies CagA coimmunoprecipitated with c-Met [19] . CD44 also coimmunoprecipitated with c-Met . C-Met is an important partner with CD44 in proliferation , but this is the first time that it has been reported that this association occurs in response to H . pylori infection . There is evidence suggesting CD44 binds to hepatocyte growth factor ( HGF ) and acts as a co-receptor by presenting HGF to c-Met and subsequently activating Met signaling [20 , 21] ( Fig . 3D ) . In response to H . pylori infection , we show for the first time , HGF also coimmunoprecipitated with c-Met , and HGF expression was significantly upregulated in response to bacterial infection ( Fig . 3C ) . Collectively , these data suggest that H . pylori induces the proliferation of the gastric epithelium by promoting an association between CagA , Met and CD44 . To test the functional role of CD44 in mouse-derived gastric organoids , we cultured organoids derived from mice that lacked the gene for CD44 ( CD44KO ) . Organoids derived from the stomachs of CD44 deficient ( CD44KO ) mice did not proliferate in response to H . pylori infection ( Fig . 3E ) . CD44KO mouse-derived organoids proliferated in response to a Wnt agonist , and thus showing that this lack of response was specific to H . pylori ( Fig . 3E ) . To determine if CD44 was required for c-Met phosphorylation , lysates from uninfected control , H . pylori infected and ∆CagA infected organoids derived from stomachs of CD44-deficient mice were collected . In H . pylori infected CD44-deficient organoids c-Met was present but not phosphorylated ( Fig . 3F ) . Importantly , in the absence of CD44 , CagA co-immunoprecipitated with c-Met and is thus likely to form a complex with this receptor ( Fig . 3F ) . In the absence of CD44 , HGF was not detected in the c-Met immunoprecipitated protein complex ( Fig . 3F ) . Collectively , these data further show that both CD44 and c-Met play a functional role in CagA dependent H . pylori-induced epithelial cell proliferation in mouse-derived fundic gastric organoids . In addition to increased epithelial cell proliferation , we observed striking morphological changes in response to H . pylori infection that were consistent with epithelial-to-mesenchymal transition ( EMT ) ( Fig . 4 ) . In uninfected control ( CON ) mouse-derived fundic gastric organoids , we observed clear membrane-expressed E-cadherin ( Fig . 4A ) . However , in the H . pylori G27 strain infected organoids there was a disruption in membrane-expressed E-cadherin and transition of the spheroid morphology to a cell monolayer was observed ( Fig . 4B ) . In control mouse-derived fundic gastric organoids infected with the ∆cagA strain , we also observed clear membrane-expressed E-cadherin ( Fig . 4C ) . EMT was also documented by increased expression of markers that included alpha smooth muscle actin ( αSMA ) , SNAIL2 , TWIST1 , N-cadherin and Zeb1 ( Fig . 4D ) . Organoids derived from the stomachs of CD44KO mice were also microinjected with Brucella broth ( uninfected control ) , H . pylori G27 strain or H . pylori ∆cagA strain . In all three groups membrane-expressed E-cadherin was observed ( Fig . 4E ) . In addition changes in gene expression of αSMA , SNAIL2 , TWIST1 , N-cadherin and Zeb1 was not observed in infected CD44KO mouse-derived organoids ( Fig . 4F ) . These data show that H . pylori induces an EMT phenotype in the mouse-derived fundic gastric organoids and that this response may be mediated by CD44 . To identify the role of CD44 as a mediator of H . pylori-induced proliferation in human epithelium , we developed the human-derived fundic gastric organoids ( hFGOs ) . Gastric glands were enzymatically dissociated from the human fundic tissue embedded in Matrigel , provided gastric organoid growth media and cultured to form hFGOs epithelial spheres over 7 days ( Fig . 5A ) . These 3 dimensional epithelial spheres contained markers specific for the fundic epithelium , including H+ , K+ ATPase ( HK ) , Muc5ac , and Muc6 , but do not contain gastrin , a marker for the antral region of the stomach ( Fig . 5B ) . Flow cytometric analysis showed that the majority of cells in the human organoids were positive for the marker for parietal cells , H+ , K+ ATPase , as would be expected in oxyntic glands in the human stomach ( Fig . 5C ) . Quantification of the flow cytometry histograms revealed that the hFGOs were comprised of approximately 5% UEAI+ cells ( surface mucous pit cells ) , 15% GSII+ cells ( mucous neck cells ) , 12% pepsinogen C ( PgC ) + cells ( chief cells ) , 7% chromogranin A+ cells ( ChgA endocrine cells ) and 60% H+ , K+ ATPase+ cells ( HK , parietal cells ) ( Fig . 5D ) . As detailed for the mouse-derived fundic gastric organoids , hFGOs were microinjected with H . pylori G27 strain and bacterial adhesion was confirmed by Warthin-Starry stain ( Fig . 5E , F ) . Therefore , the hFGOs provided a method by which bacterial-host cell interactions were studied in the context of an intact normal human gastric epithelium in vitro . To determine if the co-receptor role of CD44 and c-Met was intact in human tissue , hFGOs were infected with H . pylori G27 strain . Compared to the control treated ( CON ) hFGOs , infection with H . pylori triggered a significant induction in proliferating cells that was not seen in organoids injected with the ∆CagA mutant strain ( Fig . 6A , B ) . Consistent with the response observed in the mouse-derived organoids , the proliferative response to H . pylori was significantly blocked when hFGOs were pretreated with the c-Met inhibitor ( c-MetI PF04217903 mesylate ) ( Fig . 6A , B ) . As detailed for the mouse-derived fundic gastric organoids , lysates were prepared from uninfected and hFGOs infected with either H . pylori ( G27 strain ) or ∆CagA and immunoprecipitated using an anti-c-Met antibody . Immunoprecipitates analyzed by western blot using an anti-phosphotyrosine antibody showed an increase in phosphorylated c-Met in response to H . pylori ( Fig . 6C ) . CagA , CD44 and HGF co-immunoprecipitated with c-Met . It is known that c-Met is an important partner for CD44 in proliferation , but this is the first report of the CD44/c-Met association occuring in response to H . pylori infection of the human gastric epithelium . To test the function of CD44v6 in hFGOs , we used a neutralizing antibody that specifically targets human CD44v6 . We found H . pylori-induced proliferation was blocked in hFGOs pre-treated with the CD44v6 neutralizing antibody for one hour prior to H . pylori injection ( Fig . 6D , E ) . Treating hFGOs with the CD44v6 neutralizing antibody alone did not significantly change the baseline of proliferation seen in the control hFGOs ( Fig . 6D , E ) . Consistent with data shown in Fig . 6C , immunoprecipitates analyzed by western blot showed an increase in phosphorylated c-Met in response to H . pylori and CagA , CD44 and HGF co-immunoprecipitated with c-Met ( Fig . 6F ) . Interestingly , inhibition of CD44 binding to hyaluronic acid using the CD44v6 peptide inhibited the co-immunoprecipitation of CD44 and HGF with c-Met ( Fig . 6F ) . These data indicate that CD44v6 plays a functional role in H . pylori-induced proliferation in hFGOs , and that c-Met may collaborate with CD44 in this proliferative response to H . pylori infection in human tissue . We next investigated the role of CD44 in cancer progression using a well-established Mongolian gerbil model of gastric cancer [35 , 36] . We administered an inhibitory peptide , Pep1 , that prevents binding of CD44 ligand hyaluronic acid ( HA ) , thus blocking CD44 downstream signaling . Gerbils were infected with H . pylori strain 7 . 13 , an in vivo adapted strain originally isolated from a gastric ulcer [37] and reported to reproducibly induce gastric cancer in Mongolian gerbils [13 , 37] . We found that gerbils infected with H . pylori strain 7 . 13 , developed atrophic gastritis ( Fig . 7B ) , as documented by neutrophil ( Fig . 7E ) and lymphocyte ( Fig . 7F ) infiltration , development of lymphoid follicles ( Fig . 7G ) and atrophy ( Fig . 7H ) , 6 weeks of infection , compared to control ( Brucella broth administered ) gerbils ( Fig . 7A , E-H ) . Gerbils that received Pep1 injections three times a week for the duration of the 6 week infection with H . pylori , did not exhibit the development of atrophic gastritis ( Fig . 7D , E-H ) , compared to the gerbils who received H . pylori and the scrambled control peptide ( cPep1 ) ( Fig . 7C , E-H ) . Our studies show that CD44 plays a functional role in H . pylori-induced proliferation of mouse and human epithelial tissues in vitro . Thus , we next examined the proliferative state in response to H . pylori and Pep1 treatment in the Mongolian gerbils in vivo . Within 6 weeks of infection with H . pylori ( 7 . 13 ) there was a significant induction in the number of proliferating cells in the gastric epithelium ( Fig . 8B , E ) when compared to Brucella broth controls ( Fig . 8A , E ) . To determine if CD44 signaling was required for this H . pylori induced proliferative response , we treated the H . pylori infected gerbils three times a week with Pep1 , and found that the induction in proliferation was blocked ( Fig . 8C , E ) . The gerbils infected with H . pylori and treated with the scrambled control peptide ( cPep1 ) exhibited the expected induction in proliferation ( Fig . 8D , E ) . Collectively , these data show that CD44 signaling mediates H . pylori-induced atrophic gastritis and hyperproliferation in the Mongolian gerbil model of gastric cancer .
Prolonged cell proliferation in the gastric mucosa is a precursor to the progression from chronic inflammation to gastric cancer in response to H . pylori infection [2] . However , the mechanism by which H . pylori induces epithelial cell proliferation is not well defined . Here we report the culture of primary human- and mouse-derived gastric epithelial cells as 3-dimensional structures called gastrointestinal organoids for the study of H . pylori pathogenesis . We utilized mouse-derived fundic gastric organoids ( mFGOs ) , and for the first time we demonstrated that CD44 and c-Met form a complex with virulence factor CagA in response to H . pylori infection . To determine if this mechanism was present in human gastric epithelium , we developed a novel protocol for the culture of differentiated human-derived fundic gastric organoids in culture ( hFGOs ) . In the hFGOs , we found that inhibiting CD44 splice variant 6 , a specific marker for early invasive carcinoma [38] , blocked H . pylori-induced proliferation . As in the mouse organoids , H . pylori infection triggered the formation of a complex containing CD44 , c-Met , and CagA in the hFGOs . In addition , in the well-established Mongolian gerbil model of gastric cancer , animals treated with CD44 peptide inhibitor Pep1 , resulted in the inhibition of H . pylori-induced proliferation and associated atrophic gastritis . Our experiments report for the first time that CD44 acts as a mediator of H . pylori-induced proliferation both in vitro and in vivo . In vivo , we observed that CD44 regulates baseline proliferation whereby , as previously reported [39] , basal rate of proliferation was approximately half that of control mice . However , H . pylori infection did not induce epithelial cell proliferation within the CD44KO mice , suggesting a functional role of CD44 as a mediator of bacterial-induced proliferation . Consistent with previous findings in gastric cancer cell lines [10] , we also report that the proliferative response to H . pylori was CagA- and c-Met-dependent within the gastric epithelium . Western blot analysis using lysates collected from H . pylori infected mFGOs and hFGOs also showed that CagA coimmunoprecipitates with c-Met , as previously shown in human gastric cancer cells [19] . Although we report that H . pylori-induced proliferation was dependent on CD44 expression and signaling , the CagA/c-Met association was also observed in the absence of CD44 expression and signaling . Suzuki et al . [19] demonstrated that CagA CM motifs interact with Met leading to sustained PI3K-AKT signaling in response to H . pylori , leading to β-catenin activation and cellular proliferation . We advance our understanding of H . pylori-induced epithelial cell proliferation by demonstrating that CD44 acts as a coreceptor for c-Met in response to bacterial infection . The isoform containing exon v6 ( CD44v6 ) acts as the coreceptor for c-Met in tumor cell lines [20 , 21] . The coreceptor function of CD44v6 for c-Met is of particular interest given that studies pinpoint CD44v6 as a marker of early invasive intramucosal gastric carcinoma [18] . Whether CD44v6 acted as a coreceptor for the function of c-Met in response to H . pylori infection was unknown . However , we found that H . pylori-induced proliferation was blocked in hFGOs pre-treated with the CD44v6 neutralizing antibody , and thus indicating that CD44v6 plays a functional role in H . pylori-induced proliferation in the human gastric epithelium . Our data suggest a coreceptor function of CD44 for c-Met in response to H . pylori infection . The coreceptor function of CD44v6 for c-Met is shown to be dependent on HGF binding [20] . Indeed , we find that in the absence of CD44 expression and signaling HGF does not coimmunoprecipiate with c-Met and H . pylori fails to induce c-Met phosphorylation . In support of our findings , HGF binding to c-MET results in receptor homodimerization and phosphorylation of two tyrosine residues ( Y1234 and Y1235 ) located within the catalytic loop of the tyrosine kinase domain [40] , and subsequent , phosporylation of tyrosines 1349 and 1356 in the carboxy-terminal tail [41] . When these tyrosines become phosphorylated , they recruit signaling effectors that includes phosphatidylinositol 3-kinase ( PI3K ) . When Y1313 is phosphorylated , it binds and activates PI3K , which promotes cell viability , motility and proliferation [42] . The extracellular domain of CD44v6 is necessary for c-Met activation , and this is dependent on HGF binding [20] . Indeed in the mFGOs , in response to H . pylori infection , HGF also coimmunoprecipitated with c-Met . The collaboration between c-Met and CD44v6 contributes to another bacterial infection , for example the invasion of Listeria monocytogenes into target cells [43] . Based on evidence demonstrating a functional role of CD44 in bacterial infection [43] including H . pylori infection , we conclude that the activation of c-Met is not only dependent on binding of H . pylori but in addition requires adhesion molecule CD44v6 as a co-receptor . In addition , although the coreceptor function of CD44v6 for c-Met is shown to be dependent on HGF binding , we show here for the first time that this signaling pathway mediates H . pylori-induced epithelial cell proliferation . In the well-established Mongolian gerbil model of gastric cancer [35 , 36] , animals treated with CD44 peptide inhibitor Pep1 , resulted in the inhibition of H . pylori-induced proliferation and associated atrophic gastritis . Consistent with previous findings , infection with H . pylori for 6 weeks induced a significant inflammatory response , accompanied with the development of atrophic gastritis and metaplasia [35 , 44] . Using Pep1 , a peptide that blocks the interaction between hyaluronic acid and CD44 , we found that H . pylori-induced atrophic gastritis was inhibited . Importantly , infected gerbils that received Pep1 exhibited a significant reduction in actively proliferating cells . In support of our studies , a previous report by Khurana et al . [39] found that CD44 is a key coordinator of cell proliferation in a model of chemically-induced parietal cell atrophy , through downstream activation of STAT3 . These findings strongly suggest that future therapeutic targets could include CD44 inhibition , to prevent H . pylori-induced hyperproliferation and cancer progression . Our knowledge of H . pylori pathogenesis is predominantly based on data generated from gastric cancer cell lines or in vivo animal models of inflammation . Thus , limitation in acquiring such knowledge has been attributed to the inability to evaluate molecular mechanisms of bacterial and host cell interactions in a setting of a sustained gastric epithelial cell diversity and polarity . Our current work reports the development and use of a model of primary human and mouse cultured gastric epithelial cells . These models recapitulate key features of the gastric environment including the presence of the major gastric cell lineages and a polarized epithelium . In particular , we and others [25] report that the hFGO culture system represents a vital new technique for modeling H . pylori infection within the normal human tissue in vitro . As reported in similar mouse- and human-derived gastric organoid cultures [22 , 24 , 29] , our study takes advantage of the presence of a defined lumen in these models , allowing us to inject live H . pylori directly into the gastric organoid and assay the epithelial response without the influence of host recruited factors . This contribution is significant because it provides knowledge required to potentially develop techniques to disrupt bacterial colonization and prevent disease progression . Our study helps to uncover a potential mechanism of H . pylori-induced proliferation in rodent and human tissue , using both in vivo established models of H . pylori infection and gastric cancer , as well as a novel epithelial cell culture system .
All mouse studies were approved by the University of Cincinnati Institutional Animal Care and Use Committee ( IACUC ) that maintains an American Association of Assessment and Accreditation of Laboratory Animal Care ( AAALAC ) facility . Human fundus was collected during sleeve gastrectomies ( IRB protocol number: 2013–2251 ) . Helicobacter pylori ( H . pylori ) strain LSH100 , a descendant of the clinical isolate G27 [30] , G27 wild type [31] ( CagA+ ) , a mutant G27 strain bearing a cagA deletion ( ΔcagA::cat ) [45] and H . pylori 7 . 13 strain were grown on blood agar plates containing Columbia Agar Base ( Fisher Scientific ) , 5% horse blood ( Colorado Serum Company ) , 5 µg/ml vancomycin and 10 µg/ml trimethoprim as previously described [46] . Plates were incubated for 2–3 days at 37°C in a humidified microaerophilic chamber [47] . Bacteria were harvested and resuspended in filtered Brucella broth ( BD biosciences ) supplemented with 10% fetal calf serum . After 12 hours of growth at 37°C in a humidified microaerophilic chamber , bacteria were harvested , resuspended in filtered Brucella broth and C57/BL6 mice ( The Jackson Laboratory , stock number: 000664 ) , B6 . 129 ( Cg ) -Cd44tm1Hbg/J ( CD44-defiecient ) mice ( The Jackson Laboratory , stock number stock number: 005085 ) or Mongolian gerbils ( Charles River ) were inoculated by oral gavage with 108 bacteria per 200 µl of Brucella broth . Mice were infected with either G27 wild type or a mutant G27 strain bearing a cagA deletion ( ΔcagA::cat ) . Mongolian gerbils were inoculated with H . pylori strain 7 . 13 . Uninfected control mice were administered 200 µl of Brucella broth . Mice were analyzed 4 weeks post-inoculation . Mongolian gerbils were analyzed 6 weeks post-inoculation . H . pylori colonization was quantified using the culture method previously published [47] . Briefly , the wet weight of gastric tissue collected from uninfected and infected animals was measured . Tissue was homogenized in 1 ml saline and diluted 1/100 and spread on blood agar plates containing Campylobacter Base Agar ( Fischer Scientific ) , 5% horse blood ( BD Diagnostic Systems ) , 5µg/ml vancomycin and 10µg/ml trimethoprim . Plates were incubated for 7–10 days at 37oC in a humidified microaerophilic chamber . Single colonies from these plates tested positive for urease ( BD Diagnostic Systems ) , catalase ( using 3% H2O2 ) and oxidase ( DrySlide , BD Diagnostic Systems ) . Colonies were counted and data normalized using the tissue wet weight and expressed and colony forming units ( CFU ) /g tissue . Specific hyaluronan ( HA ) blocking peptide Pep 1 ( NH2-GAHHWQFNALTVRGGGS-CONH2 ) and scrambled control ( NH2-WRHGFALTAVNQGGGS-CONH2 ) [48] peptides were synthesized by Pierce Biotechnology ( Thermo Scientific-3747 , Rockford , IL ) . Pep1 and control peptides were administered by intraperitoneal injection to Mongolian gerbils at a concentration of 10 mg/kg of body weight three times a week for 6 weeks after H . pylori infection . Mongolian Gerbils were euthanized 6 weeks post-infection , stomachs were divided longitudinally for representative sections from both sides of the tissue and stained for Hematoxylin & Eosin ( H&E ) . H&E stains were analyzed by using the updated Sydney classification system for histological scoring of gastritis [49] . Mice and Mongolian gerbils were injected with BrdU ( 300 mg/kg ) 24 hours prior to analysis . Stomach sections spanning both the fundic and antral regions collected from experimental animals were fixed for 16 hours in Carnoy’s Fixative , paraffin embedded and sectioned at 5 µM . Prepared slides were deparaffinized with antigen retrieval performed by submerging in boiling solution ( 1:100 dilution Antigen Unmasking Solution in dH2O , Vector Laboratories , H-3300 ) for 10 minutes followed by 20 minutes at room temperature . Sections were then blocked with 5% BSA/PBS for 20 minutes at room temperature . BrdU color development was performed according to manufacturer’s protocol ( Roche , Cat . No . 11 296 736 001 ) . Immunohistochemical slides were dehydrated and mounted using Permount and images viewed and captured under light microscopy ( Olympus BX60 with Diagnostic Instruments “Spot” Camera ) . Gastric fundic organoids were prepared based on our recently reported protocol [22 , 50] . Primary epithelial cells from adult stomach tissue were cultured as 3-dimensional structures called mouse-derived fundic gastric organoids ( mFGOs ) . Stomachs were dissected from mice along the greater curvature and washed in ice-cold Ca2+/Mg2+-free Dulbecco’s Phosphate Buffered Saline ( DPBS ) . The stomach was stripped from muscle and visible blood vessels . Gastric fundus was further separated and cut into approximately 5 mm2 pieces . Tissue was then incubated in 5 ml of 5 mM EDTA for 2 hours at 4°C with gentle shaking . The EDTA was replaced with 5 ml chelation buffer ( 1 g D-sorbital , 1 . 47 g sucrose in 100 ml DPBS ) . Next tissue was shaken vigorously for approximately 2 minutes to dissociate glands . Dissociated glands were centrifuged at 150 g for 5 minutes then were embedded in Matrigel ( BD Biosciences ) supplemented with Advanced DMEM/F12 medium ( Invitrogen ) , Wnt conditioned medium [22] , R-spondin conditioned medium [22] supplemented with gastric growth factors including bone morphogenetic protein inhibitor , Noggin ( PeproTech ) , Gastrin ( Sigma ) , Epidermal grow factor ( EGF , PeproTech ) and Fibroblast growth factor 10 ( FGF-10 , PeproTech ) as previously described [22 , 50] . Glands grew into organoids by 1–2 days . After 4 days in culture mFGOs were cultured on the inner well of polyester Transwell inserts ( 0 . 4 µm pore size , catalogue number 3460 , Corning Lifescience ) , while immortalized stomach mesenchymal cells ( ISMCs ) were cultured in the base of the chamber of the Transwell according to our published protocol [22] . Organoids were co-cultured for a further 3 days prior to H . pylori infection and treatments . Human fundic gastric organoids ( hFGOs ) were generated independently of the recently reported protocol [25] . The fundic mucosa was stripped away from the muscle layer , and then cut into 5 mm2 pieces and washed 3 times in sterile DPBS without Ca2+ and Mg2+ . The mucosa was transferred to DMEM/F12 ( catalogue number 1263–010 , Gibco Life Technologies ) supplemented with 10mM HEPES , 1%Penicillin/Streptomycin and 1X Glutamax , and incubated while stirring and oxygenated in a 37oC water bath with Collagenase ( from Clostridium histolyticum , Sigma C9891 , 1 mg/ml ) and bovine serum albumin ( 2 mg/ml ) to release glands from the tissue . After 15–30 minutes of incubation collected glands were washed in sterile phosphate buffered saline with Kanamycin ( 50 mg/ml ) and Amphotericin B ( 0 . 25 mg/ml ) /Gentamicin ( 10 mg/ml ) , centrifuged at 200 xg , resuspended in the appropriate volume of Matrigel ( 50 µl of Matrigel/well ) , and subsequently cultured in human gastric organoid media ( DMEM/F12 supplemented with 10mM HEPES , 1X Glutamax , 1% Pen/Strep , 1X N2 , 1X B27 , 1mM N-Acetylcystine , 10mM Nicotidamide , 50ng/mL EGF , 100ng/mL Noggin , 20% R-Spondin Conditioned Media , 50% Wnt Conditioned Media , 200ng/mL FGF10 , 1nM Gastrin , 10uM Y-27632 , Kanamycin ( 50 mg/ml ) and Amphotericin B ( 0 . 25 mg/ml ) /Gentamicin ( 10 mg/ml ) ) ( Table 1 ) . Glands grew into organoids by 7 days at which time hFGOs were H . pylori infected and treated . We did not observe significant variations in organoid growth between donor gastric glands . H . pylori strains G27 and ΔCagA were grown on blood agar plates as described above , and prior to injection , a group of organoids were pre-treated for one hour with a highly selective , high affinity c-Met inhibitor ( 100 µg/ml , PF04217903 mesylate , Tocris Bioscience Cat . # 4239 ) . CD44-deficient mouse-derived organoids were treated with Wnt agonist ( 1µM CALBIOCHEM , catalogue number 681665 ) . The hFGOs were pretreated with CD44v6 antibody 1 hour prior to H . pylori infection at a concentration of 100 ng/ml . Organoids ( mFGOs and hFGOs ) cultured for 7 days were injected with 200 nl of Brucella broth using a Nanoject II ( Drummond ) microinjector , such that each organoid received approximately 2x105 bacteria . Twenty-four hours after injection , organoids were harvested by washing in ice-cold DPBS to remove Matrigel , followed by either RNA isolation by TRIzol , protein isolation by M-PER Mammalian Protein Extraction Reagent ( Thermo Scientific , IL ) or EdU labeling . The EdU solution was added to the organoid medium of either mFGOs or hFGOs for uptake for 1 hour . EdU staining was performed using the Click-iT Alexa Fluor 594 Imaging Kit , according to the manufacturer’s instructions ( Life Technologies ) . The mFGOs were fixed with 4% formaldehyde for 20 minutes , followed by permeabilization with 0 . 5% Triton X-100 in DPBS for 20 minutes at room temperature . Blocking was done with 2% normal goat serum for 20 minutes at room temperature . The organoids were then incubated at 4oC overnight with an antibody specific for E-cadherin ( Santa Cruz Biotechnology sc-59778 , 1:100 ) . Following this , organoids were incubated with Alexa Fluor 488 again overnight at 4oC . After incubation with the secondary antibodies organoids were counterstained using nuclear stain ( Hoechst 33342 , 10 µg/ml , Invitrogen ) for 20 minutes at room temperature . Organoids were visualized using the Zeiss LSM710 . Both mFGOs and hFGOs were harvested from Matrigel using ice-cold DPBS without Ca2+/Mg2+ and lysed in M-PER Mammalian Protein Extraction Reagent ( Thermo Scientific , IL ) supplemented with protease inhibitors ( Roche ) according to the manufacturer’s protocol . For the immunoprecipitation , c-Met antibody ( AbCam ab59884 ) was added to the cell lysates ( 1:50 dilution ) overnight at 4oC . Protein A/G Plus Agarose beads ( Santa Cruz sc-2003 ) were washed with PSB and added to the cell lysate mixture overnight at 4oC . Cell lysate mixtures were resuspended in 40 µl Laemmli Loading Buffer containing eta-mercaptoethanol ( Bio-Rad Laboratories , CA ) before western blot analysis . Samples were loaded onto 4–12% Tris-Glycine Gradient Gels ( Invitrogen ) and run at 80 V , 3 . 5 hours before transfer to nitrocellulose membranes ( Whatman Protran , 0 . 45 µM ) at 105 V , 1 . 5 hours at 4oC . Membranes were blocked for 1 hour at room temperature using KPL Detector Block Solution ( Kirkegaard & Perry Laboratories , Inc . ) . Membranes were incubated for 16 hours at 4°C with a 1:2000 dilution of either anti-GAPDH ( Millipore , MAB374 ) , 1:100 dilution of anti-phospho Tyrosine ( Santa Cruz sc-7020 ) , anti CagA ( Abcam ab90490 ) , anti-CD44 ( Abcam ab51037 ) , anti c-Met ( Abcam ab59884 ) , or anti-HGF ( Abcam ab83760 ) followed by a 1 hour incubation with a 1:1000 dilution anti-mouse or anti-rabbit Alexa Fluor 680 ( Invitrogen ) . Blots were imaged using a scanning densitometer along with analysis software ( Odyssey Infrared Imaging Software System ) . Total RNA was isolated from mFGOs , hFGOs and gastric glands using TRIzol ( Life Technologies ) according to manufacture’s protocol . A High Capacity cDNA Reverse Transcription Kit synthesized cDNA from 100 ng of RNA following protocol provided by Applied Biosystems . Real-time PCR assays were utilized for the following genes in the mouse-derived organoids: GAPDH , alpha-Smooth Muscle Actin ( Mm00725412_s1 ) , Zeb1 ( Mm00495564_m1 ) and 2 ( Mm0049713_m1 ) , TWIST 1 ( Mm04208233_g1 ) and 2 ( Mm000495564_m1 ) , SNAIL 1 ( Mm01249564_g1 ) and 2 ( Mm00441531_m1 ) . Cell lineage markers were determined by RT-PCR for HK-ATPase ( Hs01026288_m1 ) , gastrin ( Hs01107047_m1 ) , Muc5ac ( Hs00873651_mH ) , and Muc6 ( Hs01674026_g1 ) . PCR amplifications were done with pre-validated 20X TaqMan Expression Assay primers , 2X TaqMan Universal Master Mix ( Applied Biosystems ) , and cDNA template , in a total volume of 20 µL . Amplifications were performed with duplicate wells in a StepOne Real-Time PCR System ( Applied Biosytems ) , and fold change was calculated at ( Ct-Ct high ) = n target , 2ntarget/2nGAPDH = fold change where Ct = threshold cycle . The hFGOs were washed with cold DPBS without Ca2+/Mg2+ to remove Matrigel , and suspended in 2 ml Accutase ( Innovative Cell Technologies , Cat . # AT-104 ) for 5 minutes at 37oC . Organoids were broken into single cells using an 18G needle , passed through 20 times . Single cells were fixed and permeablized using the FIX&PERM Kit from Invitrogen ( catalogue number GAS004 ) , according to manufacturer’s instructions . A first tube of cells was co-stained with lectin FITC labeled Ulex europaeus ( UEAI , Sigma Aldrich ) , lectin Griffonia simplicifolia Alexa Fluor 647 ( GSII , Molecular Probes ) and then rabbit anti-pepsinogen C ( Abcam , ab104595 ) followed by an anti-rabbit IgG PE secondary antibody ( Abcam , ab7070 ) , all at a 1:100 dilution . A second tube of cells was co-stained for chromogranin A antibody ( Abcam , ab15160 ) followed by an anti-rabbit IgG PE conjugated secondary antibody , and anti-HK-ATPase ( MA3–923 , Affinity Bioreagents ) followed by an anti-mouse IgG FITC conjugated ( Abcam , ab6785 ) , all at a 1:100 dilution . The antibody incubations were done for a period of 20 minutes at room temperature . The stained cells were analyzed using the FACSCalibur flow cytometer ( BD Biosciences ) followed by the FloJo software ( Tree Star , Ashland , OR ) . The significance of the results was tested by two-way ANOVA using commercially available software ( GraphPad Prism , GraphPad Software , San Diego , CA ) . A P value <0 . 05 was considered significant . | Chronic gastric inflammation , typically caused by Helicobacter pylori ( H . pylori ) , is the most consistent lesion leading to cancer . During a well-choreographed interaction between H . pylori and the host , the progression from chronic inflammation to cancer involves gastric epithelial changes with evidence of hyperproliferation . Our knowledge of H . pylori pathogenesis is predominantly based on data generated from gastric cancer cell lines or animal models of inflammation . We report the development and use of a novel model of primary human and mouse cultured gastric epithelial cells that are organized into three-dimensional spheroid units containing a lumen , known as gastric organoids . To assay changes in gastric epithelial cell proliferation in relation to the direct interaction with H . pylori , human- and mouse-derived gastric organoids were infected with the bacteria . Cluster-of-differentiation gene ( CD44 ) is a transmembrane receptor responsible for epithelial cell proliferation . We show that CD44 plays a functional role in H . pylori-induced proliferation . In a Mongolian gerbil animal model of H . pylori-induced gastric cancer , we show that inhibiting CD44 blocks epithelial proliferation and subsequently cancer progression in response to bacterial infection . Thus our study provides new insights into the role of CD44 in H . pylori-induced hyperproliferation and progression of gastric disease . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | CD44 Plays a Functional Role in Helicobacter pylori-induced Epithelial Cell Proliferation |
The separation of the optic neuroepithelium into future retina and retinal pigment epithelium ( RPE ) is a critical event in early eye development in vertebrates . Here we show in mice that the transcription factor PAX6 , well-known for its retina-promoting activity , also plays a crucial role in early pigment epithelium development . This role is seen , however , only in a background genetically sensitized by mutations in the pigment cell transcription factor MITF . In fact , a reduction in Pax6 gene dose exacerbates the RPE-to-retina transdifferentiation seen in embryos homozygous for an Mitf null allele , and it induces such a transdifferentiation in embryos that are either heterozygous for the Mitf null allele or homozygous for an RPE–specific hypomorphic Mitf allele generated by targeted mutation . Conversely , an increase in Pax6 gene dose interferes with transdifferentiation even in homozygous Mitf null embryos . Gene expression analyses show that , together with MITF or its paralog TFEC , PAX6 suppresses the expression of Fgf15 and Dkk3 . Explant culture experiments indicate that a combination of FGF and DKK3 promote retina formation by inhibiting canonical WNT signaling and stimulating the expression of retinogenic genes , including Six6 and Vsx2 . Our results demonstrate that in conjunction with Mitf/Tfec Pax6 acts as an anti-retinogenic factor , whereas in conjunction with retinogenic genes it acts as a pro-retinogenic factor . The results suggest that careful manipulation of the Pax6 regulatory circuit may facilitate the generation of retinal and pigment epithelium cells from embryonic or induced pluripotent stem cells .
The vertebrate retinal pigment epithelium or RPE is a monolayer of melanin-containing cells that are intimately juxtaposed to the photoreceptor layer of the retina where they function as light screen and provide metabolic support for the retina's photoreceptors . Like the retina , the RPE is developmentally derived from the optic neuroepithelium and can be converted experimentally to retina , usually only during early development but in some species , such as newts , even in adults [1]–[3] . The differentiation of the optic neuroepithelium into precursors for retina and RPE is controlled by numerous growth factors and transcription factors . Among them is the microphthalmia-associated transcription factor MITF , which is a basic-helix-loop-helix-leucine zipper protein that together with the closely related proteins TFE3 , TFEB and TFEC forms the Mi-T subfamily of MYC-related transcription factors [4] , [5] . During mouse eye development , when the budding optic vesicle has not yet touched the surface ectoderm , low levels of MITF are found throughout the vesicle's epithelium . Mitf then is downregulated in the distal part of the vesicle , the future retina , in a pathway that involves FGF signaling and the paired-like homeodomain transcription factor VSX2 [6] , [7] , and it is upregulated in the proximal part , the future RPE , in a pathway that , based on work mostly in the chick , involves bone morphogenetic proteins ( BMPs ) , several transforming growth factor-ßs including ACTIVIN , and WNT signaling [8]–[12] . If Mitf is missing , an optic cup still forms and retina development continues , but the domain that normally becomes RPE hyperproliferates , remains unpigmented , and in its dorsal part turns into a second , fully laminated retina in a process commonly referred to as “transdifferentiation” [7] , [13] . Conversely , if Mitf fails to be downregulated in the distal optic vesicle , the domain that normally becomes retina hypo-proliferates and turns into a pigmented monolayer resembling the RPE [6] , [7] , [14] . A second transcription factor initially expressed in both distal and proximal optic vesicle is the paired domain protein PAX6 . Unlike MITF , PAX6 initially remains high in both future retina and RPE but later fades away in the RPE [14]–[16] . PAX6 has widespread roles in vertebrate eye development , both in structures derived from the surface ectoderm , such as cornea and lens , and in those derived from the neuroepithelium , such as iris , ciliary body , and retina [17] . Indeed , if in mice Pax6 is missing entirely , optic vesicles fail to form properly and eye development is aborted [18] . Conditional ablation specifically in the neuroretina allows for neuroretinal development to proceed but retinal progenitor cells can only mature into amacrine interneurons instead of the many different cell types of a normal retina [19] . In the RPE , however , the roles of Pax6 are less clear as an early RPE-specific ablation is not yet available and overexpression has no major effects on RPE development [20] , [21] . Nevertheless , studies involving chimeras between Pax6 wild-type and mutant mouse embryos showed that Pax6-deficient cells can contribute to the RPE although their differentiation ( pigmentation ) is impaired [22] . There are a number of previous studies that have suggested cross-regulations between Mitf and Pax6 . Biochemical studies , for instance , have indicated that PAX6 and MITF interact at the protein level , and that this interaction results in a mutual inactivation of both proteins [23] . In vitro , PAX6 stimulates at least one of the multiple Mitf promoters , the 5′ most human A-MITF promoter , and the combined absence of PAX6 and PAX2 , though not the absence of PAX6 alone , abolishes MITF expression in the optic neuroepithelium of mouse embryos as visualized by immunofluorescence [24] . Furthermore , a reduction of functional MITF in the RPE leads to an increase in PAX6 expression , most prominently in the transdifferentiating but also in the non-transdifferentiating portion [7] . Given the above findings , we reasoned that a decrease in Pax6 gene dose would alleviate RPE transdifferentiation in Mitf mutants , and upregulation further expand this transdifferentiation . Hence , we crossed mice carrying mutant Mitf alleles either with mice carrying the Pax6Sey-Neu allele , which represents a functional null allele [18] , or mice overexpressing human PAX6 from a yeast artificial chromosome ( YAC ) transgene [20] , [21] . Contrary to expectation , however , we find that a reduction of Pax6 enhances RPE transdifferentiation in Mitf mutants and overexpression suppresses it . In fact , overexpression of PAX6 increases the expression of the RPE-specific Mitf paralog Tfec which helps to compensate for the loss of the anti-proliferative , though not the pigmentary , function of Mitf . By gene expression profiling , we identified two major targets of the Pax6/Mitf/Tfec pathway that act extracellularly: Fgf15 , a member of the FGF family of ligands that serves as a positive regulator of retinal development , and Dkk3 , which together with FGFs inhibits RPE-promoting WNT signaling . Hence , our results show that despite the common neuroepithelial origin of retina and RPE , the two tissues differ in the way they utilize the transcription factor PAX6 . These findings may shed light on the evolution of Pax6 as a major regulator of eye development and may have implications for designing optimal methods to obtain RPE or retinal cells for cell-therapeutic purposes from embryonic or induced pluripotent stem cells .
As seen previously [7] , [15] , [16] and documented in Figure S1 , PAX6 and MITF protein are co-expressed in the developing RPE of mice , suggesting that the two genes interact . In fact , although the lack of Pax6 does not alter the onset of Mitf expression in the optic neuroepithelium [24] , the lack of Mitf results in increased Pax6 expression in the RPE [7] , [14] , prompting us to test whether manipulating Pax6 gene dose in embryos lacking Mitf would alter the development of the RPE . As shown in Figure 1A–1F , the future RPE of embryos homozygous for the Mitf null allele Mitfmi-vga9 showed dorsal thickening and PAX6 upregulation at E13 . 5 while heterozygotes showed no corresponding abnormalities ( for gene structures and alleles used in this study , see Table S1 ) . When Mitfmi-vga9 heterozygous embryos were also heterozygous for the non-functional Pax6Sey-Neu allele , however , they displayed a dorsal RPE thickening reminiscent of that found in Mitfmi-vga9 homozygotes ( compare Figure 1H , 1K with Figure 1C , 1F ) . Moreover , when Mitfmi-vga9 homozygotes were heterozygous for Pax6Sey-Neu , they displayed a massive enlargement of the RPE domain , which in its dorso-proximal part showed a morphology and expression pattern of the neuronal marker TUJ1 similar to those of the adjacent normal retina ( Figure 1I , 1L ) . In contrast , when PAX6 was overexpressed from a homozygous human PAX6 YAC transgene ( line Pax77; [21] , there was no RPE thickening , not even in Mitfmi-vga9 homozygotes , and this phenotypic correction persisted even after birth ( Figure 1M–1R ) . Similar observations were made at E11 . 5 ( Figure S2A–S2F ) . That the thickening of the RPE indeed resulted from cellular hyperproliferation is based on the observation that the percentage of phosphohistone H3-positive cells was significantly increased in RPEs with dorsal thickening ( Figure S2G–S2K ) while no changes in cell death were found ( not shown ) . These results indicate that in the RPE domain , Pax6 shares with Mitf an RPE-promoting function , in contrast to the retina domain where Pax6 is known to promote retina development [19] . Furthermore , because the Pax6-dependent RPE alterations were seen in the total absence of MITF protein , PAX6/MITF protein interactions as seen in vitro [23] cannot account for the observed phenotypes . To test whether the Pax6-dependent RPE alterations were specific to the Mitfmi-vga9 null allele , we repeated the above experiments with a newly generated RPE-specific allele of Mitf . We have previously found that of the many Mitf mRNA isoforms ( see Table S1 ) , the D-Mitf isoform is the only one specific to the RPE , and that this isoform contributes up to one third of total Mitf RNA in the RPE depending on the developmental time point [14] . Hence , by specifically eliminating the D-Mitf promoter/D-Mitf exon by targeted recombination , we expected to obtain mice with an RPE-specific hypomorphic allele of Mitf . The generation and characterization of such mice , labeled Mitfmi-ΔD/Mitfmi-ΔD or , for short , D-Mitf knock-outs , is documented in Figure 2 . To test for RNA expression in their developing eyes , we used microdissected embryonic or P0 RPE or retina for quantitative RT-PCR assays as previously described [14] ( Figure S3 ) . The assays showed that D-Mitf RNA was completely eliminated in the RPE fraction of D-Mitf knock-out embryos but that other isoforms , in particular H-Mitf , were compensatorily upregulated , though not before E13 . 5 ( Figure 2C ) . Consequently , total Mitf mRNA , quantified separately using primers common to all isoforms , was reduced only at E11 . 5 but not at E13 . 5 and thereafter ( Figure 2C , see also Figure 3N ) . This likely explains why D-Mitf knock out embryos show a slight reduction in the RPE expression of the MITF target gene Tyrosinase and in eye pigmentation only at E11 . 5 ( Figure 2D–2G ) , but normal eye development thereafter . In fact , on visual inspection , adult D-Mitf knock-outs look completely normal ( Figure 2H , 2I ) . In contrast to homozygous D-Mitf knock-outs carrying wild-type Pax6 alleles ( Figure 3A , 3D ) , homozygous D-Mitf knock-outs that were heterozygous for the Pax6Sey-Neu allele showed a dorsal RPE thickening at E11 . 5 ( Figure 3B , 3E ) , whereby some of the cells in the thickened RPE were positive for the retinal progenitor cell marker CD138 [25] ( Figure 3E , arrow ) . At E13 . 5 , the dorsal thickening became prominent ( Figure 3H ) , co-stained with TUJ1 ( Figure 3K ) , and resembled the transdifferentiating RPE domain of Mitfmi-vga9 homozygous mutants ( Figure 1C , 1F ) . No such RPE thickening or transdifferentiation were seen in D-Mitf knock-outs overexpressing PAX6 from the YAC transgene ( Figure 3C , 3F , 3I , 3L ) . Quantitative RT-PCR measurements of total Pax6 RNA , wild-type mouse-specific Pax6 RNA and pan-Mitf RNA in the E11 . 5 RPE fraction of the different mutant combinations are shown in Figure 3M , 3N . Interestingly , the presence of the Pax6Sey-Neu allele led to a reduction in wild-type mouse Pax6 RNA when Mitf was wild type , but to an increase when D-Mitf was eliminated . This increase was likely a consequence of the increase in expression of retinogenic genes including Pax6 in the transdifferentiating portion of the RPE ( Figure 3E , 3K; Figure S4 ) . Nevertheless , the expression of just PAX6 provided by the YAC transgene interfered with dorsal thickening or transdifferentiation . This phenomenon was likely linked to the differential regulation of the Mitf paralog Tfec ( see below ) . As documented in detail in Figure 4 and Figures S4 , S5 , S6 , S7 , the abnormal hyperproliferation of the dorsal RPE was associated with the expression of specific retinal markers and a concomitant downregulation of RPE markers ( Figure S4 ) . The abnormal RPE re-specification seems to begin in Pax6/Mitf double mutants already at E10 . 0–E10 . 25 , just when the RPE and retina become distinct domains . As shown in Figure S5 , Six3 , which is normally expressed in the future RPE domain at E9 . 5 but downregulated when the cells become committed at around E10 . 0 , continues to be expressed at low levels in this domain in both Pax6Sey-Neu/Pax6+; Mitfmi-ΔD/Mitfmi-ΔD and Pax6Sey-Neu/Pax6+; Mitfmi-vga9/Mitfmi-vga9 mutants . Nevertheless , the retina-specific marker Vsx2 starts to be expressed at this early time point only in the RPE domain of Pax6Sey-Neu/Pax6+;Mitfmi-vga9/Mitfmi-vga9 mutants and not in that of Pax6Sey-Neu/Pax6+;mi-ΔD/Mitfmi-ΔD . Hence , the assessment of whether in the double mutants an RPE fate is never initiated or is first initiated and then reversed depends on the retinal markers used and the particular allelic combination studied . In any event , however , the abnormal specification of the RPE domain as a future retina always lags behind the specification of the normal retina . For reasons of simplicity , we therefore use the term “transdifferentiation” not only for the single Mitf mutants but also for the double mutants . Despite the above mentioned exacerbation of the RPE-to-retina specification , dorso-ventral polarity was maintained because neither did a dorsal marker , Tbx5 , extend ventrally nor a ventral marker , Vax2 , extend dorsally in E13 . 5 Pax6Sey-Neu/Pax6+; Mitfmi-vga9/Mitfmi-vga9 double mutants ( Figure S6G–S6J ) . That the ventral RPE was unaffected in the Pax6/Mitf double mutants ( Figure S6I , S6J ) was likely due to the fact that Vax1/2 and Nr2f1/2 , two genes whose alterations lead to ventral and central RPE transdifferentiation [26] , [27] , were wild type in these mice . Furthermore , at P0 , the transdifferentiating RPE domain of Pax6Sey-Neu/Pax6+; Mitfmi-ΔD/Mitfmi-ΔD ( Figure 4 ) and Pax6Sey-Neu/Pax6+; Mitfmi-vga9/Mitfmi-vga9 mutants ( Figure S7 ) eventually generated differentiated retinal cells , including ISL1-positive ganglion cells , NF160-positive horizontal cells , and VC1 . 1-positive amacrine cells . They assumed a laminar structure similar to that seen in the normal retina , except that , as expected based on the developmental topology , the lamination was inverted relative to that of the retina [7] . In contrast , in Mitf mutants carrying the Pax6 YAC transgene , the RPE maintained its normal morphology as a monolayer . Taken together , the above results indicated that in two distinct Mitf mutants , a reduction in Pax6 gene dose promotes formation of a retina from the neuroepithelial domain that normally becomes RPE while an increase in Pax6 gene dose interferes with such a transition . It has previously been shown that the Mitf paralog Tfec is expressed in the developing RPE and that its expression is increased by homozygosity for Mitfmi-vga9 [28] . Tfec shares with Mitf a similar multi-promoter gene structure and encodes proteins with a basic-helix-loop-helix domain that is nearly identical with that of MITF [4] , suggesting the two proteins can form heterodimers and share an overlapping set of target genes . We confirmed the increased expression of Tfec in the RPE of E11 . 5 Mitfmi-vga9 homozygotes by in situ hybridization ( Figure 5E , 5F ) and determined by qRT-PCR that this increase was 2 . 5-fold ( data not shown ) . Tfec was , therefore , an excellent candidate gene that might participate in the above interplay between Mitf and Pax6 mutations . In fact , by in situ hybridization , Tfec was weakly upregulated in the RPE of E10 . 5 homozygous D-Mitf knock-outs ( Figure 5A , 5B ) , downregulated when they were heterozygous for Pax6Sey-Neu ( Figure 5C ) , and upregulated when they were transgenic for PAX6 YAC ( Figure 5D ) . Quantitative RT-PCR confirmed that Tfec was upregulated about 1 . 7-fold in the E11 . 5 RPE of homozygous D-Mitf knock-outs , slightly , but significantly less when they carried the Pax6Sey-Neu allele , and significantly more when they carried the PAX6 YAC transgene ( Figure 5I ) . Similar observations were made for Mitfmi-vga9/Pax6-mutant combinations by in situ hybridization even though the PAX6 YAC-mediated upregulation was not as striking as that in D-Mitf knock-outs ( Figure 5E–5H ) . These results suggest that MITF regulates Tfec negatively in the RPE while PAX6 regulates it positively . To directly demonstrate that TFEC can rescue eye defects in an Mitf mutant , we generated lines of transgenic mice that express a V5-tagged Tfec cDNA under control of an RPE-specific enhancer of the Tyrosinase gene [29] , and crossed them with microphthalmia red-eyed white ( Mitfmi-rw ) mice ( Figure 5J–5Q ) . We chose this hypomorphic Mitf allele ( rather than Mitfmi-vga9 ) because Mitfmi-rw homozygotes express small amounts of Tyrosinase and functional , though aberrant MITF protein , likely enough to stimulate the transgenic Tyrosinase enhancer , and yet they have an obvious small-eye phenotype [14] . Indeed , the Tyr-Tfec transgene was able to rescue eye size ( Figure 5L , 5M ) and RPE abnormalities ( Figure 5N–5Q ) of Mitfmi-rw homozygotes . Importantly , this rescue was not due to upregulation of the endogenous MITF ( Figure 5O ) . Only minimal rescue in eye size was observed with transgenic lines expressing TFEC at lower levels ( not shown ) . Hence , transgenic TFEC can replace MITF in the RPE at least with respect to its anti-proliferative functions . To analyze whether PAX6 and MITF regulate Tfec directly , we employed chromatin immunoprecipitation ( ChIP ) and reporter assays , using a 5 kbp Tfec enhancer/promoter sequence . Based on expression analysis of Tfec RNA isoforms [30] , this sequence corresponds to the upstream region of the only isoform we found expressed in RPE . It contains a number of conserved potential PAX6 and MITF binding sites that we arbitrarily grouped into six amplicons . ChIP analyses of wild-type RPE fractions showed PAX6 and MITF binding to amplicon I , PAX6-only binding to amplicon II ( Figure 5R , 5S ) , and binding of neither protein to the other four amplicons ( not shown ) . In reporter assays in the human RPE cell line ARPE19 , a 700 bp promoter element containing amplicons I and II was spontaneously active . To test whether this reporter fragment could be further stimulated by PAX6 , we expressed either one of two alternatively spliced Pax6 isoforms , Pax6 ( +5a ) and Pax6 ( −5a ) , which differ by a 14-amino acids insertion in the paired domain and exhibit unique DNA binding properties [22] , [23] . In fact , the Tfec promoter was stimulated by a vector expressing the PAX6 ( −5a ) but not the PAX6 ( +5a ) isoform . Transfection of a mutant Pax6 ( −5a ) cDNA that represents the Pax6Sey-Neu allele ( stop codon at amino acid position 301 ) ( 18 ) did not stimulate the Tfec promoter . Importantly , transfection of a D-Mitf expression vector reduced the activity of the Tfec promoter in a dose-dependent manner , suggesting that Mitf can indeed repress Tfec ( Figure 5T ) . These results suggest that in vivo , Tfec is regulated at least in part directly by PAX6/MITF and participates in the circuit that can partially compensate for the reduction or loss of MITF function , thus supporting Pax6's anti-retinogenic activities in the RPE . To gain deeper insights into the underlying mechanisms of PAX6/MITF/TFEC-mediated RPE transdifferentiation , we performed a microarray analysis on the Affymetrix platform ( Affymetrix Mouse Gene 1 . 0 ST ) . For this analysis , we used cDNAs prepared from E11 . 5 RPE fractions from the indicated D-Mitf knock-outs differing in their Pax6 gene dose , and from their respective controls . We centered the evaluation of up- and downregulated genes on Pax6Sey-Neu/Pax6+; Mitfmi-ΔD/Mitfmi-ΔD RPE fractions because their RPEs display the greatest extent of transdifferentiation . Table 1 and Table 2 show a selection of genes upregulated at least 1 . 49-fold in such RPE fractions and downregulated at least 1 . 43-fold when compared to wild-type RPE fractions ( for full data and technical details , see Table S2 ) . The microarray assay confirmed upregulation of retinal genes , including Vsx2 , Rax , Pax6 and Six6 , and downregulation of RPE genes , including Tyrp1 , Silv , Tyr and Mitf , and showed the expected changes of these genes under the additional genetic configurations . Importantly , among the many genes whose expression changes in Pax6Sey-Neu/Pax6+; Mitfmi-ΔD/Mitfmi-ΔD RPEs , we found prominent upregulation of two genes that encode extracellular ligands and that are potentially involved in eye development . One , Fgf15 , is a fibroblast growth factor whose paralogs have previously been shown to promote retinal at the expense of RPE development [31] , [32] . The other , Dkk3 , is a member of a family of genes involved in the inhibition of WNT signaling [33] , which is known to promote RPE development [8] , [11] . Dkk3's role in WNT signaling is not entirely clear , however , as in several cancer cell lines , WNT signaling is increased after Dkk3 downregulation [34] while in a Müller glia cell line ( though not in cos7 cells ) , it is increased after Dkk3 upregulation [35] . This suggests that Dkk3 acts in a context-dependent way , prompting us to focus specifically on FGF15 and DKK3 in RPE transdifferentiation . We first confirmed by qRT-PCR that Fgf15 and Dkk3 are indeed upregulated in RPE fractions of Pax6Sey-Neu/Pax6+; Mitfmi-ΔD/Mitfmi-ΔD embryos . As shown in Figure 6A , 6B , Fgf15 expression was increased approximately 15-fold and Dkk3 expression approximately four-fold compared to wild-type or Mitfmi-ΔD/Mitfmi-ΔD RPEs . Although gene expression profiling showed that another member of the Dkk family , Dkk1 , was also upregulated in Pax6Sey-Neu/Pax6+;Mitfmi-ΔD/Mitfmi-ΔD RPEs ( see Table S2 ) , it did not show the prominent Pax6 gene dose-dependent difference observed for Dkk3 and was therefore not further analyzed . Strong expression of both Fgf15 and Dkk3 was also seen by in situ hybridization in the transdifferentiating RPEs of Pax6Sey-Neu/Pax6+;Mitfmi-vga9/Mitfmi-vga9 embryos at E11 . 5 , but not in the RPEs of Pax6YAC/YAC;Mitfmi-vga9/Mitfmi-vga9 embryos ( Figure 6C–6N , arrows in F , L pointing to the transdifferentiating RPE ) and not at E9 . 5–E10 ( data not shown ) . These results suggest that PAX6 and MITF/TFEC together normally suppress Fgf15 and Dkk3 in the developing RPE . Nevertheless , single reductions of either PAX6 or MITF alone have only mild effects on Fgf15 or Dkk3 expression , consistent with their milder phenotypes ( Figure 6A–6N ) . The results also suggest that ectopically expressed Fgf15 and Dkk3 help to induce the RPE-to-retina transdifferentiation . We next used ChIP and luciferase reporter assays similar to those shown for the Tfec promoter to test whether the regulation of Fgf15 and Dkk3 by Pax6 and Mitf/Tfec is direct or indirect . Although a 5 kbp region upstream of the Fgf15 translational start site showed various conserved potential binding sites for PAX6 and MITF , only one region , represented by amplicon I , which contains one potential PAX6 and one potential MITF binding site ( Figure 6O , left side ) , gave ChIP signals on wild-type RPE fractions ( Figure 6P and data not shown ) . For Dkk3 , conserved potential binding sites for PAX6 or MITF , arbitrarily grouped into amplicons I and II ( Figure 6O , right side ) were present in a region between position −6026 and −6306 of the annotated start site of translation . ChIP assays with wild-type RPE fractions gave signals for both PAX6 and MITF with Dkk3 amplicon I , but only for MITF with amplicon II ( Figure 6P ) . Luciferase reporter assays in RPE cells ( ARPE19 ) showed that PAX6 , MITF and TFEC negatively regulate these sequence elements , though in the case of Fgf15 only when PAX6 and MITF , or PAX6 and TFEC , were co-expressed with the reporter ( Figure S8 ) . As done for the Tfec promoter , we then evaluated the chromatin status of these elements in the mutants , again using whole optic vesicles . Although part of the ChIP signal in optic vesicles was likely contributed by the high level of expression of Fgf15 and Dkk3 in the retina , there was a substantial increase in the anti-acetyl H3 ChIP signal in Pax6Sey-Neu/Pax6+; Mitfmi-vga9/Mitfmi-vga9 for Fgf15 amplicon I and Dkk3 amplicon II ( lane 3 in Figure 6Q ) when compared to the respective controls . Conversely , the anti-dimethyl H3K9 ChIP signal was only seen in wild-type and not in Pax6Sey-Neu/Pax6+; Mitfmi-vga9/Mitfmi-vga9 . These results suggest that the tested promoter regions are indeed subject to in vivo regulation by PAX6 and MITF and that this regulation is at least in part direct . The above results show that Dkk3 and Fgf15 are major targets of the circuits that regulate RPE development . To test whether DKK3 and FGF signaling are also actively involved in RPE transdifferentiation , we employed wild-type optic vesicle explant cultures into which beads soaked in human recombinant DKK3 or FGF2 ( which , like FGF15 or its human ortholog FGF19 , binds the same isoforms of all four main FGF receptors , [36] ) were implanted . To test if FGF and DKK3 signaling can induce transdifferentiation in RPE cells , we used explant cultures established at E10 . 0 , when the RPE fate is just specified . If left untreated or implanted with a bead coated with bovine serum albumin , these explant cultures normally develop within 48 hours into optic cups with a clearly pigmented RPE [Figure 7A , 15/15 cultures developing a pigmented RPE; see also [7]] . The implantation of beads soaked in human recombinant DKK3 at 0 . 65 or 1 µg/ml had little effect on RPE development ( Figure 7B , 0 . 65 µg/ml , 8/8 cultures with pigmented RPE ) . Therefore , we tested for cooperation between DKK3 and FGF signaling . Previous results showed that beads soaked in 1 µg/ml of FGF2 are capable of inducing RPE transdifferentiation on their own [6] , [7] . Nevertheless , a dose response curve indicated that implanting beads soaked in 0 . 35 µg/ml or less of FGF2 was without effect on pigmentation ( Figure 7C , 0 . 35 µg/ml , 10/10 cultures with pigmented RPE ) . Beads soaked in a mixture of 0 . 65 µg/ml of DKK3 and 0 . 35 µg/ml of FGF2 , however , markedly reduced pigmentation in the vicinity of the bead ( Figure 7D , 8/10 cultures showing segmental RPE depigmentation ) . Furthermore , in situ hybridization for Vsx2 and Six6 showed that only the double DKK3/FGF2 treatment led to expression of these retinal genes in the RPE ( Figure 7E–7L ) . To test whether DKK3 would inhibit the canonical WNT signaling pathway , we used optic vesicle cultures from mice transgenic for the WNT reporter Tcf-LacZ [37] . As shown in Figure 7M–7P , DKK3 beads alone could not repress ßGAL expression ( nor could , for that matter , FGF2 beads alone: 8/8 and 4/4 cultures , respectively , retained ßGAL staining ) . However , the combination of DKK3 and FGF2 at the above concentrations efficiently repressed ßGAL expression ( 5/7 cultures lacking ßGAL staining ) . The above results suggested that DKK3 and FGF cooperate to effect RPE transdifferentiation through the inhibition of WNT signaling and are sufficient to exert this effect . To test whether the Pax6/Mitf mutations ( which , as shown above , upregulate Dkk3 and Fgf15 ) also inhibit WNT signaling , we crossed the Tcf-LacZ transgene into Pax6Sey-Neu/Pax6+ mice heterozygous or homozygous for Mitfmi-vga9 . Double immunolabeling of E10 . 5 optic cups for VSX2 and ßGAL clearly showed that RPE transdifferentiation in Pax6Sey-Neu/Pax6+; Mitfmi-vga9/Mitf+ and Pax6Sey-Neu/Pax6+; Mitfmi-vga9/Mitfmi-vga9 embryos was associated with suppression of ßGAL expression , suggesting that enabling WNT signaling by inhibiting DKK3 and FGF is the common pathway through which the PAX6/MITF/TFEC regulatory circuit operates in the RPE ( Figure 7Q–7T ) .
Here we provide genetic evidence that the transcription factor PAX6 , which is known in vertebrates to be crucial for the development of cornea , iris and retina , is also critical for early RPE development when tested in an Mitf mutant background . In fact , we find that overexpression of PAX6 in the Mitfmi-vga9 null mutant background efficiently suppresses the RPE-to-retina transdifferentiation caused by Mitf-downregulation while a reduction in Pax6 enhances this transdifferentiation . Hence , in the RPE , Pax6 shares with Mitf an anti-retinogenic effect while in the retina it is pro-retinogenic . Although Pax6 and , for that matter , many other transcription factors have evolved to play different roles in different tissues , we have to keep in mind that future retina and RPE are both derived from the same , seemingly uniform optic neuroepithelium and maintain a remarkable capacity to switch from one into the other during development and , in some vertebrates , even in adulthood [2] . As outlined below , the regulatory circuit that we here describe and schematically depict in Figure 8 may provide an explanation for the differential function of Pax6 in retina and RPE and for the easy phenotypic switch between the two . There is ample evidence that the initial separation of the optic neuroepithelium into future retina and RPE is effected by extracellular signals , predominantly FGFs , that emanate from the surface ectoderm . FGFs induce Vsx2 in the distal optic vesicle domain , which in turn downregulates Mitf and Tfec and initiates the retinal developmental program [6] , [28] . In contrast , BMP , ACTIVIN and WNT signaling in or around the dorso-proximal optic vesicle may stimulate Mitf in the dorso-proximal neuroepithelium and initiate the RPE developmental program [8] , [10]–[12] . As soon as a slight bias in gene expression patterns is established to support either RPE or retina differentiation , the regulatory circuit shown in Figure 8 will reinforce this bias in the following way . Coexpression of Pax6 with either Mitf or Tfec suppresses the expression of the retinogenic genes Fgf15 and Dkk3 , thus enabling canonical WNT signaling , which in turn leads to upregulation of Mitf and Tfec expression . This positive feed back loop potentiates RPE differentiation and suppresses the retinal fate . If , on the other hand , PAX6 , MITF and TFEC together are unable to inhibit Fgf15 and Dkk3 expression , then canonical WNT signaling is inhibited and several retinogenic genes including Six6 , Lhx2 , and Vsx2 are upregulated . These retinogenic genes will further stimulate the expression of Pax6 in pro-retina and anti-RPE feed-forward loops [38] , and Mitf expression is suppressed . Likewise , Tfec expression is suppressed under these conditions because the repressive effect of other retinogenic genes overcomes Pax6's stimulatory effect on Tfec ( for instance , based on the analysis of Vsx2 mutants , Vsx2 suppresses Tfec in the retina despite the presence of Pax6 [6] , [28] . This regulatory circuit has the typical features of a bi-stable loop as it may assume only one of two states: either Mitf and/or Tfec are high and retinogenic genes ( with the exception of Pax6 ) are low , or Mitf and/or Tfec are low , and retinogenic genes are high ( Figure 8 ) . Consequently , the neuroepithelium will develop either as a retina or as an RPE . Perturbations of the circuit , such as mutations in Mitf [14] or Vsx2 [6] , [28] or gene dose changes in Pax6 could then flip the switch and lead to an inversion of cell fates rather than an indeterminate , mixed phenotype . Experimental evidence for the above consideration is also provided by the earlier observation that physical removal of the surface ectoderm in optic vesicle cultures , and hence removal of retina-inducing signals , does not lead to the formation of two RPE-like monolayers as one might predict , but rather to a distal pigmented monolayer and a proximal retina-like tissue [7] . As this happens without deliberate addition of a source of FGFs on the proximal side , we have to assume that removal of the surface ectoderm establishes a new initial bias that is again reinforced by the Pax6/Mitf/Tfec regulatory circuit . It is conceivable that this is so because removal of the surface ectoderm removes not only FGF signals but other signals as well . In the chick , for instance , the dorsal portion of the surface ectoderm expresses bone morphogenic protein-4 and -7 which help initiate RPE development [10] , and neural crest cells normally provide TGF-ß1 and ACTIVINS that stimulate WNT signaling in the dorsal ectoderm [9] , [12] . Perhaps even more intriguing is the recent observation that optic neuroepithelial vesicles established entirely in vitro from mouse embryonic stem cells spontaneously self-organize into optic cups with properly oriented retina and RPE , without any overt presence of a typical surface ectoderm [39] . This phenomenon , too , could be explained if any initial developmental bias is reinforced by the proposed bi-stable nature of the circuit . The above considerations , though , should not lead one to assume that the Pax6/Mitf/Tfec circuit is uninfluenced by other genes . A recent study showed , for instance , that in the absence of the neurogenic transcription factor SOX2 , Pax6 expression in the distal optic neuroepithelium increases , leading to conversion to ciliary epithelium but not RPE [40] . Perhaps this observation is linked to the fact that in Mitf single or Pax6/Mitf double mutants , the anterior most portion of the dorsal RPE never undergoes transdifferentiation into retina , likely because of the above mentioned RPE-inducing signals from the adjacent dorsal surface ectoderm [9] , [10] . Furthermore , what we observe in mice need not necessarily apply to other vertebrates . For instance , removal of the surface ectoderm from chicken optic vesicles led to a salt and pepper structure of intermingled pigmented and non-pigmented cells , not to cleanly separated pigmented monolayers and hyperproliferating retinas as seen in mice [41] . We would predict , therefore , that even though the molecular players may be the same across all vertebrates , they may not be interconnected in the same way in all vertebrates . In fact , the PAX6 binding sites that we identified in the mouse Tfec promoter , although conserved across several mammalian species , are not conserved in birds and reptiles . Such differences could well explain the differences in the developmental time frame during which RPE and retina can interconvert in different species . Finally , we would like to draw attention to the potential importance of the role of Pax6 , Mitf and Tfec and their regulatory targets for the establishment of retinal and RPE cells from embryonic or induced pluripotent stem cells . The ability to obtain such cells in vitro has recently generated much excitement for the study of the pathogenesis of blindness caused by primary retinal or RPE malfunctions , such as adult onset macular degeneration , as well as for the eventual cell-based therapy of such diseases [42] . Although the methods to generate such cells are rapidly improving , the process is still not very well controlled , and it remains to be seen whether the in vitro generated cells represent truly authentic cell types [42] . As the production of such cells has many of the hallmarks of development , we think a careful consideration of the normal developmental pathways is paramount for their successful generation . In the light of the results presented in this paper , it would seem important , therefore , that the activity levels of PAX6 and its upstream regulators and downstream targets be carefully monitored during the in vitro generation of retinal and RPE cells . In conclusion , we have shown that positioning of PAX6 in the center of a bi-stable regulatory loop allows this single transcription factor to be bi-functional and to participate either in a pro-retinogenic or a pro-RPE developmental pathway .
Extant Pax6 and Mitf mutants and transgenics are described in Table S1 and were kept on a C57BL/6J background ( Backcrosses: 20 for Pax6Sey-Neu , 2 for PAX6Yac , 7 for Mitfmi-vga9 , strain of origin of Mitfmi-rw is C57BL/6J ) . C57BL/6J served as wild-type controls . Mitfmi-ΔD/Mitfmi-ΔD targeted mice were generated using the recombineering technology . To generate the targeting construct , the D-Mitf promoter/D-Mitf exon and its flanking regions ( 15 , 801 kbp ) were cloned using plasmid rescue from BACRP23-9A13 . A floxed neomycin resistance expression cassette flanked by 200 bp of sequence flanking the D-Mitf promoter/D-Mitf exon was used to replace 5 . 8 kbp of the D-Mitf promoter/D-Mitf exon from the above plasmid and used for standard targeting of LC3 ES cells ( genotype [C57BL/6Nx129S6]F1 ) , giving 6 correctly targeted colonies/40 colonies tested . Of several germline transmitting lines , one , officially designated Mitftm3Arnh; MGI: 5050698 , was selected and crossed with C57BL/6J•129S4-Prm1-Cre deleter mice ( Jackson Laboratories , stock 003328 , backcrossed twice to C57BL/6J ) . Offspring lacking the neo-cassette ( Mitftm3 . 1Arnh; MGI: 5050699 ) were backcrossed to C57BL/6J twice and then bred to homozygosity with or without the corresponding Pax6 alleles or transgenes . Tyr-Tfec transgenic mice ( strain C57BL/6N ) were generated using a construct composed of ( 5′-3′ ) a 4721 bp Tyr RPE-specific enhancer , a 985 bp hsp70 minimal promoter [29] , and a 1203 bp V5-tagged Tfec cDNA . Four transgenic lines were obtained and used for crosses with Mitfmi-rw mice . Genotyping of mice was performed by Southern blot and/or PCR using primers shown in Table S3 . All animal experiments were covered by approved animal protocols . Immunostaining and in situ hybridizations were performed as described previously , using 16 µm thick coronal cryostat sections [14] . All eye sections are shown with the dorsal side up . Cell proliferation analysis was done by phosphohistone H3 staining of three different embryos . Immunostaining of adult mouse RPE flat-mounts was performed as described by [43] . A Zeiss LSM510 confocal ( Zeiss , Thornwood , NY ) and a Nikon E800 ( Nikon , Melville , NY ) microscope were used to record immunostainings . A Polyvar microscope ( Reichert Jung , Depew , NY ) was used for recording in situ hybridizations . For antibodies and in situ probes , see Tables S4 and S5 . For RNA expression , E9 . 5 and E10 . 5 optic vesicles or RPE and retinal fractions from E11 . 5–E18 . 5 embryos and P0 mice were prepared as described . RNA was extracted from pools of 20–40 individual samples and RT-PCR and real time PCR was performed according to previously published protocols [14] . For primers , see Table S1 . Statistical analysis was done using two-tailed Student's t-test . Statistical significance of the data is represented as ( ø ) non-significant; ( * ) p<0 . 05; ( ** ) p<0 . 01; and ( *** ) p<0 . 001 . For microarray analysis , the Affymetrix platform was used in collaboration with the NHGRI/NINDS microarray core facility and bioinformatics core . Pools of optic vesicles from ∼20 individual E10 . 5 eyes or RPE fractions from ∼30 individual E11 . 5 eyes were prepared as previously described [14] . ChIP assays using ChIP IT kit from Active Motif ( Carlsbad , CA ) was performed as described . Primers and antibodies are given in Tables S3 and S5 . Reporter assays were performed using ARPE19 ( CRL-2302 , ATCC , Manassas , VA ) and the dual luciferase assay kit ( Promega , Madison , WI ) . Statistical analysis was done using two-tailed Student's t-test . Details about the plasmids used for reporter assays are given in Table S3 . Cultures were established from E10 . 0 embryonic heads and maintained for 48–72 hours using published protocols [7] . Polyacrylamide beads used for implantation were soaked for an hour with BSA , recombinant DKK3 ( R&D Systems , Minneapolis , MN ) , recombinant FGF2 ( R&D Systems , Minneapolis , MN ) , or their combinations at the indicated concentrations . In situ hybridization/immunostaining of the vesicles was performed on 16 µm thick cryostat sections . | The retinal pigment epithelium or RPE in the back of the eye is critical for the normal function of the retina , and its abnormalities can lead to retinal disorders such as adult-onset macular degeneration . Insights into the pathogenesis of such disorders , and potential therapies , may come from using RPE cells generated in vitro from induced pluripotent stem cells . To obtain authentic RPE cells in vitro , we need to thoroughly understand the normal process of their development in vivo . Here we find that the potent retina-inducing transcription factor PAX6 plays a critical anti-retinogenic role in the RPE of mice . But how can PAX6 be pro-retinogenic in the retina and anti-retinogenic in the RPE ? To address this question , we used gene expression studies and combined them with chromatin immunoprecipitation assays , which analyze the interaction of transcription factors with chromatin in vivo . Our findings show that , in the RPE , PAX6 cooperates with either one ( or both ) of two related RPE transcription factors , MITF and TFEC , to suppress extracellular signals that in the normal retina induce a signaling cascade promoting retina formation . Hence , this study provides mechanistic insights into RPE development that may become important for the efficient generation of retina and RPE from induced pluripotent stem cells . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"developmental",
"biology",
"genetics",
"biology",
"molecular",
"cell",
"biology",
"neuroscience",
"genetics",
"and",
"genomics"
] | 2012 | A Regulatory Loop Involving PAX6, MITF, and WNT Signaling Controls Retinal Pigment Epithelium Development |
RNA polyadenylation ( pA ) is one of the major steps in regulation of gene expression at the posttranscriptional level . In this report , a genome landscape of pA sites of viral transcripts in B lymphocytes with Kaposi sarcoma-associated herpesvirus ( KSHV ) infection was constructed using a modified PA-seq strategy . We identified 67 unique pA sites , of which 55 could be assigned for expression of annotated ∼90 KSHV genes . Among the assigned pA sites , twenty are for expression of individual single genes and the rest for multiple genes ( average 2 . 7 genes per pA site ) in cluster-gene loci of the genome . A few novel viral pA sites that could not be assigned to any known KSHV genes are often positioned in the antisense strand to ORF8 , ORF21 , ORF34 , K8 and ORF50 , and their associated antisense mRNAs to ORF21 , ORF34 and K8 could be verified by 3′RACE . The usage of each mapped pA site correlates to its peak size , the larger ( broad and wide ) peak size , the more usage and thus , the higher expression of the pA site-associated gene ( s ) . Similar to mammalian transcripts , KSHV RNA polyadenylation employs two major poly ( A ) signals , AAUAAA and AUUAAA , and is regulated by conservation of cis-elements flanking the mapped pA sites . Moreover , we found two or more alternative pA sites downstream of ORF54 , K2 ( vIL6 ) , K9 ( vIRF1 ) , K10 . 5 ( vIRF3 ) , K11 ( vIRF2 ) , K12 ( Kaposin A ) , T1 . 5 , and PAN genes and experimentally validated the alternative polyadenylation for the expression of KSHV ORF54 , K11 , and T1 . 5 transcripts . Together , our data provide not only a comprehensive pA site landscape for understanding KSHV genome structure and gene expression , but also the first evidence of alternative polyadenylation as another layer of posttranscriptional regulation in viral gene expression .
Kaposi sarcoma-associated herpesvirus ( KSHV ) , also referred as human herpesvirus 8 ( HHV-8 ) , is a member of gammaherpervirus subfamily [1] . KSHV infection in healthy individuals is well-controlled by host immune system and other host factors , and hence , it is usually asymptomatic . However prolonged immunosuppression may lead to occurrence of KSHV-induced malignancies . KSHV has been linked to three malignancies including all forms of Kaposi sarcoma , a complex solid tumor of endothelial origin , and two rare B-cell lymphomas , primary effusion lymphoma ( PEL or body cavity-based large cell lymphoma ) and multicentric Castleman disease [2]–[4] . KSHV , like other herpesviruses , exhibits two distinguishable states of infection , latent and lytic infection . During latency only a small fraction of viral genes are expressed to facilitate the maintenance of viral genome , drive cell proliferation , and mediate immune invasion . Various external and internal stimuli cause a disruption of KSHV latency and induction of KSHV lytic infection with the expression of all viral lytic genes and replication of viral progeny [5]–[7] . KSHV has a large DNA genome ( ∼168 kb ) encoding more than 90 genes for production of viral structural and non-structural proteins , small peptides , long non-coding RNAs ( lncRNAs ) and small regulatory miRNAs [1] , [8]–[10] . Like many DNA viruses , KSHV has a complex gene organization and depends on host cell machinery for its gene expression . However , a full compendium of viral genome annotation is still unknown and the true nature of viral gene expression and its regulation remains to be fully understood . RNA polyadenylation ( pA ) of nascent transcripts is a critical posttranscriptional step in maturation of eukaryotic transcripts [11] . A primary role of RNA polyadenylation is to release newly synthesized RNA from DNA template through endonuclease cleavage and protect it from degradation by addition of a poly ( A ) tail to the RNA 3′ end . The presence of a poly ( A ) tail also promotes nucleocytoplasmic export and efficient protein translation of mRNAs [12] , [13] . RNA polyadenylation is carried out by a large protein complex composed of at least 85 protein factors and binds to specific sequences within nascent transcripts surrounding the cleavage site [14] . An A/U-rich element upstream , recognized by cleavage and polyadenylation specificity factor ( CPSF ) , and an U/GU-rich element downstream , recognized by cleavage stimulatory factor ( CstF ) [15] , [16] , of the cleavage site are two major determinants of RNA polyadenylation , although other auxiliary cis-elements may be also involved in pA site definition [17] , [18] . After assembly of polyadenylation complex the pre-mRNA is generally cleaved at “CA” dinucleotide followed by addition of a poly ( A ) tail [19] . While the process of polyadenylation itself is well characterized , the selection of pA site remains a puzzle . Recent genome-wide studies on polyadenylation of host transcripts in various organisms revealed highly promiscuous polyadenylation of large population of RNAs from multiple pA sites [20]–[22] . As a result , genes affected by alternative polyadenylation produce a subset of transcripts with different coding potentials or 3′ untranslated regions ( 3′ UTRs ) [23] . A polyadenylation landscape of human herpesviruses has not been reported at the genome-wide level . In this study , we performed a genome-wide analysis on RNA polyadenylation of KSHV transcripts from B cells with latent or lytic viral infection by using a modified polyadenylation- sequencing ( PA-seq ) technology [24] . We identified that KSHV utilizes 67 active pA sites for the expression of its latent and lytic genes and a few novel or unannotated genes . We also found alternative RNA polyadenylation of several known KSHV genes and revealed pA site cis-elements in regulation of KSHV gene expression .
To elucidate a role of RNA polyadenylation in KHSV gene regulation , we performed a genome-wide analysis of viral pA sites to monitor their usage during KSHV infection . Three KSHV-positive B-cell lines ( JSC-1 , BCBL-1 and TREx BCBL-1 ) , which support both latent and lytic virus infection , were chosen in this study . For each cell line , polyadenylation events were compared between latent and lytic infection ( Figure S1A ) . The cells with lytic infection were harvested at 48 h after virus reactivation by chemical induction to allow full viral replication cycle and sufficient expression of viral late transcripts . We observed dramatic reduction of cell viability associated with virus reactivation ( 31% vs 88% for JSC-1 , 50% vs 89% for BCBL-1 and 20% vs 82% for TREx BCBL-1 cells [further referred as TREx] ) by trypan blue exclusion analysis . Poly ( A ) + RNA fraction from each sample was used for preparing 3′-end cDNA libraries with a modified PA-seq method followed by Illumina paired-end sequencing [24] , [25] . In total , we obtained more than 119 million of paired reads from all samples ( Figure S1B ) . KSHV- and human-specific reads were extracted by alignment of obtained sequence reads to the reference KSHV ( GenBank acc no U75698 . 1 ) and human ( UCSC version hg19 ) genomes . More than 100 million ( ∼84% ) of all reads were uniquely mapped , with about 35 million ( ∼29% ) to KSHV and approximately 65 million ( ∼55% ) to human genome . The remaining 19 million ( 16% ) are unmapped reads . As expected , a remarkable correlation was noticed between KSHV-specific reads and the stat of KSHV infection in all three cell lines , with less KSHV reads ( 0 . 10–0 . 47% ) in the cells with viral latent infection and much more KSHV reads ( 20–77% ) in the cells with KSHV lytic infection ( Figure S1B and S1C ) . For KSHV pA site analysis we focused only on the sequence reads uniquely mapped to KSHV genome and further clustered to identify PA-peaks . We pooled KSHV sequence reads from all samples and performed a peak calling analysis using F-seq algorithm with significant enrichment over a background model [26] and a threshold of >50 read counts per peak ( Figure S2 ) . As expected , only a handful of PA-peaks were found in the cells with latent KSHV infection , but significantly more peaks were detected in the cells with lytic KSHV infection ( Figure S3A ) . To further analyze the PA-peak distributions in the context of selected viral genes and to ensure the PA-peaks obtained from our PA-seq representative of authentic pA site regions , we looked into a PA-peak distributed in a well-characterized ORF50 ( RTA ) /K8/K8 . 1 locus which encodes three collinear KSHV genes . Although each gene in the locus has its own promoter , their transcripts are all polyadenylated at a single pA site downstream of K8 . 1 ( see diagram in Figure S3B ) . We found a prominent PA-peak in all lytic samples , but less so in the latent samples , overlapping with the mapped pA site ( Figure S3B ) reported in previous studies [27] , [28] . No other PA-peaks were seen either upstream or downstream of this pA site . These data indicate that PA-seq libraries were in high quality and suitable for comprehensive analysis of viral pA sites . Subsequently , we determined the nucleotide ( nt ) position with the highest number of reads within individual peaks in the peak calling analysis as a PA mode ( Figure S2 ) and designated such a PA mode as an unique pA site ( Table S1 ) . We also determined a PA peak from the peak start to the peak end , and the total number of the reads within a PA peak was used to approximate the usage of a pA site ( Figure S2 ) . With this approach we identified 67 pA sites on both viral DNA strands of the KSHV genome ( Figure 1A ) . The pA sites mapped by PA-seq in this study were remarkably close to several known pA sites previously mapped by traditional methods both in terms of mapped nucleotide position and strand specificity ( Table S2 ) . A higher prevalence of pA sites shows strand bias , with 43 pA sites in the minus strand and 24 in the plus strand of the KSHV genome . The majority of the mapped pA sites are positioned in the intergenic regions of KSHV genome , outside of annotated ORFs , with exception of the pA sites in the coding regions of ORF7 at nt 7032 and ORF61 at nt 98274 and of K12 at nt 118012 , 118032 and 118087 . Our genome-wide pA site analysis allowed us to correlate each mapped pA site to annotated KSHV genes and to identify novel KSHV gene ( s ) . We assigned each pA site to a known viral gene or gene cluster region based on the following criteria: ( 1 ) both of the gene ( s ) and the corresponding pA site must be on the same strand of viral genome , ( 2 ) the pA site must be positioned outside of the coding region of the viral gene ( s ) , and ( 3 ) the gene ( s ) assigned to a mapped pA site must be positioned upstream of the pA site . These criteria assume that viral transcripts originated from a promoter ( s ) upstream of the gene will be polyadenylated from the first available pA site downstream . Accordingly , we assigned 55 pA sites to all known KSHV genes ( Figure 1B , Table S3 ) . The remaining 12 pA sites unable to assign would indicate the presence of transcripts from unknown KSHV genes for further validation . Interestingly , the majority of unassigned pA sites are positioned antisense to known KSHV genes , suggesting the existence of putative antisense transcripts to these viral genes [29] . Among 55 pA sites assigned to known KSHV gene transcripts , 20 are positioned immediately downstream of a single KSHV gene for polyadenylation of a monocistronic mRNA , and the remaining 35 have multiple upstream KSHV genes ranging from 2 to 5 for polyadenylation of bicistronic or polycistronic transcripts ( Figure 1B ) . Interestingly , we found two or more pA sites mapped to a region downstream of the same gene . These include two alternative pA sites downstream of ORF54 , K2 ( vIL6 ) , K9 ( vIRF1 ) , K10 . 5 ( vIRF3 ) , K11 ( vIRF2 ) , and K12 ( Kaposin A ) , three downstream of T1 . 5 RNA and PAN ( nut-1 ) RNA or in an internal K12 region , and five downstream of vnct internal repeats ( Figure 1A ) . From protein-coding genes , ORF54 showed the highest usage of alternative pA sites ( ∼24% ) followed by K10 . 5 ( ∼17% ) and K11 ( ∼11% ) , but K2 , K9 , and K12 did so much less frequently ( Table S4 ) . Thus , our analyses provide not only the first comprehensive landscape of functional pA sites in the context of KSHV genome , but for the first time the alternative polyadenylation of KSHV transcripts during virus infection . We next aimed to determine the length and composition of 3′ UTR for each KSHV protein-coding gene . The unassigned pA sites and the pA site for viral non-coding RNA genes were excluded from this analysis . A total of 50 pA sites were used to calculate the 3′UTR length from a pA site to the adjacent termination codon of the closest upstream ORF . We found that the calculated 3′ UTR length of KSHV genes varies greatly in size from 2 nts ( ORF38 ) to 1925 nts ( ORF62 ) ( Table S3 ) . The distribution of KSHV 3′UTR is shown in Figure 1C , with a median size of the 3′UTR in ∼80 nts which is significantly shorter than human 3′ UTR with a median size of ∼300 nts [20] . Based on the number of sequence reads obtained at each pA site , one can infer the relative steady-state level ( pA site usage ) of the pA site-associated transcripts . The limitation of this approach is cluster genes utilizing a single pA site , in which the number of sequence reads reflects a combined level of all gene transcripts . The pA site usage was compared from latent to lytic infection . First , we determined each pA site usage in individual samples to obtain a sample-specific pA site usage and then normalized the number of sequence reads within each viral pA peak to the total sequence reads mapped to KSHV and host genome in each sample ( Figure S4 , Table S5 ) . Combination of the normalized sequence reads from all latent samples was compared with that from all lytic samples ( Figure 2A , 2B , Table S6 ) . The pA site 122069 ( + ) of latent polycistronic RNA ORF73 ( LANA ) , ORF72 ( vCyclin ) , and K13 ( vFLICE ) was served as a reference ( red bar in Figure 2A–C ) . Surprisingly in the samples with latent infection , the top 5 sites based on the pA site sequence counts were PAN ( nut-1 ) , ORF2/K2 cluster , K12 , ORF50/K8/K8 . 1 cluster , and T1 . 5 ( Figure 2A ) which supposed to be KSHV lytic genes , but spontaneously reactivated in a small fraction of cells with latent infection ( Figure S3 , Table S6 ) . The usage of pA site for the expression of ORF73/ORF72/K13 ranked the 6th during the latency . In the samples with lytic infection , the top 5 pA site usage was the pA sites for abundant expression of PAN , K12 , ORF62-58 cluster , T1 . 5 , and K4 . 2/4 . 1/4 cluster and usage of the reference latent pA site for expression of ORF73/ORF72/K13 dropped to the 41st . When the changes in utilization of each identified pA site from lytic to latent infection were calculated , however , the usage of mapped pA sites for virus lytic gene expression became remarkable , with more than 500-fold increase from latent to lytic infection for ORF62 , ORF24/23 , ORF44 , PAN , and K12 ( Figure 2C , Table S6 ) . The smallest usage change ( <10 fold ) during virus lytic infection was the pA sites for the expression of ORF2/K2 , K10 . 6/10 . 5 , and ORF73/ORF72/K13 , and the transcripts antisense to ORF50 ( RTA ) and K15/ORF75 ( Figure 2C , inset ) . The smallest change in the pA site usage was the reference pA site of ORF73/ORF72/K13 , with only 2 . 1-fold increase . Thus , these pA sites are truly used to express viral latent genes . Notably , the peak sizes of pA sites vary considerably ranging from 3 ( pA site at nt 17227 ) to 98 nts ( pA site at 29740 ) ( Table S7 ) , and represents heterogeneity of the cleavage sites within each mapped pA site [30] . This heterogeneity of a given pA site , such as the pA site at nt 29740 , nt 76738 or nt 122069 , remained invariable either from JSC-1 to BCBL-1 cells or from latent to lytic infection ( data not shown ) . Based on their bimodal distribution ( Figure 3A ) , we grouped 38 pA sites with a narrow peak ( ≤30 nts , with a median size of 17 nts ) , 24 pA sites with a broad peak ( >30 , ≤45 nts , with a median size of 36 . 5 nts ) , and 5 pA sites with a wide peak ( >45 nts , with a median size 61 nts ) ( Figure 3A ) . Interestingly , we found a strong positive correlation of pA site usage by sequence reads in the order from narrow ( 4 , 013 reads ) , broad ( 62 , 764 reads ) , to wide ( 425 , 475 reads ) peaks with Spearman correlation coefficient rs = 0 . 86 ( Figure 3B , Table S8 ) . Because each transcript could produce only one read in PA-seq , whereas RNA-seq relies on the read coverage on the entire region of a transcript , the read count in a given PA peak of a pA site simply reflects the abundance of the corresponding transcript . To investigate the regulatory elements responsible for polyadenylation of KSHV viral transcripts , we analyzed flanking sequences ( ±50 nts ) of all 67 pA sites identified . Prevalence of each nucleotide at individual position was calculated and followed by motif analysis using WebLogo software ( Figure 4A ) . A high prevalence of “A” residues between 10 to 30 nts upstream of the cleavage site was identified , representing the upstream A/U-rich polyadenylation signal . The cleavage site itself was also enriched in A residues , followed by a ∼30 nt long , mostly U-rich element . This distribution of RNA cis-elements around viral pA sites is in agreement with what has been found in human transcripts [16] . To better understand the role of cis-elements in regulation of KSHV polyadenylation , we performed similar analyses separately for three groups of pA sites with a narrow , broad , or wide peak ( Figure 4A ) . The profiles of pA sites with a narrow and broad peak showed the highest similarity to the canonical pA site , with a defined upstream A-rich and a downstream U-rich element . The pA sites with a broad peak also exhibit a U-rich region further upstream . However , there is no significant U-rich element downstream of the pA site with a wide peak , nor other sequence motifs could be seen . These differences in sequence context surrounding the pA sites with different peaks could devote to their notable abundance of the associated transcripts , and was further reiterated by analysis of top 10 pA sites with the highest numbers of sequence reads and bottom 10 pA sites with the lowest number of sequence reads among all 67 pA sites . As shown in Figure 4B , the top 10 pA sites show highly conserved polyadenylation signals upstream and an U-rich region downstream . In contrast , the bottom 10 pA sites only exhibit less conserved polyadenylation signals and lack an U-rich region downstream . Analysis of upstream poly ( A ) signal ( PAS ) strength of KSHV pA sites further reaffirmed this conclusion . The canonical ( AAUAAA ) and non-canonical ( NNAUNA ) PAS were identified within 50 nts upstream of the mapped 59 pA sites ( Figure 5A , Table S9 ) . Two most common PAS in KSHV as seen in human polyadenylation are canonical AAUAAA ( 69% ) followed by AUUAAA ( 9% ) . The usage of other non-canonical PAS for viral RNA polyadenylation ranges from 1% to 3% ( Figure 5A , Table S10 ) . Similar to human transcripts [20] , about 12% of pA sites mapped in this study have no PAS . Surprisingly , we found that most of the non-canonical PAS were associated with a narrow peak and low level of expression . In contrast , the broad and wide peaks use predominantly canonical AAUAAA and AUUAAA PAS . This became even more obvious with PAS in the top and bottom used 10 pA sites . We found that all top 10 pA sites , but only 60% of the bottom 10 pA sites , contain the canonical AAUAAA ( Figure 5B , Table S10 ) . Given that the pA sites obtained by PA-seq , in general , showed a high correlation with previously mapped KSHV pA sites , we carried out a series of experiments to reconfirm several novel pA sites discovered in this study by 3′ RACE . These include the pA site downstream of ORF27 , the pA site mapped within the coding region of ORF61 , the alternative pA sites downstream of ORF54 and T1 . 5 , and a cluster of 5 unassigned pA sites downstream of the vnct internal repeats , in addition to the known pA sites and unknown alternative pA sites of K11 , K2/vIL6 , and K12 . Most of the selected pA sites determined by PA-seq were verified by sequencing the expected 3′ RACE products in the predicted size ( s ) ( Figure 6 , Table S11 ) . The alternative pA site at nt 25192 within T1 . 5 lncRNA was not experimentally confirmed because of its <1% usage among T1 . 5 transcripts and lack of a searchable PAS upstream , nor the alternative pA sites at nt 17227 for K2/vIL6 and at nt 117868 for K12 because of their lower level usage . We were also unable to detect any 3′RACE product in the predicted sizes from five pA sites downstream of vnct internal repeats , despite their moderate usage based on the number of associated read counts . These pA sites identified by PA-seq are in proximity to the short internal 13-bp repeats region “vnct” between nt 29775 and nt 29942 of the KSHV genome . It is worth noting that four of the five pA sites have no detectable PAS upstream and a pA site at the nt 29615 has a non-canonical AAUAUA PAS . A reported pA site at nt 18200 for an RNA antisense to K2 ( vIL6 ) [29] , [31] which was not revealed by PA-seq was also not detectable by 3′ RACE in this study ( Figure 6 ) . We further verified the PA-seq-identified pA sites from the mRNAs antisense to ORF21 , ORF34 , and ORF K8 by 3′ RACE and confirmed the production of the antisense RNAs in B cells during viral lytic infection ( Figure 7A ) . The read abundance of these novel pA sites associated with each antisense RNA was correlated , as predicted , to the amount ( measured by band intensity ) of the 3′ RACE products derived from its corresponding RNA transcript ( Figure 7B ) . KSHV T1 . 5 RNA is a long non-coding RNA , which is transcribed from nt 24243 in the KSHV genome , next to the left lytic origin of replication ( oriL ) ( Figure 8A ) . The expression of T1 . 5 RNA is strongly inducible by viral transactivator RTA [32] . While the expression of T1 . 5 is required for viral DNA replication , its functional characteristics remain unknown [33] . Our study showed that T1 . 5 RNA is one of the most abundant transcript expressed during KSHV infection . The T1 . 5 RNA 3′ end was mapped to nt 25440 [34] and we mapped it to nt 25441 by PA-seq and by 3′RACE ( Figure 6 ) . In addition , we found that about 10% of T1 . 5 transcripts were also polyadenylated from two additional pA sites upstream of the nt 22541 pA site ( Figure 8A ) , leading to the production of ∼300 nts shorter transcripts as verified by Northern blot analysis of BCBL-1 total RNA ( Figure 8B ) . Because the antisense probe used in the assay was derived from a upstream region of the mapped pA sites , this probe could detect all transcripts running over this region: a strong band corresponding to the reported size of the inducible T1 . 5 RNA , a smaller size band ( ∼1 . 2 kb ) with weaker intensity representing the alternatively polyadenylated T1 . 5 transcripts , and a much larger T6 . 1 transcript . The T6 . 1 RNA does not use T1 . 5 pA sites [34] , but rather a PAN pA site for its expression ( Figure 1A ) . T1 . 5 RNA contains a few short ORFs and has potential to encode small peptides [34] . We thus assumed that T1 . 5 might be exportable to the cytoplasm . As expected , we demonstrated by Northern blot analysis its partial presence in the cytoplasm ( Figure 8B ) . RNA FISH assays further showed T1 . 5 RNA distribution both in the cytoplasm and nucleus of KSHV infected PEL cells using an antisense RNA probe to the 3′ end of T1 . 5 ( Figure 8A , Figure S5 ) . In these two assays , nuclear PAN RNA served as a control ( Figure 8B–C , Figure S5 ) and displayed , as expected , predominantly in the nucleus overlapping with Hoechst DNA staining [8] , [35] . Interestingly , the nuclear coexpression of T1 . 5 and PAN RNA appears mutually exclusive . We found that the cells expressing high level of nuclear T1 . 5 RNA display much less nuclear PAN RNA or vice versa ( Figure 8C ) . Compared with the subcellular distribution profile of PAN RNA , we saw more B cells with both cytoplasmic and nuclear distribution of T1 . 5 RNA during virus lytic infection ( Figure 8D ) . The usefulness of PA-seq was further extended to examine the expression of a few host genes for its possible application to unveil a pA site landscape of the host genome before and after KSHV lytic infection . Human IL6 ( hIL6 ) and GAPDH were initially chosen because B cells with lytic KSHV infection exhibit increased expression of human IL6 [36] , [37] , but decreased expression of GAPDH ( Figure 8B ) . As shown in Figure 9 , the results from PA-seq on GAPDH and hIL6 were comparable with that from RT-qPCR . The decreased expression of GAPDH RNA could be found by both methods in all three tested B cell lines with KHSV lytic infection and a significant increase of hIL6 expression in TREx BCBL-1 cells with lytic KSHV infection . However , we did not see in either method an increased hIL6 expression in JSC-1 cells with butyrate ( a very potent inducer ) -induced KSHV and EBV lytic coinfections , but observed the increased hIL6 expression in BCBL-1 cells with valproate ( a weak inducer ) -induced lytic KSHV infection by RT-qPCR . Human IL6 is a cytokine highly sensitive to ( vulnerable for ) RNA degradation and PA-seq detects the transcripts carrying an intact 3′-end poly ( A ) tail , while RT-qPCR detects only a small region of the IL6 RNA . Thus , multiple factors could contribute to the variations in detection of hIL6 gene expression from one cell line to another .
In this report we present the first viral genome landscape of polyadenylation sites from three PEL cell lines with KSHV latent or lytic infection . The comprehensive pA site landscape for the entire KSHV genome was revealed by using a modified PA-seq strategy which conveys single nucleotide resolution and strand specificity [24] , [25] . The mapped pA sites have been annotated to all known KSHV genes and four putative novel genes in the KSHV genome . The steady-state expression level of every gene in the KSHV genome from viral latent to lytic infection was quantified by PA-seq reads associated with each mapped pA site and was used to distinguish viral latent genes from lytic genes . By analyzing the flanking sequences of each mapped pA site , we determined the regulatory elements governing viral RNA polyadenylation and gene expression . More importantly , we identified several viral genes utilizing alternative polyadenylation as a mechanism for their expression during KSHV infection . In general , the mapped viral pA sites in this study have high accuracy both in terms of nucleotide position and strand orientation , when compared with the known viral pA sites identified by the conventional methods ( Table S3 ) [8] , [34] , [38]–[42] . However , we were unable to verify a few pA sites previously reported in other studies , including a pA site at nt 124061 for a C-terminal truncated LANA [43] and a pA site at nt 18200 ( + ) for the expression of a 0 . 7 kb transcript antisense to K2 ( vIL-6 ) [29] , [31] . The 0 . 7 kb transcript was discovered using custom-made tiling arrays covering the entire KSHV genome [29] , [31] and a T7-Oligo ( dT ) primer for sample cDNA synthesis . The likelihood internal priming of the oligo primer used in the study might create aberrant synthesis of cDNA probes hybridizing to the tiling arrays . In our study the detection of any pseudo pA sites resulting from internal priming was largely avoided by exclusion of the sequence reads upstream of an A-stretch in the KSHV genome . It is worth noting that the expression of the 0 . 7 kb transcript antisense to K2 was originally discovered only in viral lytically-infected endothelial iSLK . 219 cells derived from Kaposi sarcoma , but not detected in PEL-derived B cells [29] . In addition to assigning the known pA sites and many novel viral pA sites from this study to the KSHV genes being previously annotated , we also identified a few novel viral pA sites ( Table S3 ) that could not be assigned to any known KSHV genes . These unassigned pA sites are often found in the opposite strand to known KSHV genes , including ORF8 , ORF21 , ORF34 , K8 and ORF50 . Some of those antisense transcripts were described in other reports [29] and the existence of these RNAs antisense to ORF21 , ORF34 , and ORF K8 transcripts could be confirmed by 3′ RACE in this study ( Figure 7 ) . Their potential roles in KSHV biology are now under active investigation . KSHV has been evolved to use one pA site for the expression of multiple genes in many regions of the genome . Supporting this notion , our PA-seq analysis identified numerous regions of the KSHV genome with several viral genes ( up to 5 genes ) sharing a common pA site ( Figure 1B ) . As a consequence , many KSHV genes are expressed as bicistronic or polycistronic transcripts with a long 3′ UTR covering the coding region ( s ) of downstream gene ( s ) . These RNA structures are vulnerable to viral and cellular miRNAs [44]–[46] and all transcripts from the gene cluster regions could be regulated even by a single miRNA . Others could avoid this regulation by RNA splicing of the downstream ORF ( s ) as shown in ORF50/K8/K8 . 1 and K1 transcript [28] , [47] , [48] . Thus , understanding the gene organization and pA site position is critical for knocking-out or knocking-down studies of various virus genes from the KSHV genome in order to make appropriate interpretation on the function of individual viral genes in a cluster region . The usage of each mapped pA site in this study was determined by counting the sequence reads associated with each pA site to approximate the steady-state expression level of the associated gene ( s ) . When the sequence-reads of a given pA site in viral lytic infection were compared with that in viral latent infection , we could distinguish pA site usage from viral lytic genes to viral latent genes . The pA site for lytic gene expression could be used 100-fold more in lytic infection than in latent infection , whereas the pA site usage for latent gene expression displays only little increase ( less than 10-folds ) in lytic infection . Two pA sites downstream of K12 , a classical viral latent gene , could be an exception because both showed an increased usage in viral lytic infection . The increased usage of two K12 pA sites is consistent with the finding that a lytic inducible promoter could be activated for K12 expression [49] . Analysis of pA site usage in lytic viral infection also confirmed PAN RNA being an extremely abundant RNA species , with sequence-read counts in the mapped pA site at nt 29740 ( + ) from viral lytic infection alone representing more than 80% of the total sequence-reads for all pA sites . Moreover , the efficient expression of viral RNA transcripts was found being related to the peak size of a pA site in this study . It should note that each transcript could give rise to only one read in PA-seq . Thus , the read count in a PA peak simply reflects the abundance of the corresponding transcript . In fact , more sequence reads are not expected to inflate the size of a PA peak , especially when the pA cleavage events are precise . Therefore , the positive correlation we observed between the sizes of PA peaks and the expression levels of corresponding transcripts may suggest some degree of “slippage” in polyadenylation of viral transcripts to ensure high-level expression at the lytic stage . When compared to the pA sites falling into a broad or wide peak , an RNA transcript carrying a pA site with a narrow peak was less expressed , with fewer PA-seq sequence reads . Although this difference in the pA sites with a narrow peak might be attributable partially to their frequent usage of non-canonical PAS , there must be other unknown mechanisms governing the utilization of a pA site with a narrow peak , other than canonical vs non-canonical PAS per se . Previous reports showed that the PAS strength directly affects the overall level of mature transcripts [50] , [51] and is determined by conservation of RNA cis-elements UGUAN upstream and an run of U/G downstream of the PAS AAUAAA . For example , the presence of a weaker early SV40 PAS leads to lower expression of a reporter gene than the construct containing a stronger SV40 late PAS when both were driven by the same promoter [52] . Therefore , the PAS strength governing polyadenylation of individual viral transcripts may provide additional level of regulation to fine tune their proper expression during viral infection . In addition , the length of the 3′ UTR could be another factor to affect RNA expression level . A shorter 3′ UTR in KSHV transcripts would provide expression advantage of viral genes in escaping from miRNA-mediated RNA degradation [53] , [54] . Recent studies unveiled highly prevalent alternative RNA polyadenylation in various organisms and its profound role in regulation of gene expression [23] . We identified several KSHV genes , including both non-coding and protein-coding genes , exhibit alternative RNA polyadenylation ( Table S4 ) . These alternative pA sites were previously ignored because of their relatively lower prevalence and the conceptual bias toward the longest detectable transcripts . All alternative pA sites identified in our study were located in the 3′ UTR of the respective transcripts and thus , their utilization does not affect coding potential of these variant transcripts . Notably , alternative polyadenylation was identified in two most abundant viral lncRNAs PAN and T1 . 5 , each of which harbors three alternative pA sites . We experimentally verified the two alternative pA sites for the expression of corresponding T1 . 5 transcripts in B cells with viral lytic infection ( Figures 6 , 8 ) . In addition , alternative polyadenylation of PAN RNA expression had been reported in our earlier study [35] . Therefore , the role of alternative polyadenylation in PAN and T1 . 5 expression will become an attractive subject for better understanding the function of PAN and T1 . 5 lncRNAs . Two unusual clusters of pA sites located downstream of the internal repeat regions were identified by PA-seq , but could not be validated by 3′ RACE in this study . The first cluster is located in the minus strand of the KSHV genome , downstream of “vnct” 13-bp repeats and composed of 5 individual pA sites within a ∼250-bp region from nt 29376 ( - ) to 29615 ( - ) . The second cluster of three pA sites from 118012 ( - ) to 118087 ( - ) is also located in the minus strand downstream of “zppa” repeat region containing two 23-bp repeats ( Figure 1A ) . These pA sites are located within the coding region of K12 , but no transcripts associated with these mapped pA sites were detected in previous studies [40] . None of them has a canonical PAS upstream . The sequence reads detected by PA-seq are more likely associated with cryptic transcription from the internal regions [55] . However , these transcripts are unstable and their degradation by cellular exosome is initiated by addition of a short pA tail , which is mediated by a non-canonical pA polymerase and is therefore is not dependent on PAS [56] . These transcripts with the rapid turnover may not be detectable by 3′RACE , but could be picked up by our high sensitive PA-seq .
Primary effusion lymphoma cells lines ( JSC-1 [KSHV+ , EBV+] , BCBL-1 [KSHV+ only] and TREx BCBL-1-vector and –RTA [BCBL-1 derived] ) [57] , [58] were used in this study The viral lytic replication was induced for 48 h by 3 mM sodium butyrate ( Bu ) for JSC-1 cells , 0 . 6 mM sodium valproate ( VA ) for BCBL-1 cells , or 1 µg/ml doxycycline ( DOX ) for both TREx BCBL-1-vector and –RTA cells . Total RNA was isolated by TRIzol ( Invitrogen ) and genomic DNA contamination was removed by RNeasy Mini kit ( Qiagen ) using on-column DNase I digestion step . The 3′end library for each sample was constructed using a modified PA-seq strategy [25] , [24] . Briefly , 10 µg of DNA-free total RNA from each sample described above was sheared into 200–300 nt fragments by heating ( 94°C for 3 minutes ) with magnesium . After precipitation a reverse transcription was carried out using a modified oligo ( dT ) primer ( 5′-bio-T16dUTTTVN-3′ , ‘bio’ denotes duo biotin group , ‘dU’ stands for deoxyuridine , ‘V’ represents any nucleotide except T and ‘N’ denotes any nucleotide ) . After second strand synthesis , resulted dsDNA was pulled down by Dynabeads MyOne C1 ( Invitrogen ) and dephosphorylated with APex Heat-Labile Alkaline Phosphatase ( Epicentre ) enabling PCR strand specificity for selective adaptor ligation . Dephosphorylated dsDNA was released from beads by USER enzyme digestion ( NEB ) and end-repaired , followed by an “A” base addition at the ends . Notably , only the first-stand cDNA contains a 5′ phosphate , and thus can be ligated to bar-coded Illumina paired-end Y linker without a nick . The usage of a dUTP in the oligo ( dT ) primer and the de-phosphorylation step reinforce strand-specificity , and allow precisely mapping of pA cleavage site at singe-base resolution . Ligation products between 250 bp and 450 bp were gel purified and PA-seq libraries were generated by 16-cycle PCR with Phusion Hot Start High-Fidelity DNA Polymerase ( Finnzymes ) . The obtained libraries were subjected to two technical replicate sequencing by an Illumina HiSeq2000 sequencer . Obtained raw reads were first aligned to KSHV genome ( GenBank acc no U75698 . 1 ) , EBV B95-8 strain genome ( GenBank acc no V01555 . 2 ) and human genome ( UCSC version hg19 ) by Burrows-Wheeler Alignment tool ( BWA ) [59] allowing two mismatches and processed by SAMtools [60] . All uniquely mapped KSHV-specific sequence pairs were used for downstream analyses . First the distribution of obtained reads along KSHV genome was visualized using IGV genome browser ( www . broadinstitute . org/igv/ ) to assure their suitability for pA site analysis . Individual KSHV pA sites were then designated by peak calling using F-Seq program [26] on combined libraries . The PA-seq peaks above the threshold of 50 reads were considered as true peaks . The peaks were further refined by removing pseudo pA sites resulting from “internal priming” due to continuous “A-stretch” in the template . After the peak calling the sequence reads were assigned back to individual samples to obtain the reads-counts for both latent and lytic infection . To obtain a relative expression level the total reads-counts were normalized per million to overall reads mapped to both KSHV and human [61] . The sequence surrounding the mapped pA sites was covered from 50 nts upstream and 50 nts downstream of each identified pA site for the motif analysis . The percentage of occurrence for each nucleotide was calculated , plotted and smoothed with the loess function in R software ( R version 2 . 12 . 1 ) . Polyadenylation signals ( PAS ) occurred within 50 nts upstream of pA site were assigned manually . Graphical representation of sequence conservation was generated by Weblogo v3 ( http://weblogo . berkeley . edu/ ) [62] , [63] . Transcript 3′ end was identified by SMARTer RACE cDNA Amplification Kit ( Clontech ) . The primer sequences used in 3′RACE are listed in Table S11 . The obtained 3′RACE products were sequenced directly or after cloning in pCR2 . 1-TOPO vector ( Invitrogen ) . Total RNA was isolated using TRIzol reagent . The cytoplasmic and nuclear fractions of RNA were isolated as described [64] . Obtained RNA ( 5 µg ) was separated on agarose gel and analyzed by Northern blot analysis with 32P labeled oligo probes: oVM 208 ( 5′-CGTGGCTGTGCTTCTCATCAT-3′ ) for T1 . 5 lncRNA , oJM7 ( 5′-GTTACACAACGCTTTCACCTACA-3′ ) for PAN lncRNA , oZMZ270 ( 5′-TGAGTCCTTCCACGATACCAAA-3′ ) for GAPDH and oST197 ( 5′-AAAATATGGAACGCTTCACGA-3′ ) for U6 snRNA . The single stranded sense and antisense RNA probes were prepared by FISH Tag RNA Multicolour Kit ( Invitrogen ) by in vitro transcription using DNA fragment of KSHV genome ( nt 24906–25375 for T1 . 5 and nt 29018–29481 for PAN lncRNAs ) as templates . The hybridization was performed as previously described [65] . After immobilization the cells were fixed with 2% paraformaldehyde , permeabilized with 0 . 5% Triton X-100 and blocked with hybridization buffer ( 50% formamid , 5×SSC , 0 . 1% Tween-20 , 50 µg/ml heparin , 100 µg/ml salmon DNA ) . The hybridization was carried out overnight at 55°C . The nuclei were counterstained with Hoechst dye . The pictures were collected using a Zeiss LSM510 META laser-scanning microscope ( Zeiss ) . Total cell RNA isolated by TRIzol ( Invitrogen ) was treated with Turbo DNA-free DNase to remove DNA . Five micrograms of total cell RNA was used to synthesize cDNA using SuperScript First-Stand Synthesis System ( Invitrogen ) . The GAPDH and human IL6 ( hIL6 ) transcript levels were determined by RT-qPCR using ΔCt method [44] , [66] , [67] . | A genome-wide polyadenylation landscape in the expression of human herpesviruses has not been reported . In this study , we provide the first genome landscape of viral RNA polyadenylation sites in B cells from KSHV latent to lytic infection by using a modified PA-seq protocol and selectively validated by 3′ RACE . We found that KSHV genome contains 67 active pA sites for the expression of its ∼90 genes and a few antisense transcripts . Among the mapped pA sites , a large fraction of them are for the expression of cluster genes and the production of bicistronic or polycistronic transcripts from KSHV genome and only one-third are used for the expression of single genes . We found that the size of individual PA peaks is positively correlated with the usage of corresponding pA site , which is determined by the number of reads within the PA peak from latent to lytic KSHV infection , and the strength of cis-elements surrounding KSHV pA site determines the expression level of viral genes . Lastly , we identified and experimentally validated alternative polyadenylation of KSHV ORF54 , T1 . 5 , and K11 during viral lytic infection . To our knowledge , this is the first report on alternative polyadenylation events in KSHV infection . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2013 | A Viral Genome Landscape of RNA Polyadenylation from KSHV Latent to Lytic Infection |
Asymptomatic persons infected with the parasites causing visceral leishmaniasis ( VL ) usually outnumber clinically apparent cases by a ratio of 4–10 to 1 . We assessed the risk of progression from infection to disease as a function of DAT and rK39 serological titers . We used available data on four cohorts from villages in India and Nepal that are highly endemic for Leishmania donovani . In each cohort two serosurveys had been conducted . Based on results of initial surveys , subjects were classified as seronegative , moderately seropositive or strongly seropositive using both DAT and rK39 . Based on the combination of first and second survey results we identified seroconvertors for both markers . Seroconvertors were subdivided in high and low titer convertors . Subjects were followed up for at least one year following the second survey . Incident VL cases were recorded and verified . We assessed a total of 32 , 529 enrolled subjects , for a total follow-up time of 72 , 169 person years . Altogether 235 incident VL cases were documented . The probability of progression to disease was strongly associated with initial serostatus and with seroconversion; this was particularly the case for those with high titers and most prominently among seroconvertors . For high titer DAT convertors the hazard ratio reached as high as 97 . 4 when compared to non-convertors . The strengths of the associations varied between cohorts and between markers but similar trends were observed between the four cohorts and the two markers . There is a strongly increased risk of progressing to disease among DAT and/or rK39 seropositives with high titers . The options for prophylactic treatment for this group merit further investigation , as it could be of clinical benefit if it prevents progression to disease . Prophylactic treatment might also have a public health benefit if it can be corroborated that these asymptomatically infected individuals are infectious for sand flies .
Visceral leishmaniasis ( VL ) is a vector borne infectious disease that is endemic in parts of Asia , Africa , South America and Europe , and is caused by different species of the genus Leishmania . In South America and the Mediterranean region , the disease is mainly caused by Leishmania infantum , which is a parasite of dogs opportunistically transmitted to humans . In East Africa and on the Indian subcontinent the disease is caused by L . donovani . L . donovani is essentially a parasite of humans , though recently there have been some reports of a possible animal reservoir [1] . Insect vectors are different species of phlebotomine sand flies . The symptomatic disease VL is a severe systemic disease characterized by fever , hepatosplenomegaly , pancytopenia and cachexia . VL has a protracted course and ultimately leads to death if it is not treated [2] . However , as in tuberculosis , not every infected individual develops disease , and there is a spectrum ranging from asymptomatic infection and subclinical disease to the full-blown VL syndrome [3] . Subclinical or mild forms of visceral disease have been described due to L . infantum in a Brazilian study , and due to L . tropica in veterans of Operation Desert Storm [4] , [5] . Epidemiological surveys showed that asymptomatic and subclinical forms are frequent and outnumber clinical VL cases [6] , . The prevalence of asymptomatic infections of L . infantum varied from 0 . 6 to 71% in endemic areas [6] . Prospective studies cite ratios of incident asymptomatic infections with L . donovani or L . infantum to symptomatic clinical cases as 1∶2 . 4 in Sudan , 4∶1 in Kenya , 5 . 6∶1 in Ethiopia , 6∶1 to 18∶1 in Brazil , 4∶1 in Bangladesh and 8 . 9∶1 in India and Nepal [8] , [9] , [10] , [11] , [12] . These large numbers of latent leishmania infections raise the questions on whether there is any clinical relevance of asymptomatic infection for an individual , and whether these latent infections play any role in disease transmission . If asymptomatically infected persons are indeed infectious for sand flies , theoretically they could contribute to an epidemic [13] . Studies of the risk of progression from infection to disease have yielded contradictory results . In a study by Das et al . 21 out of 91 polymerase chain reaction ( PCR ) and/or recombinant K39 ( rK39 ) positive but asymptomatic individuals ( 23 . 1% ) progressed to clinical VL over a follow-up period of one year [14] . Singh et al . reported 38 VL cases among 55 asymptomatic seropositive subjects followed up for one year ( 69% ) [15] . Gidwani et al . followed up a cohort of 870 endemic subjects of whom 230 were Direct Agglutination Test ( DAT ) positive at onset and 120 were rK39 positive . Over a 2-year follow-up , 15 VL cases occurred among 583 persons negative on both tests ( 2 . 6% ) versus 10 cases among 287 DAT and/or rK39 positive individuals ( 3 . 5% ) ; the difference was not statistically significant [16] . Ostyn et al . followed up a cohort of 9034 DAT negative subjects in high VL incidence villages in India and Nepal . The ratio of incident infection to incident disease was 8 . 9 to 1 , the risk of disease was 11 . 5 times higher ( 95% confidence interval 4 . 5–28 . 3 ) in those that had converted from seronegative to seropositive during the first year when compared to the risk of those that remained seronegative at the start of follow-up [7] . The absolute risk was not very high though , with 7 out of 375 seroconvertors ( 1 . 9% ) progressing to disease . Two studies from Brazil found no association between positive results in serological tests among asymptomatic persons and progression to VL disease [17][18] . Whereas the antibody-positive cases reported by Gidwani et al . included a mixture of individuals with remote and recent seroconversion , Ostyn et al . were able to establish an association between positive serology and disease in individuals who were recently infected ( recent seroconversion ) . Similarly , the asymptomatic subjects reported upon by Singh et al . were contacts of patients with VL , and as such they may have been recently infected and at higher risk of progression to disease . There has been recent controversy about the role of various antibody tests as markers of infection , and about the link between serologic status and disease risk in humans [19][20] . To explore the association between serologic status and disease risk , we analyzed prospective cohort data from two population-based research projects in endemic areas of India and Nepal . The main objective was to estimate the risk of progression to disease in asymptomatically infected persons and particularly among recent seroconvertors .
We pooled prospectively obtained data from two population-based research projects focused on the epidemiology of VL in India and Nepal . The data were divided into four distinct cohorts . The first two cohorts belong to the KALANET research project which was concluded in India and Nepal in 2009; the two other cohorts belong to the NIH/TMRC research project that started in 2008 in India [21][22] . The main aim of the KALANET project was to investigate the effectiveness of insecticide-treated bed nets ( ITN ) as an intervention against VL . A paired-cluster randomized trial was conducted in 26 highly VL endemic villages in the Muzaffarpur district of Bihar , India ( n = 16 clusters ) and the Terai region of Nepal ( n = 10 clusters ) . In the 13 intervention villages , ITN were distributed at the start of the trial , and the incidence rate of leishmania infection and VL in those villages was compared with that in 13 non-intervention villages over a two-year period . Incident infection , measured by seroconversion according to the DAT , was the primary outcome measure . Importantly , no differences in primary outcome were found between intervention and non-intervention villages , ( seroconversion 5 . 4% vs . 5 . 5% ) , nor in the secondary outcome , incident VL ( OR 0 . 99 ) [21] . During the current study we used available data from the 13 , 286 subjects who had been enrolled in the baseline survey: 7 , 950 from India ( cohort 1 ) and 5 , 336 from Nepal ( cohort 2 ) . In the framework of the NIH/TMRC research project , a population of approximately 85 , 000 persons living in 50 rural villages of Muzaffarpur district , a highly VL endemic district in Bihar India , have been under active surveillance since April 2008 . From this population , hereafter called ‘TMRC - old area’ , data are available for 13 , 134 subjects enrolled in the baseline serosurvey conducted in the villages with the highest VL incidence in the 3-year period preceding the survey ( cohort 3 ) . In 2010 the study population was expanded to an additional 10 villages named ‘TMRC-new area’ . The additional villages were selected based on high reported VL incidence during the year preceding their entry . All subjects providing informed consent were included in a serosurvey . Data for 7 , 934 individuals from this ‘TMRC-new area’ ( Cohort 4 ) were included in the present study . Altogether we thus have serology data available on 34 , 354 individuals enrolled in one of the four cohorts with a total of 76 , 113 person years follow-up . HIV prevalence in the area is assumed to be very low . In the TMRC study populations HIV testing was conducted among a subgroup of 1 , 787 persons of whom only 4 ( 0 . 2% ) tested positive . Study procedures are extensively described in Picado et al ( 2010 ) and Hasker et al ( 2012 ) [21][22] . Briefly , the KALANET project conducted three sero surveys separated by 12 month intervals starting in November 2006 . Two sero surveys were conducted in each of the TMRC areas , separated by a 12 month interval in the ‘old area’ , and a 6-month interval in the ‘new area’ . The TMRC baseline surveys took place between December 2009 and February 2010 in the ‘old area’ , or in April–May 2011 in the ‘new area’ . During all surveys , each subject aged over 2 years and residing in the selected villages was asked for a capillary blood sample collected by finger prick and stored on filter paper . These samples were tested with DAT and rK39 ELISA . Since enrolment , all subjects have been under regular follow-up by a network of village volunteers and have been visited on an annual basis as part of household surveys . Follow-up ended in May 2009 for the KALANET cohorts and is still ongoing for the TMRC cohorts . DAT and rK39 ELISA assays of serologic status were done from eluted filter paper blood as described elsewhere in the same laboratory at BHU India for the three Indian cohorts ( Cohorts 1 , 3 and 4 ) and at BPKIHS Nepal for the Nepal samples ( Cohort 2 ) [23] . For the purpose of this analysis , DAT seropositivity was defined as a DAT end titer of 1∶1 , 600 or above . Seropositives were again subdivided into moderately seropositive ( DAT titer ≥1∶1 , 600 & <1∶25 , 600 ) and highly sero-positive ( DAT titer ≥1∶25 , 600 ) . A seroconvertor was defined as an individual who was DAT negative in the baseline survey but DAT positive in the follow-up survey with an increase of at least 2 titer steps . DAT seroconvertors were also subdivided into moderate and high using the cut offs described above . For subjects from the KALANET study ( cohorts 1 and 2 ) we considered only the follow-up results of the second round of serosurvey and not those of the third round to define DAT seroconversion . For the KALANET cohort ( Cohorts 1 and 2 ) results for rK39 were available from the baseline survey only . The cut-off for a seropositive test for rK39 ELISA was determined for the 4 cohorts separately , using the percentage point optical density ( OD , set at 100% for a positive control ) with the highest Youden index as a cut off with recent VL cases ( diagnosed in the 24 months period preceding the survey ) as reference . As a second cut off point , to distinguish moderately seropositive from highly seropositive , we used the mean OD plus two standard deviations among those who had never suffered VL at the time of the baseline surveys . A seroconvertor was defined as an individual who was rK39 negative in the baseline survey but rK39 positive in the follow-up survey with an increase in optical density of at least three percentage points . RK39 seroconvertors were also subdivided into moderate and high , using the cut offs described above for seropositivity . A case of VL was defined as a person with a clinical history typical for VL ( fever of more than 2 weeks' duration , not responding to anti malaria treatment ) and in whom the presence of L . donovani has been confirmed; or a person with the combination of a clinical history typical for VL and a positive result of the rK39 test ( Inbios International , Seattle , WA , USA ) and a good response to specific VL treatment . All cases reported were ascertained by study physicians . Data were analyzed in Stata/IC V10 . 1 ( Stata Corp . , College Station Tx , USA ) . For analytical purposes we considered the four cohorts pooled and separately . Kaplan-Meier curves of DAT or rK39 seropositive or seroconversion titer levels were made for the combined study population . Cox regression models were fitted to each cohort separately and the combined cohorts to test for ( 1 ) associations between initial DAT or rK39 sero-positivity and clinical VL , and for ( 2 ) associations between DAT or rK39 seroconversion and clinical VL . The latter was done only for the TMRC cohorts ( cohort 3 and 4 ) since for the KALANET cohorts ( cohort 1 and 2 ) rK39 results were only available at baseline . In the models for combined cohorts , the factor ‘cohort’ was included as a fixed effect . We tested for interactions , and if there were no interactions we presented a combined effect for the different study cohorts . During follow-up visits , information on all household members was obtained irrespective of whether they were present or absent . Deaths were censored on the reported date of decease . Both research projects ( KALANET and TMRC ) sought consent at three levels: community , household and individual . Communities were duly informed about the purpose of the studies and consent was sought from village leaders . For the KALANET trial ethical clearance was obtained from the ethical committees of BHU ( India ) , the BPKIHS ( Nepal ) , the London School of Hygiene and Tropical Medicine ( UK ) , and the University of Antwerp ( Belgium ) . For the TMRC study ethical clearance was obtained from the review committee of the U . S . National Institutes of Health ( NIH ) , as well as Institutional Review Boards of the Institute of Medical Sciences , Banaras Hindu University , Varanasi , India , and the University of Iowa . All subjects provided written informed consent; in case of illiterate subjects , a thumb print plus a signature of an independent witness were used . For minors under the age of 18 , written informed consent was obtained from a parent or guardian . Informed consent procedures for each of the studies were approved by the respective review boards .
Among the total eligible study population of 34 , 354 subjects , 1 , 825 were excluded from further analysis because of a reported history of VL prior to the first serosurvey . Thus 32 , 529 subjects were included in the analysis , of whom 28 , 073 ( 86 . 3% ) were also enrolled in a follow-up survey . Participants were divided over study populations and countries as indicated in table 1 . Based on the criteria described in the ‘study procedures’ section , cut off values for rK39 ELISA expressed in percentage point optical density of a positive control ( pp ) varied between cohorts . The cut offs that distinguished seronegative from moderately seropositive individuals in the KALANET cohorts were 27 pp for Nepal and 34 pp for India . The cutoffs for strongly seropositive individuals were 44 and 42 pp respectively . The cut offs for individuals in the TMRC cohorts were lower , i . e . , 15 and 21 pp for the ‘old area’ , and 11 and 27 pp for the ‘new area’ . Overall 235 VL cases occurred among the72 , 169 total person years observed , i . e . 3 . 3 per 1 , 000 person year at risk . As shown in the Kaplan-Meier graphs in Figure 1 , in all cohorts combined the probability of developing VL was strongly increased in individuals with the highest titers of either DAT or rK39 . Because the interactions between titer levels and study cohorts were statistically significant ( p = 0 . 0001 both for DAT and for rK39 ) , hazard ratios are presented for each of the populations separately ( Table 2 ) . There was little difference in risk of progressing to clinical VL between individuals who were seronegative for DAT versus individuals in the moderately seropositive group . Only in the cohort 4 ( new TMRC area ) there was a statistically significant hazard ratio of progression among individuals in the moderately seropositive group ( hazard ratio 3 . 8 , 95% CI 1 . 9–7 . 3 ) . Among the other cohorts there were either no cases in the moderate titers group or the hazard ratio was very close to 1 when comparing with the seronegative individuals . Among the strongly seropositive individuals however , all hazard ratios were very high and statistically significant , ranging from 7 . 9 ( 95% CI 2 . 3–26 . 6 ) in cohort 3 ( old TMRC ) to 35 . 6 ( 95% CI 17 . 0–74 . 7 ) in cohort 2 ( KALANET-Nepal ) ( Table 2 ) . The pattern of progression to VL according to the baseline rK39 status was similar to that according to DAT ( Figure 1B and Table 3 ) . Individuals in cohorts 3 and 4 ( the two TMRC areas ) who were moderately seropositive had a significantly increased risk of progressing to symptomatic disease , although the association was much more pronounced among the strongly seropositive persons . Hazard ratios were very high and highly significant among strongly seropositive individuals in all four cohorts , ranging from 7 . 7 ( 95% CI 2 . 3–26 . 0 ) in cohort 3 ( old TMRC ) to 39 . 6 ( 95% CI 25 . 2–62 . 3 ) in cohort 4 ( new TMRC ) ( Table 3 ) . The associations between seroconversion and progression to symptomatic VL were very strong for those with high titers , measured with either DAT or rK39 ( Figure 2 ) . The risk of progression to VL among seroconvertors with intermediate titers was not very different from the risk in those who had remained seronegative . As illustrated by the Kaplan-Meier curves in Figure 2 , most cases occurred during the first 6 months of follow-up after the second serosurvey . The overall hazard ratio of developing symptomatic VL when comparing non-convertors to convertors with moderate DAT titers was 6 . 9 ( 95% CI 3 . 4–14 . 3 ) . Among seroconvertors with high titers the hazard ratio was 97 . 4 ( 95% CI 53 . 5–177 . 4 ) , with 12% ( 15 out of 123 ) of subjects progressing to disease ( table 4 ) . There was no significant interaction between cohort and DAT titer level ( p = 0 . 58 ) . Unlike the data for DAT titers , there was no significant association between moderate titer rK39 conversion and progression to symptomatic VL in the two TMRC cohorts . However there was a very strong and highly significant association with progression to symptomatic VL among those who converted with high rK39 titers . Hazard ratios ranged from 15 . 9 ( 95% CI 2 . 1–121 . 4 ) to 123 . 9 ( 95% CI 52 . 7–291 . 2 ) among the cohorts ( table 5 ) . There was significant effect modification by cohort ( p = 0 . 03 ) , precluding us from calculating an overall effect .
At population level , specific antibodies against L . donovani as measured in DAT or rK39 are strongly associated with risk to disease progression . This relationship is even more pronounced among individuals who have recently seroconverted . | To study the association between positive serology for VL among asymptomatic persons and progression to disease , we analyzed combined data from two different cohort studies in which over 32 , 000 subjects were enrolled . All subjects were screened for presence of antibodies to L . donovani with DAT and rK39 . After an interval of 6–12 months , a second round of surveys was conducted in which seroconvertors were identified , i . e . persons seronegative at baseline that had become seropositive in the interval between surveys . We subdivided the subjects according to serologic titers scored as negative , moderately positive or strongly positive . During follow-up we observed that the risk of progressing to disease was not very much different between seronegative and moderately seropositive subjects , but strongly increased for high titer seropositive subjects . This pattern was true for the initial serology , and even more pronounced among seroconvertors . Given the strongly increased risk of progression to disease among those with high DAT or rK39 titers , these data raise the question whether prophylactic treatment would be beneficial . The case for prophylactic treatment would be further strengthened if it could be shown that asymptomatically infected individuals are infectious to sand flies . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"infectious",
"diseases",
"epidemiology"
] | 2014 | Strong Association between Serological Status and Probability of Progression to Clinical Visceral Leishmaniasis in Prospective Cohort Studies in India and Nepal |
It has been reported that feeding mice resveratrol activates AMPK and SIRT1 in skeletal muscle leading to deacetylation and activation of PGC-1α , increased mitochondrial biogenesis , and improved running endurance . This study was done to further evaluate the effects of resveratrol , SIRT1 , and PGC-1α deacetylation on mitochondrial biogenesis in muscle . Feeding rats or mice a diet containing 4 g resveratrol/kg diet had no effect on mitochondrial protein levels in muscle . High concentrations of resveratrol lowered ATP concentration and activated AMPK in C2C12 myotubes , resulting in an increase in mitochondrial proteins . Knockdown of SIRT1 , or suppression of SIRT1 activity with a dominant-negative ( DN ) SIRT1 construct , increased PGC-1α acetylation , PGC-1α coactivator activity , and mitochondrial proteins in C2C12 cells . Expression of a DN SIRT1 in rat triceps muscle also induced an increase in mitochondrial proteins . Overexpression of SIRT1 decreased PGC-1α acetylation , PGC-1α coactivator activity , and mitochondrial proteins in C2C12 myotubes . Overexpression of SIRT1 also resulted in a decrease in mitochondrial proteins in rat triceps muscle . We conclude that , contrary to some previous reports , the mechanism by which SIRT1 regulates mitochondrial biogenesis is by inhibiting PGC-1α coactivator activity , resulting in a decrease in mitochondria . We also conclude that feeding rodents resveratrol has no effect on mitochondrial biogenesis in muscle .
Resveratrol has been reported to have a number of remarkable effects in mice . These include protection against high-fat-diet-induced obesity and insulin resistance [1]–[3] , marked improvements in running endurance and maximal oxygen uptake capacity ( VO2max ) [3] , increased muscle strength [3] , improved motor coordination [1]–[3] , and antiaging effects [1] , [2] . Subsequent studies have shown that resveratrol does not have antiaging effects in mice , as evidenced by no increases in average or maximum longevity [4] , [5] . The protection against obesity and insulin resistance was attributed to an increase in , and uncoupling of , mitochondria in brown fat , and the increase in running endurance and VO2max were attributed to an increase in muscle mitochondria [3] . The increase in mitochondria induced by resveratrol was explained by activation of the protein deacetylase SIRT1 , resulting in deacetylation and activation of the transcription coactivator PGC-1α [3] . PGC-1α regulates mitochondrial biogenesis [6] . The pharmaceutical agent SRT1720 has also been reported to activate SIRT1 , resulting in PGC-1α activation , and an increase in enzymes of the mitochondrial fatty acid oxidation pathway in muscle and improved running performance , muscle strength , and coordination [7] . However , Pacholec et al . [8] have reported that SRT1720 does not activate SIRT1 , and that it does not induce an increase in mitochondrial enzymes in mice . Based on studies on yeast and in vitro , it was initially thought that resveratrol directly activates SIRT1 [9] . However , Kaeberlein et al . [10] showed that , although resveratrol binds and deacetylates peptide substrates that contain a Fluor de Lys , it does not bind or deacetylate acetylated peptides lacking the flurophore . They also found that resveratrol has no effect on SIRT2 activity in yeast . Similarly , Bora et al . [11] found that resveratrol activation of SIRT1 was completely dependent on the presence of a covalently attached flurophore . Evidence that resveratrol can activate AMP activated protein kinase ( AMPK ) [12]–[14] led to further studies that indicated that the activation of SIRT1 by resveratrol is indirect , and is mediated by activation of AMPK [15] . The mechanism by which AMPK is thought to activate SIRT1 is by increasing NAD concentration [15] . We have a long-standing interest in the adaptive responses to endurance exercise , such as running and swimming , which include an increase in muscle mitochondria [16] , [17] . Endurance exercise training also results in increases in endurance and in maximal oxygen uptake capacity . Endurance exercise does not , by itself , result in increases in either muscle strength , which occurs in response to heavy resistance exercise , or improved motor coordination , which occurs in response to activities that require various motor skills . A sedentary lifestyle greatly increases the risk of developing obesity , insulin resistance , type 2 diabetes , atherosclerosis , and frailty [18] . Therefore , in addition to being necessary for successful competition in sports , regular exercise is necessary for maintenance of health and functional capacity . Because it is difficult to motivate people to exercise , an effective , nontoxic exercise mimetic—that is , an “exercise pill”—could have great public health value . Therefore , the reports that , in addition to protecting against obesity and insulin resistance , resveratrol feeding mimics not only the adaptive response to endurance exercise but also the adaptations to strength training and motor skill exercise training were of great interest to us . The present study was undertaken to further evaluate the adaptive response of skeletal muscle mitochondria to resveratrol treatment .
Feeding rats resveratrol in a chow diet containing 4 g resveratrol per kg diet [3] for 8 wk had no effect on the expression of PGC-1α or on a number of mitochondrial proteins in rat skeletal muscle as shown in soleus muscle ( Figure 1A ) . A similar lack of effect was found in the gastrocenemius muscle . Feeding rats a high fat diet containing 4 g resveratrol per kg diet also had no effect on the expression of a range of mitochondrial enzyme proteins ( Figure 1B ) . To rule out the possibility that the lack of effect of resveratrol on the mitochondrial content of skeletal muscle in rats was due to a species difference , we fed mice a high fat diet containing 4 g resveratrol per kg/diet as in the study by Lagouge et al . [3] . As in the rats , resveratrol feeding had no effect on the expression of PGC-1α or a number of mitochondrial proteins in skeletal muscle of mice ( Figure 1C ) . To evaluate the possibility that the lack of effect of resveratrol on mitochondrial biogenesis is due to an inadequate increase in plasma resveratrol , we measured plasma resveratrol concentration . Plasma resveratrol concentration at 9:00 am in rats in the fed state averaged 1 . 56±0 . 28 µM . This plasma resveratrol concentration is higher than that reported by Lagouge et al . [3] in their resveratrol fed mice , in which the highest concentration attained was ∼0 . 5 µM . Most of the information regarding the effects of resveratrol on , and the role of SIRT1 in , the regulation of mitochondrial biogenesis has come from studies on C2C12 myotubes or other cells in culture . Because resveratrol feeding had no effect on mitochondrial biogenesis in laboratory rodents , we evaluated the effect of resveratrol on mitochondrial biogenesis in C2C12 myotubes . The concentration of resveratrol that was routinely used in studies on C2C12 myotubes by Auwerx's group was 50 µM [3] , [15] , ∼100-fold higher than the highest plasma resveratrol level in their resveratrol fed mice [3] . In our initial experiments we found that 50 µM resveratrol is toxic , with a high proportion of the C2C12 myotubes appearing to be dead or dying after 24 h of exposure to 50 µM resveratrol . That this concentration of resveratrol is cytotoxic was born out by measurements of cytotoxicity ( Figure 2A ) and of ATP concentration , which was markedly reduced ( Figure 2B ) . Similarly , Zang et al . [13] have reported that exposure of Hep-G2 cells to 50 µM resveratrol for 60 min resulted in an 80% reduction in ATP concentration . The decrease in ATP concentration in cells exposed to a high concentration of resveratrol is mediated by toxic effects on mitochondria , with inhibition of ATP synthase [19] and NADH: ubiquinone oxidoreductase [20] . Numerous studies have shown that concentrations of resveratrol in the 30 to 100 µM range kill a variety of malignant cells [21] . These studies were uncontrolled , and it was assumed that resveratrol specifically kills cancer cells . However , the present finding and that of Zang et al . [13] show that resveratrol at the high concentrations used is also lethal for nonmalignant cells . In the study in which 50 µM resveratrol increased mitochondrial biogenesis in C2C12 myotubes [3] , the investigators used cells that overexpressed PGC-1α . We have observed that myotubes in which PGC-1α is overexpressed have increased resistance to the effect of puromycin ( DH Han and JO Holloszy , unpublished findings ) , suggesting the possibility that overexpression of PGC-1α results in a nonspecific increase in resistance to toxins . We , therefore , evaluated the effect of 50 µM resveratrol in C2C12 cells in which PGC-1α was overexpressed by infection with a virus expressing PGC-1α . As shown in Figure 2C , the toxic effect of 24 h exposure to 50 µM resveratrol on cell viability was markedly reduced . However , there was still a significant reduction in ATP concentration ( Figure 2D ) . Treatment with 50 µM resveratrol for 24 h resulted in an increase in mitochondrial biogenesis in the myotubes in which PGC-1α was overexpressed , as evidenced by increases in the expression of a number of mitochondrial proteins ( Figure 3A ) . All of our subsequent experiments in which 50 µM resveratrol was used were performed on C2C12 myotubes in which PGC-1α was overexpressed . To evaluate the effect of resveratrol in the absence of PGC-1α overexpression , we tried to identify a resveratrol concentration that induces an increase in mitochondrial proteins in wild type C2C12 cells . Resveratrol concentrations in the 1 µM to 10 µM range did not result in a decrease in ATP concentration ( Figure 2B ) . Although exposure to 20 µM resveratrol for 24 h is less toxic than exposure to 50 µM , it results in a decrease in cell viability ( Figure 2A ) and a ∼50% decrease in ATP concentration ( Figure 2B ) . Six hours of treatment with 20 µM resveratrol resulted in a smaller decrease in ATP ( ∼20% ) , and wild-type C2C12 cells treated with 20 µM resveratrol for 6 h followed by an 18 h recovery period showed no evidence of toxicity . “Training” the wild-type C2C12 cells by exposing them to 20 µM resveratrol for 6 h per day for 3 d resulted in increases in PGC-1α and a number of mitochondrial proteins ( Figure 3B ) , while the same treatment with 1 µM , 5 µM , or 10 µM resveratrol had no effect ( Figure S1 ) . As shown in Figure 4A , treatment with 20 µM resveratrol resulted in increased phosphorylation of AMPK and acetyl-CoA carboxylase ( ACC ) in C2C12 myotubes . As a first step in evaluating the relative roles of AMPK and SIRT1 in the resveratrol-induced increase in mitochondrial biogenesis , we infected C2C12 myotubes with an adenovirus encoding a dominant-negative AMPK gene construct . That the DN AMPK was effective in blocking AMPK activity is demonstrated by prevention of increases in AMPK and ACC phosphorylation in response to resveratrol treatment ( Figure 4B ) . Blocking AMPK activity prevented induction of an increase in mitochondrial proteins by resveratrol ( Figure 4C ) , showing that AMPK activation is necessary for stimulation of mitochondrial biogenesis by resveratrol . Jäger et al . [22] have shown that AMPK directly phosphorylates and activates PGC-1α . Canto et al . [15] have interpreted their data to indicate that phosphorylation of PGC-1α by AMPK constitutes a priming event for subsequent deacetylation by SIRT1 , and that deacetylation of PGC-1α is a key mechanism by which AMPK triggers PGC-1α activity . To further evaluate the relative roles of SIRT1 and AMPK in the resveratrol-induced increase in mitochondria , we used nicotinamide to inhibit SIRT1 [23] . That 10 mM nicotinamide decreases SIRT1 activity in C2C12 myotubes is evidenced by the finding of increases in the acetylation of p53 , which is a SIRT1 substrate [24] ( Figure 5A ) and of PGC-1α ( Figure 5C ) . Nicotinamide also prevented p53 deacetylation in response to 50 µM resveratrol ( Figure 5A ) . However , we were surprised to find that nicotinamide did not prevent the resveratrol-induced increase in mitochondrial proteins ( Figure 5B ) . Treatment of C2C12 myotubes with 10 mM nicotinamide had no effect on ATP concentration ( nicotinamide 5 . 3±0 . l3 µmol/g protein , Control 5 . 7±0 . 18; n = 6 per group ) . We further evaluated the role of SIRT1 in mitochondrial biogenesis by suppression of SIRT1 activity by adenovirus-mediated expression of a dominant-negative ( DN ) SIRT1 H355A [23] , and knockdown of SIRT1 with a shRNA , in C2C12 myotubes . SIRT1 H355A suppressed SIRT1 activity as evidenced by an increase in PGC-1 acetylation and inhibition of resveratrol-induced PGC-1α deacetylation ( Figure 5C ) . Both the DN SIRT1 and the SIRT1 shRNA resulted in increased PGC-1α coactivator activity , measured in C2C12 myotubes co-transfected with a PGC-1α–GAL4 fusion construct and a luciferase reporter [25] , and enhanced the resveratrol-induced increase in PGC-1α activity ( Figure 5D ) . Overexpression of wild-type SIRT1 resulted in PGC-1α deacetylation ( Figure 5C ) , reduced PGC-1α coactivator activity , and prevented the increase in PGC-1α activity induced by resveratrol ( Figure 5D ) . SIRT1 H355A expression in myotubes resulted in an increase in mitochondrial enzyme proteins ( Figure 6A ) . Expression of SIRT1 H355A in rat triceps muscle by electroporation also resulted in an increase in mitochondrial enzyme proteins ( Figure 6B ) . Furthermore , knockdown of SIRT1 by transfection of C2C12 myotubes with a SIRT1 shRNA brought about an increase in mitochondrial proteins , providing further evidence that acetylation activates PGC-1α ( Figure 6C ) . Expression of DN SIRT1 H355A in C2C12 myotubes had no effect on ATP concentrations ( Control 5 . 7±0 . 18 , DN SIRT1 H355A 6 . 0±0 . 3; n = 6 per group ) . To further evaluate the effect of SIRT1 on mitochondrial biogenesis , we determined the effect of overexpression of SIRT1 by adenovirus mediated infection of C2C12 cells , and electroporation of rat triceps muscle , with a SIRT1 gene construct . SIRT1 overexpression resulted in a decrease in cytochrome c and inhibited the resveratrol-induced increase in cytochrome c in C2C12 myotubes ( Figure 6D ) . Overexpression of SIRT1 in rat triceps muscle resulted in decreases in mitochondrial enzyme proteins ( Figure 6E ) . Interestingly , the increase in PGC-1α coactivator activity induced by acetylation does not result in an increase in PGC-1α expression ( Figure 6 ) . This is in contrast to PGC-1α activation by phosphorylation by AMPK and/or p38 MAPK , which is associated with an increase in PGC-1α expression ( Figure 3B ) [22] , [26]–[29] . A probable explanation for this difference is that AMPK and p38 MAPK do not just activate PGC-1 , but also activate the transcription factors that induce increased PGC-1α expression . P38 MAPK phosphorylates and activates ATF2 , which binds to a CREB binding site on the PGC-1α promoter , and AMPK and p38 MAPK bring about activation of MEF2 , which binds to a MEF2 binding site on the PGC-1α promoter , resulting in increased PGC-1α transcription [27] , [30]–[32] .
In the present study , resveratrol feeding had no effect on mitochondrial biogenesis in skeletal muscle even though our animals were fed a diet containing the same amount of resveratrol , 4 g/kg diet , as used by Lagouge et al . [3] , and more than the dose , 0 . 4 g/kg diet , used by Bauer et al . [1] . In studies on the effects of resveratrol on cells in culture , concentrations in the 30 µM to 100 µM range have routinely been used [3] , [12] , [14] , [15] . Based on our findings on C2C12 myotubes , the concentration of resveratrol required to induce an increase in mitochondrial biogenesis is above 10 µM , and the data shown by Bauer et al . [1] suggest the concentration of resveratrol needed to activate AMPK in CHO cells is also above 10 µM . The plasma resveratrol concentration in our rats was 1 . 56±0 . 28 µM and the highest concentration in the mice of Lagouge et al . [3] was ∼0 . 5 µM . Thus , a likely explanation for the failure of resveratrol feeding to induce mitochondrial biogenesis in rats and mice in our study is its poor bioavailability . In our experiments on C2C12 cells , we confirmed the finding of Lagouge et al . [3] that treatment of C2C12 cells with a high concentration of resveratrol results in both PGC-1α activation , evaluated using a PGC-1α–GAL4 construct together with a luciferase reporter , and an increase in mitochondrial biogenesis . The research groups of Auwerx and Puigserver have published a large number of studies , reporting that deacetylation activates and acetylation deactivates PGC-1α [3] , [7] , [15] , [33]–[39] . Phosphorylation of PGC-1α by AMPK results in PGC-1 activation and increased mitochondrial biogenesis [22] . We found that high concentrations of resveratrol activate AMPK in C2C12 cells by a toxic effect on mitochondria that reduces ATP level , and that this is the mechanism by which resveratrol activates PGC-1α . We also found that the concomitant increase in SIRT1 activity , also mediated by AMPK , results in a deacetylation of PGC-1α that causes a blunting of the increase in PGC-1α activity induced by AMPK . This is in contrast to the report by Canto et al . [15] that activation of PGC-1α by AMPK is dependent on PGC-1α deacetylation by SIRT1 . In support of this conclusion , they reported that inhibition of SIRT1 with nicotinamide or knock down of SIRT1 markedly reduced PGC-1α activation and attenuated the increase in mitochondrial proteins in response to AMPK activation . We confirmed that activation of AMPK results in SIRT1 activation , as evidenced by deacetylation of p53 and PGC-1α . We also confirmed that suppression of AMPK activity blocks the increase in mitochondrial proteins induced by resveratrol . However , we were surprised to find that inhibiting SIRT1 with nicotinamide did not prevent the resveratrol-induced increase in mitochondrial proteins in C2C12 myotubes . Furthermore , an increase in PGC-1α acetylation , mediated by suppression of SIRT1 activity using a dominant-negative SIRT1 construct , resulted in an increase in PGC-1α coactivator activity and mitochondrial biogenesis . Knockdown of SIRT1 also increased PGC-1α activity . Further evidence that PGC-1 is activated by acetylation is provided by the findings that overexpression of wild-type SIRT1 , resulting in PGC-1 deacetylation , decreases mitochondrial proteins , blunts the resveratrol/AMPK-induced increase in cytochrome c , and reduces PGC-1α coactivator activity . An additional mechanism by which the inhibitory effect of SIRT1 on PGC-1α activity might be mediated is by deacetylation and inactivation of the transacetylase p300 [40] . p300 is a transacetylase that binds to and acetylates PGC-1α [41] , and powerfully enhances its coactivator activity [42] . Thus , inactivation of p300 , resulting in decreased PGC-1α acetylation , could result in a reduction of PGC-1α activity . Our findings that SIRT1 activation decreases PGC-1α coactivator activity and that inhibition or knockdown of SIRT1 increases PGC-1α activity are in keeping with data reported by Finkel's group [41] . These investigators showed that SIRT1 binds to and deacetylates PGC-1α , and that increasing SIRT1 expression in PC12 cells results in a ∼25% reduction in O2 consumption , a ∼45% decrease in cytochrome oxidase ( COX ) subunit 2 expression , and a ∼50% decrease in activity of a GAL4–PGC-1α fusion construct [41] . They also found that overexpression of the transacetylase p300 , which activates PGC-1α [42] , dramatically increased PGC-1α acetylation [41] . Our findings also confirm the report by Gurd et al . [43] that overexpression of SIRT1 in rat skeletal muscle results in decreased expression of the mitochondrial enzyme COXIV . Gurd et al . [43] also found an inverse relationship between mitochondrial content and SIRT1 content in different types of skeletal muscle and heart muscle . SIRT1 is induced by , and appears to play a key role in the adaptive responses to , fasting , starvation , and calorie restriction [44]–[47] . The Auwerx and Puigserver research groups have interpreted their findings to indicate that SIRT1 leads to increased mitochondrial biogenesis , which provides a molecular mechanism that allows cells to survive and adapt to periods of nutrient deprivation [35] , that SIRT1 activation by SRT1720 mimics low energy levels [7] , and that “interdependent regulation of SIRT1 and AMPK provide a finely tuned amplifier mechanism for energy homeostasis under low energy availability” [15] . A key component of this concept is that mitochondrial adaptations induced by increased SIRT1 activity are necessary for the switch from carbohydrate to fat oxidation in response to fasting [7] , [35] . What was actually reported is that treatment with resveratrol [3] or SIRT1720 [7] and other interventions that activated SIRT1 [34] , [38] resulted in increases in basal oxygen consumption , heat production/body temperature , and protection against weight gain or reduced weight gain despite no decrease in food consumption . This syndrome , which resembles hyperthyroidism , was attributed by the authors to mitochondrial adaptations in brown adipose tissue and is incompatible with the large increase in running endurance reported in these mice [3] , [7] , [38] . Adaptive responses were selected for because they enhance an organism's chances of surviving environmental changes . Increases in energy expenditure and substrate oxidation resulting in more rapid depletion of energy stores , such as were reported to occur with SIRT1 activation , would be seriously maladaptive responses to fasting , starvation , or CR . Actually , it is well documented that fasting and CR result in decreases in metabolic rate , as reflected in lower resting oxygen consumption and a decrease in body temperature [48]–[51] . With regard to the claim that an increase in mitochondrial fatty acid oxidation enzymes is necessary for the switch from carbohydrate to fatty acid oxidation in muscle [7] , [35] , no increase in mitochondria is needed . Skeletal muscle has a low rate of energy utilization at rest and contains sufficient mitochondria to make possible a many-fold , acute increase in fatty acid oxidation in response to exercise that greatly exceeds the increase in fat oxidation in muscle in response to fasting . Furthermore , SIRT1-null mice are hypermetabolic , have elevated rates of fatty acid utilization , and readily switch from carbohydrate to fat oxidation in response to fasting [52] . In conclusion , our results show that resveratrol feeding does not induce an increase in muscle mitochondria in rodents . This lack of effect may be due to poor bioavailability , because the plasma levels of resveratrol attained in response to feeding large amounts of resveratrol are far below the concentration of resveratrol required to activate AMPK . This seems fortunate , because the activation of AMPK by resveratrol is mediated by a toxic effect that depletes ATP in cells exposed to AMPK-activating concentrations of resveratrol . Surprisingly , in light of the many studies reporting that deacetylation of PGC-1α results in activation of PGC-1α's coactivator activity , we find that deacetylation decreases , and PGC-1α acetylation increases , PGC-1α activity and mitochondrial biogenesis . Our results indicate that the activation of PGC-1α by resveratrol is mediated by AMPK , and that the activation of SIRT1 by AMPK acts to reduce , rather than induce , this activation .
This research was approved by the Animal Studies Committee of Washington University School of Medicine . Rats were lightly anesthetized during muscle electroporation . Rats were anesthetized with pentobarbital and , after muscles were harvested , were killed by exsanguination while under anesthesia . Antibodies against cytochrome oxidase subunit I ( COXI ) , cytochrome oxidase subunit IV ( COX IV ) , Core II , Complex III FeS , NADH ubiquinol oxidoreductase ( NADH-UO ) , and succinate ubiquinol oxidoreductase ( SUO ) ATP synthase alpha subunit #45924 and lipofectamine 2000 were purchased from Invitrogen ( Carlsbad , CA ) . Anti-cytochrome c antibody was obtained from BD Biosciences ( San Jose , CA ) . Antibodies against p53 , acetyl-p53 , AMP-activated protein kinase ( AMPK ) , phospho-AMPK , acetyl-CoA carboxylase ( ACC ) and phospho-ACC were products of Cell Signaling technology ( Beverly , MA ) . An anti-SIRT1 antibody #09844 was purchased from EMD Millipore . An anti-PGC-1α c-terminal ( 777–797 ) antibody #516557 was purchased from EMD Millipore ( Billerica , MA ) ; an antibody against acetylated lysine #9441 was purchased from Cell Signaling ( Beverly , MA ) . Horseradish peroxidase ( HRP ) –conjugated donkey anti-rabbit IgG and donkey anti-mouse IgG were purchased from Jackson ImmunoResearch Laboratories ( West Grove , PA ) . Enhanced chemiluminescence ( ECL ) reagents were obtained from Amersham ( Arlington Heights , IL ) . All other reagents were purchased from Sigma ( St . Louis , MO ) . This research was approved by the Animal Studies Committee of Washington University School of Medicine . Male Wistar rats weighing ∼95 g were purchased from Charles Rivers ( Wilmington , MD ) and housed in individuals cages . The resveratrol used in the study on rats was purchased from Stryka Botanicals ( Hillsborough , NJ ) . Control rats were fed Purina rodent chow , and the resveratrol-fed rats were given the chow diet containing 4 g resveratrol per kg diet , for 8 wk . Male c57BL/6J mice were purchased from Jackson Laboratory ( Bar Harbor , ME ) , housed 6 per cage , and fed a high fat diet , 50% of calories from fat , or the high fat diet containing 4 g resveratrol per kg diet 3 for 8 wk . The resveratrol used in the study on mice was a kind gift from DSM Nutritional Products ( Basel , Switzerland ) . ( The resveratrol used in the study by Lagouge et al . was from Orchid , Chennai , India . ) Rats or mice were anesthetized with sodium pentobarbital 5 mg/100 g body weight . Muscles were dissected out , clamp-frozen , and kept at −80°C until used for assays . ATP concentration was measured using a luminescence ATP detection assay ( ATPlite , Perkin Elmer , Waltham , MA ) ; LDH activity , as an indicator of cytotoxicity , was measured using an LDH-Cytotoxicity Assay Kit ( BioVision , Mountain View , CA ) , according to the manufacturer's instructions . For expression in skeletal muscle via electroporation , wild-type SIRT1 and dominant-negative SIRT-1 H355A constructs were purchased from Addgene ( Cambridge , MA ) and inserted into pCDNA3 . 1 ( Invitrogen , Carlsbad , CA ) . For expression in C2C12 myoblasts by transfection , a gal-4-DBD-PGC-1α plasmid was purchased from Addgene ( Cambridge , MA ) [53] , a 9×gal-4–dependent reporter plasmid was purchased from Promega ( Madison , WS ) , and a LacZ control plasmid was purchased from Invitrogen ( Carlsbad , CA ) . For expression by adenoviral infection in C2C12 myotubes , the adenoviral constructs of pAd-Track Flag-PGC-1α [36] , pAd-Track Flag-SIRT1 [23] , and pAd-Track Flag dominant-negative SIRT1 H355A [23] were purchased from Addgene ( Cambridge , MA ) . Dominant-negative Myc-AMPKα 2 DNA [54] was PCR cloned and ligated into pAd-Track plasmid . Mouse SIRT1 shRNA ( 5′-GCCCTGTAAAGCTTTCAGAA-3′ ) and scrambled control ( 5′-GATGAAGTCGACCTCCTCAT-3′ ) sequences were cloned into pRNAT-H1 . 1/adeno ( Genescript , Piscataway , NJ ) . Recombinant adenoviruses were generated employing a system described by He et al . [55] . Transfection of DNA into rat skeletal muscle was accomplished by using an electric pulse-mediated gene transfer technique [56] . Male Wistar rats weighing ∼60 g were anesthetized with isoflurane gas . A triceps muscle was injected with 100 µg of plasmid DNA containing either empty vector , pcDNA3 . 1 SIRT1 WT , or pcDNA3 . 1 Sirt1 H355A in 100 µl saline , using a 27 gauge needle , at a rate of 0 . 04 ml/min . After injection , an electric field was applied to the triceps muscle using a S88 square-pulse stimulator ( Grass ) with a 533 model two-needle array ( BTX ) . The electric field application consisted of 8 pulses of 100 ms duration , at a frequency of 1 Hz and amplitude of 100 volts , that were applied perpendicular to the muscles' long axis . Muscles were harvested 14 d after electroporation . C2C12 mouse myoblasts were grown in DMEM ( 4 . 5 g glucose/L , Invitrogen ) containing 10% fetal bovine serum , 100 µU/ml penicillin , and 100 µU/ml streptomycin . Differentiation was initiated by switching to medium containing 2% heat inactivated horse serum when the myoblasts were 90% confluent . After 48 h of differentiation , batches of myotubes were infected with adenoviruses expressing ( a ) Flag-PGC-1α , ( b ) dominant-negative Myc-AMPKα 2 , ( c ) dominant-negative Flag-SIRT1 H355A , ( d ) Flag-SIRT1 , and ( e ) SIRT1 shRNA . At 96 h after differentiation , batches of C2C12 myobutes were treated with 20 µM or 50 µM resveratrol or vehicle for the time periods given in the figures , or with 10 mM nicotinamide or vehicle for 24 h . Homogenates were prepared and Western blotting was performed as described previously [57] using the antibodies described previously [57] , [58] . To evaluate the effect of SIRT1 on PGC-1α transcription coactivator activity , C2C12 myoblasts were co-transfected with a gal-4-DBD PGC-1α plasmid , and a 9×gal-4-dependent reporter plasmid , or with a LacZ control plasmid , and with either wild-type SIRT1 , dominant-negative SIRT1 H355A , or SIRT1 shRNA-plasmids using lipofectamine 2000 . After overnight transfection the culture medium was changed to DMEM containing 10% FBS . Thirty-six hours later , some of the cells were treated with 20 µM resveratrol for 6 h and harvested after a 6 h recovery period . Dual luciferase assays were performed using a kit ( Invitrogen ) according to the manufacturer's instructions . Flag-PGC-1α was expressed in C2C12 myotubes by adenoviral infection . To evaluate the effect of SIRT1 on PGC-1α acetylation , the myotubes were co-infected with wild-type SIRT1 or SIRT1 H355A . Forty-eight hours after infection , myotubes were treated with 50 µM resveratrol or vehicle for 18 h . Wild-type C2C12 myotubes were treated with 10 mM nicotinamide for 24 h . The myotubes were then harvested , and cell extracts containing 200 µg of protein were rotated with anti-Flag antibody at 4°C overnight . The following morning , agarose G beads were added and the samples were rotated at room temperature for 2 h . The agarose beads were washed 4 times with PBS and protein was eluted from the beads with 5× SDS buffer , which was boiled for 5 min . PGC-1α was measured with an anti-PGC-1α antibody , and levels of PGC-1α acetylation were then assessed with an anti-acetyl lysine antibody ( #9441 Cell Signaling Technology ) . Results are expressed as means ± SE . The significance of differences between two groups was determined using Student's t test . For multiple comparisons , significance was determined by one-way analysis of variance followed by post hoc comparison using Tukey significant difference method . | Studies on cultured muscle cells have shown that treatment with resveratrol , a chemical famously found in the skin of red grapes , stimulates the manufacture of new mitochondria . This has been attributed to the activation of the deacetylase SIRT1 either directly by resveratrol or indirectly via the activation of AMP-activated protein kinase ( AMPK ) . SIRT1 is then thought to deacetylate and activate the transcriptional coactivator PGC-1α , which in turn stimulates mitochondrial biogenesis . It has also been reported that feeding resveratrol to mice increases muscle mitochondria and results in improved running endurance . Here we further analyze the adaptive response of muscle mitochondria to resveratrol treatment to see if it mimics the response to endurance exercise . We find that feeding rats or mice large amounts of resveratrol did not increase muscle mitochondria . In these rodents , the bioavailability of oral resveratrol is low , and the resulting plasma level of resveratrol is far below the concentration required to activate AMPK . Contrary to previous reports we find that deacetylation by SIRT1 decreases PGC-1α activity and results in a decrease in mitochondria; moreover we show that the increase in mitochondria induced in cultured muscle cells by a high resveratrol concentration is due to the toxic activation of AMPK and , in turn , PGC-1α . However , this effect requires resveratrol concentrations that are very much higher than those attained by oral administration , and we conclude that oral resveratrol has no effect on mitochondrial biogenesis in skeletal muscle . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Experimental",
"Procedures"
] | [
"biology"
] | 2013 | Effects of Resveratrol and SIRT1 on PGC-1α Activity and Mitochondrial Biogenesis: A Reevaluation |
Ebola virus ( EBOV ) caused more than 11 , 000 deaths during the 2013–2016 epidemic in West Africa without approved vaccines or immunotherapeutics . Despite its high lethality in some individuals , EBOV infection can produce little to no symptoms in others . A better understanding of the immune responses in individuals who experienced minimally symptomatic and asymptomatic infection could aid the development of more effective vaccines and antivirals against EBOV and related filoviruses . Between August and November 2017 , blood samples were collected from 19 study participants in Lagos , Nigeria , including 3 Ebola virus disease ( EVD ) survivors , 10 individuals with documented close contact with symptomatic EVD patients , and 6 control healthcare workers for a cross-sectional serosurvey and T cell analysis . The Lagos samples , as well as archived serum collected from healthy individuals living in surrounding areas of the 1976 Democratic Republic of Congo ( DRC ) epidemic , were tested for EBOV IgG using commercial enzyme-linked immunosorbent assays ( ELISAs ) and Western blots . We detected antibodies in 3 out of 3 Lagos survivors and identified 2 seropositive individuals not known to have ever been infected . Of the DRC samples tested , we detected antibodies in 9 out of 71 ( 12 . 7% ) . To characterize the T cell responses in the Lagos samples , we developed an anthrax toxin-based enzyme-linked immunospot ( ELISPOT ) assay . The seropositive asymptomatic individuals had T cell responses against EBOV nucleoprotein , matrix protein , and glycoprotein 1 that were stronger in magnitude compared to the survivors . Our data provide further evidence of EBOV exposure in individuals without EVD-like illness and , for the first time , demonstrate that these individuals have T cell responses that are stronger in magnitude compared to severe cases . These findings suggest that T cell immunity may protect against severe EVD , which has important implications for vaccine development .
Ebola viruses including the five antigenically distinct species , Bundibugyo ebolavirus ( BDBV ) , Reston ebolavirus ( RESTV ) , Tai Forest ebolavirus ( TAFV ) , Sudan ebolavirus ( SUDV ) , and Zaire ebolavirus ( EBOV ) , are enveloped filamentous viruses belonging to the family Filoviridae [1–3] . These viruses carry negative strand RNA genomes approximately 19 kb in length that code for 7 structural proteins: the nucleoprotein ( NP ) , VP35 , the matrix protein ( VP40 ) , the glycoprotein ( GP ) , VP30 , VP24 , and the RNA-dependent RNA polymerase [1] . Since the discovery of EBOV in 1976 in the Democratic Republic of Congo ( DRC ) , BDBV , SUDV , and EBOV have caused sporadic epidemics of lethal hemorrhagic fever or Ebola virus disease ( EVD ) , largely in Central Africa , with case fatality rates of 23–90% [4–7] . In December 2013 , an outbreak of EBOV was detected in Guinea , which led to the largest ever recorded epidemic spanning seven West African countries . By 2016 , more than 28 , 000 symptomatic cases of EVD were reported with a case fatality rate of 40% [8] . Although there are several in the pipeline , there are currently no licensed vaccines against EBOV and treatment remains largely supportive . Richardson et al . recently conducted a survey of close household contacts of individuals who had severe EVD and showed that a significant portion of EBOV transmission events went undetected during the West African outbreak because some individuals contracted infection but had mild illness or were asymptomatic [9] . This work adds to a growing body of evidence suggesting that despite its high lethality in some individuals , EBOV infection can produce little to no symptoms in others [10] . Seroprevalence surveys conducted in Africa have historically described asymptomatic EBOV infection , however , due to uncertainty in serologic assays , consensus on the significance of these findings has not been reached . Nonetheless , while several hypotheses might explain minimally symptomatic and asymptomatic EBOV infection , including properties of the infecting virus ( e . g . less virulent isolate ) , low inoculum , route of transmission , host factors ( e . g . resistance through viral cell receptor polymorphism ) , or a robust innate and/or adaptive immune response are potential explanations . Antibody therapy is considered an effective and powerful treatment strategy against many infectious pathogens . While studies of whole-blood transfusion or serum as passive immunity for EBOV treatment have demonstrated limited efficacy , monoclonal antibodies have shown promise in animal models and have been tested in human clinical trials [11–15] . The most successful of these is the antibody cocktail ZMapp , which comprises three humanized murine monoclonal antibodies that target the EBOV GP . ZMapp reversed advanced disease and rescued 100% of rhesus macaques up to 5 days post-viral challenge; however , it did not meet a pre-specified threshold for efficacy in humans when tested during the 2013–2016 outbreak [13 , 16] . Immune depletion studies in the macaque EBOV model demonstrated that humoral responses were beneficial in containing virus , but CD8+ T cells were essential for vaccine-induced protection . These findings suggest that humoral immunity alone may not account for full recovery or secondary protection . T cell responses of appropriate quality and magnitude were shown to be important for human protection against EBOV [17–19] . During the West African outbreak , Ruibal et al . demonstrated unique mechanisms that regulate T cell homeostasis in fatal and non-fatal EVD cases , suggesting that EBOV infections can trigger T cell responses that may be more effective in some individuals than others [20] . More studies are needed to better understand the immunopathogenesis of EVD , especially in mildly symptomatic or asymptomatic cases . In this study , we adapted the modified anthrax lethal factor ( LFn ) delivery system to enable the detection and characterization of T cell responses in previously EBOV exposed individuals from Lagos , Nigeria , three years after the outbreak . We fused the EBOV ( Makona variant ) NP , VP40 , and the receptor binding subunit of GP , known as GP1 , to LFn . The LFn-EBOV recombinant proteins were expressed and used as antigens to stimulate peripheral blood mononuclear cells ( PBMCs ) in an ELISPOT assay . We report robust T cell responses in EVD survivors and in EBOV seropositive asymptomatic individuals .
The protocols used in this study were approved by the Health Research Ethics Committee at the College of Medicine , University of Lagos , Nigeria ( CM/HREC/09/16/055 ) and by the Internal Review Board ( IRB ) at the Harvard T . H . Chan School of Public Health ( Harvard Chan ) , Boston , USA ( Protocol number IRB16-1321 ) . All study participants included in the study were adults . The participants from Lagos provided written informed consent for the collection of samples and corresponding data were banked and de-identified prior to analyses . The DRC samples were collected under surveillance in 1976 . The samples were anonymized by the US CDC prior to shipment to Harvard Chan in 1985 . On 20 July 2014 an individual who travelled from Liberia presenting with fever was transported to a private hospital in Lagos; he denied contact with known EVD cases and was treated with antimalarial drugs [21] . The individual’s condition worsened over the next three days , at which point EVD was suspected . Filovirus and EBOV-specific PCR testing was performed by Lagos University Teaching Hospital ( LUTH ) and Redeemer’s University ( RUN ) African Centre of Excellence for Genomics of Infectious Diseases ( ACEGID ) , respectively , and the patient was confirmed EVD positive on 25 July 2014; he died on that day . During this period , many were exposed to the virus and all contacts were traced , placed under surveillance , and monitored for clinical features of EVD . The chain of EVD transmission from this single individual resulted in 19 laboratory-confirmed cases , 8 of whom succumbed to infection . Nigeria was declared EBOV free by the World Health Organization on 20 October 2014 . Between August and November 2017 , we recruited 3 EVD survivors with RT-PCR positive results for EBOV , 10 individuals with documented close contact with symptomatic EVD patients , who remained healthy during the outbreak , and 6 healthcare workers ( HCWs ) from a different hospital in Lagos with a low likelihood of previous EBOV exposure based on proximity to the outbreak epicenter and subsequent questionnaires to eliminate potential recall bias . All methods described subsequently were performed on the Lagos samples from the EVD survivors , documented EVD contacts , and controls . Additionally , an epidemic of EVD erupted in Yambuku , a small village near Yandongi , Democratic Republic of Congo ( formerly Zaire ) in 1976 [4] . In the course of epidemiological investigations of this outbreak , hundreds of serum samples were collected from residents of the surrounding areas . In 1985 , these serum samples were screened for antibodies to human immunodeficiency virus ( HIV ) [22] , and in 1988 additional screening was conducted at Harvard Chan , with excess samples stored . The availability of 71 archived samples from the first known EBOV outbreak offered an opportunity to assess EBOV-specific antibodies in serum collected from potentially EBOV exposed but presumed uninfected individuals . Only Western blot testing to EBOV fusion antigens was performed on the DRC samples for EBOV serostatus . PBMCs were separated from plasma and whole blood in EDTA tubes by Ficoll-Paque gradient density ( GE Healthcare Life Sciences , Pittsburgh , PA , USA ) and cryopreserved in freezing media ( 10% dimethyl sulfoxide [DMSO , Sigma-Aldrich , St . Louis , MO , USA] , 90% fetal bovine serum [FBS , ThermoFisher Scientific , Rockford , IL , USA ) at -80°C prior to transfer to liquid nitrogen . Plasma was separately aliquoted and immediately transferred to -80°C . RNA was extracted from all plasma samples using the QIAamp RNA Viral Kit ( Qiagen , Hilden , Germany ) according to the manufacturer’s instructions and was quantified using quantitative RT-PCR ( qRT-PCR ) for both EBOV and human ribosomal RNA ( 18S ) , as previously described [23] . Briefly , the assay mix included 3ul of RNA , 0 . 03umol/L sense and anti-sense primers , 5ul of X2 Power ZYBR Green RT-PCR Mix and 0 . 08ul of RT Enzyme Mix ( ThermoFisher Scientific , Rockford , IL , USA ) . The cycling conditions were 48°C for 30 min , 95°C for 10 min , followed by 45 cycles of 95°C for 15 sec , 60°C for 30 sec with a melt curve of 95°C for 15 sec , 55°C for 15 sec , 95°C for 15 sec . qRT-PCR was performed on the LightCycler 96 ( Roche , Indianapolis , IN , USA ) . The amplicons were cleaned using AMPure XP beads ( Beckman Coulter , Brea , CA , USA ) according to the manufacturer’s instructions and amplicon concentrations were converted to EBOV copies per microliter for quantification . All plasma samples collected from the Lagos study participants were screened by the ReEBOV IgG ELISA kit ( Zalgen Lab , Germantown , MD , USA ) and the EV-IgG ELISA kit ( MyBioSource , San Diego , CA , USA ) according to the manufacturers’ instructions . All Lagos study participants’ plasma were also screened by the ReEBOV Antigen ELISA kit ( Zalgen Lab , Germantown , MD , USA ) according to the manufacturer’s instructions . Measurement of OD was performed within 5 min of stopping the reaction at 450nm ( antibody ) and 620nm ( antigen ) with the MiltiskanTM GO Plate Reader ( ThermoFisher Scientific , Rockford , IL , USA ) . Commercially synthesized amino acid fragments corresponding to the EBOV ( Makona variant ) NP , VP40 , GP1 , and sGP were cloned into the LFn expression plasmid ( pET15bLFn ) , which contains a T7 promoter , histidine tag , and the terminal domain of the anthrax lethal factor . The EBOV NP , VP40 , GP1 , and sGP sequences were derived consensus sequences generated from the initial 99 genomes initially published during the 2013–2016 outbreak [24] . The pET15bLFn containing the coding sequences of EBOV NP , VP40 , and GP1 were transformed into E . coli BLR ( DE3 ) ( Millipore , Medford , MA , USA ) for expression . Clones containing the correct reading frame as verified by sequencing were used for protein expression , as previously described [25] . The pET15bLFn was expressed and purified as described above for use as a negative control . All plasma samples collected from both the Lagos and DRC study participants were screened for the presence of IgG antibodies to bind to recombinant LFn-EBOV-GP1 and -sGP . Briefly , 50ug LFn-EBOV-GP1 and -sGP were added to reducing buffer ( 2% SDS , 0 . 5 M Tris [pH 6 . 8] , 20% glycerol , 0 . 001% bromophenol blue , and 5% 2-Mercaptoethanol ) and subjected to 12% polyacrylamide gel electrophoresis ( PAGE ) . Western Blot analysis was conducted using plasma samples ( 1:100 ) as primary antibody and anti-human IgG HRP ( ThermoFisher Scientific , Rockford , IL , USA ) as secondary antibody . Visualization was performed using SuperSignal Femto Chemiluminescent Substrate ( ThermoFisher Scientific , Rockford , IL , USA ) and with the Chemi Doc XRS+ Imaging System ( Bio Rad Technologies , Hercules , CA , USA ) . ELISPOT assays were performed as previously described [26] . Briefly , a total of 2x105 PBMCs were incubated with 0 . 1ml complete RPMI 1640 medium ( ThermoFisher Scientific , Rockford , IL , USA ) in the presence of 2 . 5ug/ml final concentration LFn fusion proteins in antibody coated 96-well polyvinylidene difluoride ( PVDF ) -backed MultiScreenHTS ( MSIP ) microtiter plates ( Millipore , Medford , MA , USA ) . After 24 hr of incubation at 37°C , the plates were washed and incubated with secondary antibodies , followed by an overnight incubation at 4°C . Plates were washed , then incubated for 2 hr at room temperature with the enzymatic conjugate . Spots were developed using Vector Blue substrate solution ( Vector Laboratories , Burlingame , CA , USA ) and counted manually using a stereozoom microscope ( 20X magnification ) . In CD8 and CD4 experiments , CD8+ and CD4+ T cells were isolated from PBMCs using the CD8+ and CD4+ T cell Isolation kits ( Miltenyi Biotec , Auburn , CA , USA ) . EBOV-specific spots were calculated by subtracting the mean of the negative control values of the replicates from the mean values of the specific stimulation . Positive responses were greater than four times the mean background , three standard deviations above the background , and ≥55 spot-forming cells per ( SFC ) /106 PBMCs . Statistical analyses , including comparison between groups of individuals and between antigens , were performed using Prism 7 ( GraphPad , San Diego , CA , USA ) . Data are expressed as geometric positive means ± standard deviation . Non-parametric Mann-Whitney test was used to compare data and determine statistical significance . p<0 . 05 was considered significant .
At the time of sample collection , all Lagos study participants recruited were healthy . The median age was 34 years ( range , 24–65 years ) , and 37% were female . Study participant characteristics are summarized in S1 Table . We assessed Lagos plasma samples for the presence of EBOV nucleic acid , EBOV antigen , and EBOV-specific antibodies ( Table 1 ) . All study participants tested negative for EBOV by qRT-PCR and antigen capture ELISA . Using the Zalgen ReEBOV IgG ELISA , which assesses antibodies specific to EBOV VP40 , 3 of 3 EVD survivors and 2 of 10 documented EVD contacts tested positive , and all 6 control HCWs were negative . Using the MyBioSource EV-IgG ELISA , which uses inactivated and homogenized EBOV , 2 of 3 EVD survivors and the same 2 out of 10 EVD contacts tested positive while the control HCWs remained negative . Consistent with the Zalgen ReEBOV IgG ELISA , Western blot analysis using the LFn-EBOV-GP1 and -sGP fusion proteins revealed GP1- and sGP-specific IgG antibodies in the 3 EVD survivors and the same 2 out of 10 documented EVD contacts , with no detectable antibodies in the control HCWs ( Fig 1 ) . Of the serum samples collected during the 1976 EVD epidemic in the DRC , 9 of 71 ( 12 . 7% ) contained IgG antibodies that were reactive to LFn-EBOV-GP1 and/or -sGP as demonstrated by Western blot ( Table 2 , Fig 2 ) . Antibodies specific to LFn-EBOV-GP1 were found in 1 . 4% ( 1/71 ) , 9 . 6% ( 7/71 ) to LFn-EBOV-sGP and 1 . 4% ( 1/71 ) to both LFn-EBOV-GP1 and -sGP . We then assessed the post-infection cellular responses in all Lagos study participants by LFn fusion protein stimulation of PBMCs in IFN-γ and TNF-α ELISPOTs . All 3 EVD survivors and the 2 seropositive EVD contacts mounted detectable IFN-γ and TNF-α cellular responses to LFn-EBOV-NP , -VP40 , and/or -GP1 ( Table 3 , Fig 3 ) . The remaining 8 documented EVD contacts and all 6 control HCWs mounted cellular responses below the positive threshold ( Table 3 ) . In most cases , the EVD survivors and the seropositive contacts mounted the strongest IFN-γ and TNF-α cellular against LFn-EBOV-NP , p<0 . 05 . ( Fig 4 ) . IFN-γ and TNF-α cellular responses against LFn-EBOV-GP1 were stronger compared to responses against LFn-EBOV-VP40 , p<0 . 05 . Additionally , the seropositive EVD contacts mounted IFN-γ and TNF-α cellular responses against LFn-EBOV-NP , -VP40 , and -GP1 that were stronger in magnitude compared to survivors; similarly , IFN-γ responses against LFn-EBOV-NP and -GP1 and for TNF-α responses against LFn-EBOV-NP and -VP40 were stronger in contacts compared to survivors , p<0 . 05 . We further evaluated the post-infection CD8+ and CD4+ T cell responses in the EVD survivors and the seropositive EVD contacts . Due to sample availability , we assessed IFN-γ and TNF-α responses against LFn-EBOV-NP and -GP1 . In nearly all cases , mean IFN-γ and TNF-α CD8+ and CD4+ responses were stronger among the seropositive EVD contacts compared to the survivors , though not statistically significant ( Fig 5 ) . As the only exception , mean IFN-γ CD4+ response against LFn-EBOV-NP was stronger in the EVD survivors compared to the seropositive contacts , though not statistically significant ( Fig 5C ) .
The characterization of the immune response in mild and asymptomatic EBOV infection may explain how certain individuals infected with this otherwise highly lethal virus avoid severe disease . This could contribute to efforts in the development of more effective vaccines and immunotherapeutics against EBOV and related filoviruses . Therefore , we adapted the modified anthrax toxin delivery system to design LFn-EBOV fusion proteins for antibody and/or T cell analyses in EVD survivors , documented close EVD contacts , and in presumed healthy individuals living in nearby areas of the 1976 EVD outbreak in the DRC . T cell responses in survivors of EVD epidemics has been relatively well studied . Four patients with acute EVD demonstrated CD8+ and CD4+ T cell activation against several viral proteins , and this activation persisted for up to one month after infection [17] . Consistent with this study , prolonged T cell activation was observed in a single patient who cleared EBOV infection without the use of experimental drugs [27] . Another study evaluating patients during the acute phase of infection , demonstrated that survivors have T cells that express lower levels of the inhibitory molecules CTLA-4 and PD-1 , compared to fatal EVD cases with high viral load [20] . Additionally , a study of SUDV survivors , 12 years post infection , demonstrated strong memory CD4+ , but not CD8+ , T cell activation and neutralizing humoral immunity [28] . Whole irradiated SUDV was used to stimulate PBMC samples in these long-recovered survivors , likely contributing to the limited CD8+ activation . Moreover , minimally symptomatic and asymptomatic EBOV infections are not new phenomena . Using an immunofluorescence assay , World Health Organization researchers identified EBOV-infected individuals who had symptoms that ranged in severity , from mild to rapidly fatal , during the first outbreaks of EBOV in Zaire and Sudan in 1976 and 1979 , respectively [4 , 5 , 29] . Since then , a number of additional studies utilizing diverse methods have identified EBOV-infected individuals who nonetheless remained asymptomatic [30–34] . However , studies examining the cellular immune responses during mild or asymptomatic infection are scarce . In one study of Gabonese individuals with asymptomatic EVD infection , high concentrations of pro-inflammatory cytokines were detected in plasma samples; yet , no T cell-derived cytokines were observed [33] . A follow-up study in the same individuals demonstrated mRNA expression of T cell cytokines and cytotoxic activation markers , suggesting cytotoxic T cell activation; however , EBOV-specific activation was not studied [35] . In this study , we demonstrated for the first time , to our knowledge , EBOV-specific cellular responses in seropositive asymptomatic individuals as well as EVD survivors . The seropositive asymptomatic individuals mounted stronger IFN-γ and TNF-α responses to all three LFn-EBOV fusion proteins ( LFn-EBOV-NP , -VP40 , and -GP1 ) compared to the EVD survivors . Cellular responses were significantly stronger for IFN-γ responses to LFn-EBOV-NP and LFn-EBOV-GP1 and for TNF-α responses LFn-EBOV-NP and LFn-EBOV-VP40 . Consistent with previous studies , cellular responses directed to the EBOV NP were strongest compared to other EBOV antigens , in both the EVD survivors and the seropositive asymptomatic individuals . We also showed that seropositive asymptomatic individuals have IFN-γ and TNF-α cellular responses that were stronger when compared to the survivors . These results suggest that T cell immunity may play a protective role against severe EVD . Additionally , we detected EBOV-specific antibodies in serum collected from individuals living in surrounding areas of the 1976 EVD epidemic in the DRC . These results further support undiagnosed EBOV infection in individuals living in EBOV endemic regions . While the study of long recovered SUDV survivors was unable to elicit CD8+ T cells , our study demonstrates robust CD8+ T cells against the LFn-EBOV fusion proteins in 3-year post-infection EVD survivors and 2 contacts not known to have ever been infected by EBOV . The LFn-EBOV fusion proteins elicit specific and sensitive T cell responses with low background signals in ELISPOT . Thus , LFn-EBOV priming systems may offer utility as an alternative and inexpensive technology for ex vivo screening of EBOV-specific T cell responses in future vaccine clinical trials . Additionally , our detection of both EBOV-specific CD8+ and CD4+ T cell responses suggests that the LFn delivery system is capable of efficiently presenting exogenous EBOV proteins to the MHC class I and II pathways; therefore , the use of LFn-EBOV fusion constructs presents an attractive technology for EBOV antigen delivery in vaccine design . Our study has several limitations . Our study population is small and we recruited only a limited number of individuals with documented close EVD contact . It is likely that there are more individuals who were infected with EBOV and remained healthy during the outbreak in Nigeria . While we used different tests to identify individuals with markers of EBOV infection , we did not perform microneutralization assays to confirm infection specificity or define antiviral function of the antibodies detected in either the EVD survivors or the seropositive asymptomatic individuals . Future studies are expected to examine the functional characteristics of antibodies in minimally symptomatic and asymptomatic EBOV infections . We also demonstrated that the seropositive asymptomatic individuals have stronger T cell responses compared to the EVD survivors; however , additional studies with more study participants are needed to validate these results . We cannot rule out the possibility that the asymptomatic individuals were exposed to a low inoculum or dead antigen , precluding the development of EVD , and generated EBOV-specific T cell responses . Whether these CD8+ and CD4+ cells remain functional during secondary exposure to EBOV remains to be elucidated . Future studies investigating T cell functional markers such as PD-1 and CTLA-4 among the seropositive asymptomatic individuals are expected . In conclusion , our findings raise new questions and highlight the need for further investigation to better understand the immune responses associated with minimally symptomatic and asymptomatic EBOV infections . By development of the LFn-EBOV ELISPOT assay , we detected cellular immune responses in EVD survivors and individuals with documented close contact with EVD patients , approximately three years after the outbreak in Lagos , Nigeria . These results suggest the importance of T cell responses in disease progression and have important implications for vaccine development , as well as for potential EBOV diagnostics based on T cell responses . | The 2013–2016 West African Ebola virus ( EBOV ) outbreak is the largest on record with over 28 , 000 reported symptomatic cases and more than 11 , 000 deaths . We developed a simple and inexpensive modified anthrax toxin-based ELISPOT assay to detect and characterize the T cell responses elicited by prior exposure to EBOV . Our data show robust T cell responses to several EBOV proteins in individuals who experienced both severe and asymptomatic EBOV infections . These results provide further evidence that EBOV transmission events can go undetected . We also show that the seropositive asymptomatic individuals have stronger T cell responses compared to survivors , which has important implications for vaccine development . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"blood",
"cells",
"cell",
"physiology",
"medicine",
"and",
"health",
"sciences",
"enzyme-linked",
"immunoassays",
"immune",
"cells",
"immune",
"physiology",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"immunology",
"microbiology",
"viruses",
"filoviruses",
"rna",
"viruses",
"cytotoxic",
"t",
"cells",
"infectious",
"disease",
"control",
"antibodies",
"immunologic",
"techniques",
"research",
"and",
"analysis",
"methods",
"immune",
"system",
"proteins",
"infectious",
"diseases",
"white",
"blood",
"cells",
"animal",
"cells",
"proteins",
"medical",
"microbiology",
"immunoassays",
"t",
"cells",
"microbial",
"pathogens",
"immune",
"response",
"biochemistry",
"cell",
"biology",
"ebola",
"virus",
"physiology",
"viral",
"pathogens",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"hemorrhagic",
"fever",
"viruses",
"organisms",
"cell",
"fusion"
] | 2018 | A modified anthrax toxin-based enzyme-linked immunospot assay reveals robust T cell responses in symptomatic and asymptomatic Ebola virus exposed individuals |
Leishmania parasites have evolved sophisticated mechanisms to subvert macrophage immune responses by altering the host cell signal transduction machinery , including inhibition of JAK/STAT signalling and other transcription factors such as AP-1 , CREB and NF-κB . AP-1 regulates pro-inflammatory cytokines , chemokines and nitric oxide production . Herein we show that upon Leishmania infection , AP-1 activity within host cells is abolished and correlates with lower expression of 5 of the 7 AP-1 subunits . Of interest , c-Jun , the central component of AP-1 , is cleaved by Leishmania . Furthermore , the cleavage of c-Jun is dependent on the expression and activity of the major Leishmania surface protease GP63 . Immunoprecipitation of c-Jun from nuclear extracts showed that GP63 interacts , and cleaves c-Jun at the perinuclear area shortly after infection . Phagocytosis inhibition by cytochalasin D did not block c-Jun down-regulation , suggesting that internalization of the parasite might not be necessary to deliver GP63 molecules inside the host cell . This observation was corroborated by the maintenance of c-Jun cleavage upon incubation with L . mexicana culture supernatant , suggesting that secreted , soluble GP63 could use a phagocytosis-independent mechanism to enter the host cell . In support of this , disruption of macrophage lipid raft microdomains by Methyl β-Cyclodextrin ( MβCD ) partially inhibits the degradation of full length c-Jun . Together our results indicate a novel role of the surface protease GP63 in the Leishmania-mediated subversion of host AP-1 activity .
Parasites of the Leishmania genus are the causative agent of leishmaniasis; a disease distributed worldwide affecting more than 12 million people in 88 countries [1] . Leishmaniasis is a complex of diseases ranging from self-healing cutaneous lesions to lethal visceral afflictions [2] . In its mammalian host , Leishmania is an obligate intracellular pathogen infecting hematopoietic cells of the monocyte/macrophage lineage . Macrophages are specialized for the destruction of invading pathogens and priming the immune response . In order to survive within these cells , Leishmania has evolved sophisticated mechanisms to subvert macrophage microbicidal functions such as inhibition of nitric oxide ( NO ) production and cytokine-inducible macrophage functions [3] . This occurs as the direct consequence of parasite-mediated activation of protein tyrosine phosphatases , alteration of signal transduction and inhibition of nuclear translocation and activity of transcription factors such as NF-κB , STAT , CREB and AP-1[4] , [5] . Activated Protein-1 ( AP-1 ) is an important transcription factor that mediates gene regulation in response to physiological and pathological stimuli , including cytokines , growth factors , stress signals , bacterial and viral infections , apoptosis , as well oncogenic responses [6] , [7] . AP-1 is formed by homodimers of Jun family members ( c-Jun , Jun B and Jun D ) , or heterodimers of Jun and Fos family members ( c-Fos , Fos B , Fra 1 and Fra 2 ) . Homodimers within the Fos family do not occur due to conformational repulsion [8] . Previous studies have reported that the AP-1 transcription factor is inactivated by Leishmania infection . For instance , activation of macrophage AP-1 and NF-κB is inhibited by L . donovani promastigotes through an increase in intracellular ceramide concentration , which leads to the down-regulation of classical PKC activity , up-regulation of calcium independent atypical PKC-ζ and dephosphorylation of Extracellular Signal-Regulated Kinases ( ERK ) [9] , [10] . Other studies have shown that Leishmania alters signal transduction upstream of c-Fos and c-Jun by inhibiting ERK , JNK and p38 MAP Kinases , resulting in a reduction of AP-1 nuclear translocation [11] , [12] . However , little is known about the molecular mechanism ( s ) by which Leishmania parasites are able to inactivate this important transcription factor . Many Leishmania-specific factors such as lipophosphoglycan ( LPG ) , A2 proteins , cysteine peptidases ( CPs ) and the protease GP63 , contribute to Leishmania virulence and pathogenicity . LPG has been implicated in altering phagosome maturation in L . donovani infection [13] . The A2 proteins of L . donovani are involved in intracellular amastigote survival [14] . The cysteine peptidases of L . mexicana are implicated in facilitating the survival and growth of the parasite [15] . Furthermore GP63 , also known as the major surface protease ( MSP ) , has been related to resistance to complement-mediated lysis , among others [16] , [17] . GP63 is a metalloprotease which belongs to the metzincin class . It is the most abundant surface glycoprotein of the parasite and accounts for 1% of the total protein content of L . mexicana promastigotes [18] . GP63 of different Leishmania species encode similar amino acid sequences , although slight substrate specificity variations have been reported [19] . Specific characteristics of this class of metalloproteases include a conserved signature motif HEXXHXXGXXH and an N-terminal pro-peptide that serves to maintain the pro-enzyme inactive during translation , which is removed upon protein maturation and activation [20] . The mature GP63 contains 3 domains: 1 ) N-terminal ( bases ∼101-273 ) which comprises a structure corresponding to the catalytic module of metzincin class zinc protease , 2 ) central domain ( bases ∼274–391 ) and 3 ) C-terminal domain containing the site of glycosylphosphatidylinositol ( GPI ) anchor addition ( bases ∼392–577 ) [17] , [18] , [20] , [21] . We have previously shown that this protease actively participates in the cleavage of NF-κB [5] , protein tyrosine phosphatases ( PTP ) [22] and actin cytoskeleton regulators [23] . In this study we have investigated how GP63 contributes to the inactivation of AP-1 and the degradation of its subunits . Herein , we report that GP63 enters the host cell via lipid raft microdomains , independently of parasite internalization , and for the first time show that it is able to reach the nuclear compartment shortly after infection where it degrades and cleaves c-Jun and other AP-1 subunits .
We have previously studied the effect of Leishmania promastigote infection on the activity of various macrophage transcription factors: STAT-1α degradation is proteasome and receptor-dependent and is mediated through a mechanism involving PKC-α [4] , and cleavage of NF-κB subunits upon Leishmania infection is in part dependent on GP63 [5] . As AP-1 is an important transcription factor regulating the expression of many genes involved in the activation of macrophage functions ( TNFα , iNOS , and IL-12 ) [24] , [25] , [26] critical for the adequate innate immune response against Leishmania infection , we investigated the mechanisms underlying AP-1 inactivation upon Leishmania infection . To evaluate nuclear translocation and DNA binding activity of macrophage AP-1 upon infection with Leishmania promastigotes , Electrophoretic Mobility Shift Assays ( EMSA ) were performed . As shown in Figure 1A , AP-1 nuclear translocation was inhibited as early as 30 min post-infection in L . donovani-infected macrophages . Furthermore , we observed that other mammalian pathogenic Leishmania species ( L . mexicana and L . major ) were able to alter AP-1 DNA binding ( Figure 1B ) . Of interest , we did not observe any effect on macrophages infected with L . tarentolae , whose pathogenicity is limited to reptilian hosts . In order to better understand the observed decrease in AP-1 activity , we performed Western Blot ( WB ) analysis in the total cell extracts to evaluate the various AP-1 subunits during infection . Five out of the seven AP-1 subunits ( c-Fos , Fra 1 , Fra 2 , c-Jun and Jun B ) showed decreased expression after infection with L . donovani , whereas Jun D presented a slight reduction and Fos B was maintained intact ( Figure 2A ) . To further confirm the presence of these subunits in the AP-1 complex we used super shift analysis . This approach uses the incorporation of specific antibodies to nuclear protein extracts , allowing the visualization of the antibody: protein: DNA complexes by retarding the migration of the specific bands in the gel . As shown in Figure 2B , inclusion of antibodies specific for Fos B , c-Fos , Fra 1 , Fra 2 , c-Jun , Jun B and Jun D , demonstrated presence of Fra 1 , Fra 2 , c-Jun , Jun B and Jun D , but not c-Fos or Fos B , within the macrophage nuclear AP-1 complex . Importantly , L . donovani infection clearly affected the AP-1 complex as the bands observed for Fra-1 , Fra-2 , c-Jun , Jun B and Jun D in the super shift assay was greatly reduced . Whereas the c-Fos protein was not detectable by super shift assay , this protein was still affected by Leishmania infection since less expression was observed by WB ( see Figures 2A and 3 ) , suggesting that the amount of c-Fos might not be enough to be detected by super shift assay . After phosphorylation AP-1 subunits are translocated into the nucleus where they dimerize with another subunit to form an active AP-1 complex [6] , [7] , [8] , [27] . To determine the level of expression of each AP-1 subunit in the different cellular compartments ( cytoplasm vs nucleus ) , we performed WB analysis on separated nuclear and cytoplasmic fractions . As shown in Figure 3 , different phenomena can be observed . c-Fos and Fra-1 expression in the cytoplasmic fraction are not altered with Leishmania infection , but their expression in the nuclear fraction is decreased in infected macrophages; Fra-2 and c-Jun have decreased expression in both cytoplasmic and nuclear fractions , and Fra-2 in the nuclear fraction presents a band with less migration than the band observed in the cytoplasmic fractions , possible due to post-nuclear translocation modifications . On the other hand , Jun-B and Jun-D were detected only in the nuclear fraction; however , only Jun-B expression is affected by Leishmania infection . The lower expression of the different subunits in the nucleus could be due to decreased complex formation and/or cleavage and further degradation of the subunits , as it is possible to detect smaller bands ( c-Jun and Jun-B ) . Leishmania surface molecules such as LPG and GP63 , among others , play important roles as virulence factors and modulators of host cell signalling . LPG , for instance , has been implicated in the interference of phagolysosome maturation and inactivation of PKC signalling [13] , [28] . GP63 has been related to resistance to complement-mediated lysis , migration of Leishmania parasites through the extracellular matrix by degradation of casein , fibrinogen and collagen [16] , [21] and inhibition of JAK/STAT signalling by modulation of PTP activities [22] . To address the role of LPG and GP63 in AP-1 inactivation we performed EMSA with extracts from cells infected with Leishmania mutants for these two surface molecules . As shown in Figure 4A , LPG is not involved in the AP-1 degradation induced by Leishmania infection since DNA binding in macrophages infected with either L . donovani or L . donovani LPG−/− promastigotes was similarly altered . Importantly , however , we observed that cells infected with an L . major strain lacking GP63 ( L . major GP63−/− ) [29] showed normal AP-1 DNA binding capacity , compared to uninfected controls . This suggests that GP63 but not LPG is highly involved in the mechanism responsible for the inactivation of AP-1 transcription factor . To further elucidate the role of GP63 , we performed WB analysis of all the AP-1 subunits of macrophages infected with L . major , L . major GP63−/− and L . major GP63 Rescued . Results obtained further revealed that in the absence of GP63 , no degradation or cleavage of any AP-1 subunit was evident ( Figure 4B ) ; supporting the finding that AP-1 activity is unaffected in L . major GP63−/−-infected macrophages . In addition , to validate the role of GP63 on AP-1 activity we verified the expression of IL-12 transcripts , as it is known that AP-1 regulates its transcription [24] . As shown in the Figure S1 , LPS-induced IL-12 expression is fully blocked by all infectious Leishmania species but not by L . major GP63−/− and L . tarentolae . As expected , the JNK/c-Jun inhibitor has completely inhibited LPS-induced IL-12 transcripts . Leishmania GP63 can be found in three different forms: 1 ) Intracellular GP63 , 2 ) Surface GPI-anchored GP63 and 3 ) secreted or released GP63 [21] , [30] . For GP63 to target its intracellular macrophage substrates , it needs to gain access to or be internalized by the macrophage . To explore whether the internalization of the parasite is necessary to deliver GP63 inside the cell , murine macrophages were pre-treated with the phagocytosis inhibitor cytochalasin D which inhibits actin polymerization , therefore blocking internalization by phagocytosis ( Figure S2A ) . We used c-Jun as a model protein to evaluate the cleavage and degradation of the AP-1 subunits . WB analysis showed that parasite phagocytosis was not necessary for c-Jun cleavage and less expression ( Figure 5A ) . To confirm this , we incubated macrophages with the culture supernatant of L . mexicana promastigotes , which is rich in soluble GP63 [18] , [31] . WB showed that even in the absence of the parasite , c-Jun degradation was observed ( Figure 5B ) . Since phagocytosis seems not to be completely required in the internalization of GP63 we addressed whether GP63 internalization could be dependent on lipid raft-mediated endocytosis , given the fact that GP63 is an excreted and membrane-GPI anchored protein . On the other hand , lipid raft microdomains are highly dynamic membrane domains rich in cholesterol and sphingolipids , and present high affinity for proteins containing GPI anchors [32] , [33] , [34] . In order to examine the possible role of host lipid raft microdomains in GP63 internalization , we pre-treated cells with a non-cytotoxic dose ( Figure S2B ) of the cholesterol chelator and inhibitor of lipid raft integrity methyl-β-clyclodextrin ( MβCD ) prior to infection . As shown in Figure 5C , full length c-Jun was not degraded in cells infected under these conditions , although interestingly , a cleavage fragment was still observed . Pre-treatment of macrophages with MβCD and subsequent incubation with L . mexicana supernatant showed that lipid raft disruption altered internalization of parasite-free soluble GP63 and also impaired c-Jun degradation ( Figure 5D ) . As shown in Figure S3 , confocal microscopy confirmed an interaction between lipid raft microdomains and GP63 , since in macrophages infected with L . major , GP63 ( green ) partially co-localized with the lipid raft marker Choleratoxin B ( red ) . Furthermore , we have previously shown that , pre-treatment with MβCD before infection abrogates GP63 internalization [22] . To determine if MβCD had any effect over the c-Jun expression we performed a time course analysis of macrophages stimulated with MβCD . As shown in Figure S4A , there was no alteration in the expression of c-Jun after 2 hr of incubation of the macrophages with the drug . Together these data strengthen the hypothesis that GP63 uses lipid raft microdomains for internalization independent of parasite entry . To evaluate the role of the GPI anchor in mediating GP63 internalization via lipid raft microdomains , macrophages were incubated with a GPI-deficient recombinant GP63 ( rGP63 ) and c-Jun degradation was monitored . WB analysis evidenced that neither degradation nor cleavage of c-Jun occurred ( Figure 5E ) in the presence or absence of MβCD , similarly to what we have previously shown for GP63-mediated PTP cleavage . Moreover , although rGP63 is still internalized in macrophages to a limited extent , perinuclear localization was never detected [22] . To demonstrate that the less expression and cleavage of c-Jun observed in this set of experiments were occurring inside the cells and not as an effect of proteolysis during the preparation of the lysates , we included two experiments as controls; first , we lysed the cells using sample loading buffer 1× and the samples were boiled right after , to stop the proteolysis; second , we added 1 mM of phenanthroline ( a Zn chelator [35] ) to the lysis buffer to abrogate post-infection GP63 activity . In both experiments , we observed that cleavage of c-Jun under these conditions still occurs , suggesting that the cleavage of c-Jun occurs inside the cell and not during the sample preparation ( Figure S4B and S4C ) . In addition , to establish whether GP63 proteolytic activity is critical for c-Jun cleavage in the macrophage , L . mexicana culture supernatant was treated with the GP63 inhibitor phenanthroline prior to its incubation with macrophages . As shown in the Figure S4D , phenanthroline fully inhibited GP63-mediated c-Jun degradation . One of the most surprising elements of the evidence presented above was the fact that GP63 is able to act on its substrate proteins within the nucleus of its host cell . In order to further demonstrate that GP63 reaches the nucleus , we separated cytoplasmic and nuclear proteins from macrophages infected with L . major , L . major GP63−/− and L . major GP63 Rescued . WB analysis using an anti-GP63 antibody revealed that this protease is present in both fractions of Leishmania-infected cell extracts . As expected , there was no GP63 in macrophages infected with L . major GP63 −/− or the uninfected control ( Figure 6A ) . Confocal microscopy of Leishmania-infected macrophages confirmed that GP63 reaches the nuclear membrane as early as 1 hr post-infection ( Figure 6B ) . In order to demonstrate the purity of our fractions , we performed WB of the cytoplasmic and nuclear proteins against the lysosomal marker LAMP-1 , the ER specific marker ( the KDEL protein - Lys-Asp-Glu-Leu endoplasmic reticulum protein retention receptor ) , histone 2B ( nuclear marker ) , and actin ( cytoplasm marker ) . Figure S5 shows that actin , LAMP-1 and KDEL are only present in the cytoplasmic fraction , in contrast , histone is only detected in the nuclear fraction , and this way we are confident to say that GP63 was present in both protein fractions . In order to confirm nuclear interaction of GP63 with c-Jun we performed a Co-Immunoprecipitation ( IP ) assay . c-Jun was immunoprecipitated from nuclear extracts of Leishmania-infected macrophages and subjected to WB analysis of GP63 . This result revealed a band around 65 kDa , confirming the interaction between nuclear c-Jun and GP63 in the macrophages infected with L . major and L . major GP63 Rescue , but not with L . major GP63 −/− as is shown in Figure 7 . To further support that degradation of c-Jun could occurs in the nucleus we performed confocal microscopy . As shown in Figure 8A ( upper panel ) , c-Jun ( red ) is localized inside the nucleus in uninfected cells . However , after 1 hr of infection GP63 was detected in the perinuclear area and the fluorescence intensity of c-Jun was considerably diminished ( Figure 8A , lower panel ) , such reduction in the fluorescence was not observed in macrophages infected either with L . major GP63−/− or L . tarentolae ( Figure 8C upper and lower panels , respectively ) . The upper panel of Figure 8B shows partial co-localization between the nuclear stain ( blue ) and GP63 ( green ) in the periphery of the nucleus , giving a light blue signal . Of utmost importance , the panel representing c-Jun ( red ) versus nucleus ( blue ) co-localization , clearly reveals that c-Jun is absent from perinuclear area as this one is solely stained in blue ( Figure 8B , lower panel ) . To discard possible unspecific signals in the confocal micrographs we included specific isotype and secondary antibody controls ( Figure S6B ) . Collectively , our results suggest that GP63 reaches the perinuclear area of the cell shortly after macrophage-parasite contact occurred leading to degradation and cleavage of various members of AP-1 subunits , leading to its inability to dimerize and bind DNA and therefore , altering AP-1 transcriptional activity on genes under its regulation . To further understand the direct effect of GP63 on c-Jun , we used a purified GST tagged-c-Jun protein and incubated it with Leishmania promastigotes of different species ( including L . donovani , L . mexicana , L . major , L . major GP63−/− and L . major GP63 Rescued ) . WB analysis showed that direct contact of parasites expressing GP63 and c-Jun protein is sufficient to induce c-Jun degradation ( Figure 9A ) . This was corroborated by the reduction of c-Jun degradation when incubated with the GP63−/− strain . Figure S7 shows that L . tarentolae has no effect on the degradation of GST-c-Jun . GP63 recognizes a four amino acid motif in its target protein substrates based on amino acid characteristics: polar/hydrophobic/basic/basic amino acids ( P1- P′1-P′2-P′3 ) [19] . Sequence analysis of AP-1 subunits revealed putative cleavage sites within c-Jun , Jun B , Jun D and c-Fos ( Figure 9B ) . Of interest , one of the sequence-identified cleavage sites of c-Jun was found between the leucine zipper and the DNA binding domain as shown in Figure 9C . In addition , this motif is found at amino acids 271-275 , which will generate cleaved fragments with molecular weight similar to the one detected by WB of lysates from infected cells ( ∼30 kDa ) . Collectively , these data indicate that the Leishmania protease GP63 actively participates in altering the DNA binding capacity of AP-1 as a consequence of the diminished expression and cleavage of its subunits . These data further corroborate a mechanism whereby GP63 can enter the cell using lipid raft microdomains , and show for the first time that GP63 reaches the perinuclear area where it proteolytically degrades AP-1 subunits .
Leishmania parasites have evolved many mechanisms to undermine macrophage signalling pathways in order to survive and replicate inside these cells . For instance , parasite-mediated activation of macrophage PTPs leads to protein dephosphorylation resulting in the inactivation of transcription factors controlling the expression of many genes required for the effective activation of the innate immune response [36] , and macrophage effector functions such as NO production [37] . We have previously reported that Leishmania promastigote infection induces degradation and inactivation of some transcription factors . For example , STAT 1 is inactivated by a proteasome mediated mechanism [4] , and NF-kB activity is altered in a cleavage-dependent fashion [5] . We show that cleavage of p65 generates an active fragment , p35 , which is able to translocate into the nucleus , where it dimerizes with p50 to induce specific chemokine gene expression . Interestingly this cleavage event was found to occur in the macrophage cytoplasm in a GP63-dependent mechanism [5] . Along with STAT and NF-κB , AP-1 is responsible for the transcription of iNOS [38] . NO is a by-product of iNOS-mediated conversion of L-arginine to L-citruline and is essential for the control of Leishmania infection [3] , [36] . Among other genes regulated by AP-1 in macrophages and known to be affected by Leishmania infection are TNFα , IL-1β and IL-12 [24] , [25] , [26] , [39] . The breadth and importance of the immunological functions of AP-1 highlights how detrimental its degradation is to host defence against Leishmania infection . Previously , Descoteaux and Matlashewski ( 1989 ) demonstrated that the c-fos gene , one of the main activators of AP-1 , was down-regulated due to abnormal PKC signalling [40] . More recently Ghosh and colleagues ( 2002 ) reported Leishmania-dependent inactivation of both AP-1 and NF-κB in a ceramide dependent mechanism , where increased levels of intracellular ceramide conducted to the down-regulation of classical PKC activity and impartment of the phosphorylation of ERK , which results in decreased AP-1 activation [10] . These previous reports have given some indication of AP-1 inactivation by Leishmania . Here we further demonstrated the molecular mechanisms involved in the AP-1 inactivation by Leishmania parasites and its impact on IL-12 expression . We have found that infection with several Leishmania species alters the DNA binding capacity of AP-1 . In particular , we have shown that the parasite metalloprotease GP63 is responsible for this DNA binding alteration and is able to induce the degradation/down-regulation and cleavage of c-Jun , the central component of the AP-1 transcription factor [41] , as well of other components including c-Fos , Fra-1 , Fra-2 and Jun B . We provide evidence that GP63 exerts its effect by its internalization into the host cell , in a mechanism that is independent of parasite internalization , and induces AP-1 proteolysis within the nucleus or in the nuclear membrane . The present study corroborates along with previous studies ( Ghosh et . al ) that AP-1 is down-regulated by Leishmania parasites . Alteration of AP-1 activity varies according to the pathogen , for instance it has been shown , that the hepatitis C virus alters MAP kinases and AP-1 to accelerate the cell cycle progression , helping the development of hepatocellular carcinoma and HCV development [42] . Another example is the Edema toxin produced by Bacillus anthracis; this toxin is able to up-regulate macrophage gene expression , among them genes that are known to be involved in inflammatory responses , regulation of apoptosis , adhesion , immune cell activation , and transcription regulation . Interestingly this up-regulation was found to correlate with induced activation of AP-1 and CAAAT/enhancer-binding protein-beta [43] . In contrast with these reports where different pathogens up-regulate AP-1 to survive inside the host cell , herein we have shown how this transcription factor is down-regulated after Leishmania infection in a cleavage-dependent manner . Whether AP-1 down-regulation is a general mechanism used by different intracellular protozoa requires further investigation . The AP-1 transcription factor is formed by dimers of Jun and Fos family members . In addition , the Jun proteins can dimerize with other proteins that share the leucine zipper region such as ATF-1 and ATF-2 [8] , [27] . Although we did not test other non-classical AP-1 subunits , we demonstrated that at least 5 of the classical subunits belonging the Jun and Fos families are degraded by the parasite within 1 hr of infection . Of interest , c-Jun subunit , one of the main activators of AP-1 along with c-Fos , is cleaved generating a GP63-mediated 30 kDa fragment . The cleaved product would be unable to dimerize and bind DNA , as it has been demonstrated that truncated c-Jun deprived of either the leucine zipper or the DNA binding domain results in only marginal AP-1 transactivation [41] , [44] , [45] . The generation of c-Jun fragment by GP63 can explain the lower AP-1 binding activity observed in the EMSA experiment . Furthermore , Fos B , which is not cleaved or degraded and also apparently absent in AP-1 complexes ( Figure 2B ) , lacks putative GP63 cleavage sites . Surprisingly Jun D presents two putative sites of cleavage by GP63 . However , we did not detect either complete degradation/down-regulation or cleavage products . One possible explanation is that the structural conformation of this protein renders these sites unavailable for GP63-Jun D interaction . GP63 is known to interact with various substrates . For instance we have recently reported that Leishmania GP63 impacts the stability of cortactin and caspase-3 , and negatively regulates p38 kinase activity [23] . Furthermore , we have shown that GP63 cleaves host PTPs resulting in enzymatic activation and leading to JAK 2 dephosphorylation , and inhibition of NO production in IFN-γ primed and infected macrophages [22] . Our current study further supports the important role of GP63 as a negative regulator of host cell functions , actively participating in the pleiotropic effects excreted by Leishmania parasite to suppress the immune response , our results showed that internalization of GP63 by cells from innate immune response is independent of parasite internalization , as our data revealed GP63 proteolytic activity was not affected by inhibition of phagocytosis , but clearly abolished by a lipid raft disruptor , strongly suggesting that lipid rafts microdomains are important for internalization of GP63 . Proteins that have a GPI anchor have affinity for lipid rafts , and it has been reported that these rafts recognize these GPI anchors allowing the entrance of GPI-bearing proteins in endocytic vesicles [33] , [34] . In addition , Brittingham and collaborators showed interaction of GP63 with the fibronectin receptor ( α4β1 ) , that also translocate into the lipid rafts microdomains [46] , suggesting that GP63 could have two different ways to get access into the cell: 1 ) GPI-anchor ( native and excreted ) and 2 ) receptor mediated ( RGP63 ) . Additionally , we have shown that the GPI anchor is important for the internalization of GP63 since recombinant GP63 ( rGP63 ) lacking the GPI anchor is less internalized [22] . Most importantly , GPI anchor seems to be required for the cellular localization of GP63 since rGP63 is localized inside intracellular compartments whereas GPI-GP63 ( native protein ) is found within nuclear membrane ( see Figure 6 ) . Despite this evidence we have not excluded the possibility that GP63 could use other mechanisms to enter the cells , such as micropinocytosis or classical endocytosis pathways . One of the main finding of this research is the macrophage nuclear localization of GP63 . One plausible mechanism is by the recognition of its GPI domain by the recently described lipid microdomains rich in cholesterol and sphingolipids in the nuclear membrane [47] . Another possible mechanism for the internalization of GP63 inside the nucleus is the presence of a nuclear localization signal ( NLS ) -like motif ( Figure S8 ) in the GP63 sequence . Nuclear proteins are usually transported inside the nucleus by recognition of a NLS , which usually consist of short chains of basic amino acids with the signature motif K-K/R-X-K/R [48] . These NSLs are recognized by the adaptor molecule importin α , which forms a hetorodimer with the transporter receptor importin β . The importin α/β-NSL-cargo complex is then translocated through the nuclear pore complex [48] , [49] . The exact mechanisms of how the nuclear translocation of GP63 occurs are currently under investigation . In summary , GP63 seems to preferentially target AP-1 subunits within the nuclear membrane , altering its DNA binding capacity . Given the critical role of this transcription factor in the transcription of several genes involved in the innate immune response , alterations in AP-1 activity can dramatically contribute to the down-regulation of innate immune functions observed during the early stages of Leishmania infection . Therefore , this novel mechanism of evasion by Leishmania further demonstrates the complex negative regulatory mechanisms developed by the parasite , which has permitted its adaptation to the harsh intracellular environment leading to its survival and propagation within its mammalian host .
Immortalised murine bone marrow derived macrophages B10R cell line were maintained at 37°C in 5% CO2 in Dulbecco's Modified Eagle medium ( DMEM ) supplemented with 10% heat inactivated FBS ( Invitrogen , Burlington , ON , Canada ) and 100 U/ml penicillin 100 µg/ml streptomycin and 2 mM of L-glutamine ( Wisent , St-Bruno , QC , Canada ) . Leishmania promastigotes ( L . donovani infantum , L . donovani 1S2D , L . donovani R2D ( LPG −/− ) , L . mexicana , L . major A2 ( WT ) , L . major GP63 −/− , L . major GP63 Rescued [29] and L . tarentolae ) were grown and maintained at 25°C in SDM-79 culture medium supplemented with 10% FBS by bi-weekly passage . Macrophages were infected at parasite to macrophage ratio 20∶1 with stationary phase promastigotes for the times specified in each Figure legend . Using this ratio of infection we normally observe around 30% and 60% of infected cells in 1 or 2 hr , respectively . When chemical inhibitors were used , 2 µM cytochalasin D ( Sigma-Aldrich , St-Louis MO , USA ) and 20 mM Methylβ-cyclodextrin ( MβCD ) ( Sigma-Aldrich , St-Louis MO , USA ) , cells were treated 1 hr prior to infection and the inhibitor remained throughout the infection time . B10R macrophages ( 2×106 ) were infected , washed three times with Phosphate Buffered Saline ( PBS ) to remove non-internalized parasites , and processed for nuclear extraction as previously described [4] , [50] . Briefly , macrophages were collected in 1 ml of cold PBS , centrifuged and pellets were resuspended in 400 µl of ice-cold buffer A ( 10 mM HEPES , 10 mM KCl , 0 . 1 mM EDTA , 0 . 1 mM EGTA , 1 mM DTT and 0 . 5 mM of PMSF ) and incubated 15 min on ice . Twenty five µl of IGEPAL 10% ( Sigma-Aldrich , St-Louis MO , USA ) were added , and samples vortexed for 30 sec . Nuclear proteins were pelleted by centrifugation and resuspended in 50 µl of cold buffer C ( 20 mM HEPES , 400 mM NaCl 1 mM EDTA , 1 mM EGTA 1 mM DTT and 0 . 5 mM PMSF ) . Protein concentrations were determined by Bradford assay ( Bio-Rad , Hercules CA , USA ) . 6 µg of nuclear proteins were incubated for 20 min at room temperature with 1 µl of binding buffer ( 100 nM Hepes pH 7 . 9 , 8% v/v glycerol , 1% w/v Ficoll , 25 mM KCl , 1 mM DTT , 0 . 5 mM EDTA , 25 mM NaCl , and 1 µg/µl BSA ) and 200 ng/µl of poly ( dI-dC ) , 0 . 02% bromophenol blue and 1 µl of γ-P32labeled oligonucleotide containing a consensus sequence for AP-1 binding complexes ( 5′-CGTTTGATGACTCAGCCGGAA-3′ ) ( Santa Cruz Biotechnology Inc , Sta Cruz CA , USA ) . After incubation , DNA-protein complexes were resolved by electrophoresis in non-denaturing polyacrylamide gel 5% ( w/v ) . Subsequently gels were dried and autoradiographed . Competition assays were conducted by adding a 100-fold molar excess of homologous unlabeled AP-1 oligonucleotide , or the non-specific competitor sequence for SP-1 binding ( 5′-ATTCGAATCGGGGCGGGGCGAGC-3′ ) . For supershift assay , 2 µg of nuclear protein extract were incubated for 1 hr at room temperature with binding buffer , poly ( dI-dC ) , 0 . 02% bromophenol blue , labeled oligonucleotide and 4 µg of individual specific antibodies ( α-c-Jun , Jun B , Jun D , c-Fos , Fos B , Fra 1 or Fra 2; Santa Cruz Biotech Inc , Sta Cruz CA , USA ) . Complexes were resolved on standard non-denaturing polyacrilamide gel 5% ( w/v ) . Infected and non infected cells ( 1×106 ) were washed 3 times with PBS and lysed with cold buffer ( 50 mM Tris ( pH 7 ) , 0 . 1 mM , 0 . 1 mM EGTA , 0 . 1% 2-mercaptoethanol , 1% NP-40 , 40 µg/ml aproptinin and 20 µg/ml of leupeptin ) . Proteins were dosed by Bradford ( Bio-Rad , Hercules CA , USA ) , and 30 µg of proteins were separated by SDS-PAGE , and transferred onto PVDF membranes ( GE healthcare , Piskataway NJ , USA ) . Membranes were blocked in 5% non-fat dry milk , washed and incubated for 1 hr with α-c-Jun ( BD Biosciences , San Jose , CA , USA ) , α-Jun B , α-Jun D , α-c-Fos , α-Fos B , α-Fra 1 and α-Fra 2 ( Santa Cruz Biotech Inc . Sta Cruz CA , USA ) . After washing , membranes were incubated 1 hr with α-mouse or α-rabbit HRP-conjugated antibody ( GE healthcare , Piskataway NJ , USA ) , and developed by autoradiography . B10R macrophages ( 10×106 ) were infected with either L . major A2 , or L . major GP63 −/− or L . major GP63 Rescued for 1 hr , and nuclear proteins were extracted as previously described in [5] . c-Jun was immunoprecipitated from pre-cleared nuclear extracts with 2 µg antibody , followed by addition of 12 . 5 µl ( packed volume ) of protein A/G PLUS agarose beads ( Santa Cruz Biotech Inc . Sta Cruz CA USA ) . Beads were washed three times and bound proteins were analyzed by WB as described above . B10R macrophages ( 0 . 5×106 ) were plated ON in glass cover slips . After infection for 30 , 60 and 180 min cells were gently washed 3 times with PBS , and then fixed with 4% p-formaldehyde for 30 min at 4°C . Slides were permeabilized for 5 minutes with PBS containing 1% BSA and 0 . 05% NP-40 and blocked with 5% non-fat dry milk for 1 hr . Incubation with primary antibody α-c Jun or α-c-Fos or α-GP63 mouse monoclonal antibody clone #96 [51] was conducted in humid dark chamber for 1 hr at room temperature . After three washes with PBS , cells were incubated with secondary antibody ( Alexa Fluor 488 or 594 , from Molecular probes , Burlington ON , Canada ) for 1 hr . Nuclei were stained with propidium iodide or DAPI for 10 min and slides were mounted in permaflour medium ( Thermo Co . Waltham MA , USA ) . Images were taken using an Olympus FV1000 confocal microscope and a Zeiss LCS 500 . B10R macrophages ( 10×106 ) were infected with either L . major A2 , L . major GP63−/− , L . major GP63 Rescued or L . tarentolae for 18 hr or treated with 20 µM of JNK inhibitor SP600125 for 1 hr . Thereafter cells were stimulated with 100 ng/ml of LPS for 6 hr and RNA extracted using TRIzol reagent according to the manufacturer's instruction ( Invitrogen Canada ) . Reverse transcriptase was performed using oligo ( dT ) primers . Quantitative relative PCR was performed using Invitrogen Platinum qPCR Super-Mixes and 0 . 4 µM primer in 25 µl and the following parameters: 50°C for 2 min and 95°C for 3 min ( 95°C for 20 s , 60°C for 30 s , 72°C for 20 s , 80°C ( reading step ) for 20 s ) for 40 cycles followed by a melting curve . Annealing temperature was 60°C . IL-12 primer sequences: 5′-GGA AGCACG GCA GCA GAA TA-3′ and 3′-AAC TTG AGG GAG AAG TAGGAA TGG-5′ . | Leishmaniasis is a tropical disease affecting more than 12 million people around the world . The disease is caused by the Leishmania parasites that are transmitted to the mammalian host by a sandfly vector when it takes a blood meal . The parasites are able to survive and multiply inside of cells that comprise the primary defence of the host , the macrophages . We have extensively studied the mechanism whereby Leishmania escapes from macrophage microbicidal functions . Herein we report that the parasite can inactivate these cells by decreasing the activity of transcription factors such as Activated Protein-1 ( AP-1 ) that are involved in transcription of genes coding for antimicrobial functions of macrophages . In this study , we showed that Leishmania parasites use their most abundant surface protein GP63 to inactivate the AP-1 transcription factor . Furthermore , we found that GP63 enter into the macrophages independently of parasite internalization using lipid rich microdomains localized in the cellular membrane . In addition , GP63 was observed to reach the nuclear compartment where it cleaves AP-1 subunit proteins . Collectively , our findings reveal a novel mechanism used by Leishmania to facilitate its survival and propagation within its mammalian host cells . Better knowledge concerning the mechanisms whereby this pathogen can escape the innate immune response may help to develop new anti-Leishmania therapy . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"immunology/immune",
"response",
"immunology/innate",
"immunity",
"microbiology/parasitology",
"infectious",
"diseases/protozoal",
"infections",
"immunology/immunity",
"to",
"infections"
] | 2010 | Leishmania-Induced Inactivation of the Macrophage Transcription Factor AP-1 Is Mediated by the Parasite Metalloprotease GP63 |
Conventional and regulatory T cells develop in the thymus where they are exposed to samples of self-peptide MHC ( pMHC ) ligands . This probabilistic process selects for cells within a range of responsiveness that allows the detection of foreign antigen without excessive responses to self . Regulatory T cells are thought to lie at the higher end of the spectrum of acceptable self-reactivity and play a crucial role in the control of autoimmunity and tolerance to innocuous antigens . While many studies have elucidated key elements influencing lineage commitment , we still lack a full understanding of how thymocytes integrate signals obtained by sampling self-peptides to make fate decisions . To address this problem , we apply stochastic models of signal integration by T cells to data from a study quantifying the development of the two lineages using controllable levels of agonist peptide in the thymus . We find two models are able to explain the observations; one in which T cells continually re-assess fate decisions on the basis of multiple summed proximal signals from TCR-pMHC interactions; and another in which TCR sensitivity is modulated over time , such that contact with the same pMHC ligand may lead to divergent outcomes at different stages of development . Neither model requires that T and T are differentially susceptible to deletion or that the two lineages need qualitatively different signals for development , as have been proposed . We find additional support for the variable-sensitivity model , which is able to explain apparently paradoxical observations regarding the effect of partial and strong agonists on T and T development .
Conventional T cells ( T ) and T regulatory cells ( T ) are essential components of the adaptive immune system . Conventional T cells develop effector function in response to foreign antigens , while natural T regulatory cells produced in the thymus play a key role in the maintenance of tolerance to self-antigens and prevent autoimmune diseases ( reviewed in , for example , [1] ) . Both populations are derived from precursors in the thymus that develop , undergo selection and differentiate into different T cell lineages . The differentiation of a thymocyte into the mature T cell repertoire is dependent on the engagement of its T cell receptor ( TCR ) with endogenous peptides presented by major histocompatibility complex ( MHC ) molecules on thymic antigen presenting cells . Continued very weak or null interactions between the TCR and peptide-MHC ligands ( pMHC ) lead to failure to positively select ( ‘death by neglect’ ) while excessively strong TCR-pMHC interactions lead to negative selection , removing highly autoreactive cells from the T cell repertoire . However , the precise rules underlying T cell precursor fate are not well understood; based on its exposure to a sample of pMHC , how and when does a thymocyte decide to become a T , a T , or be deleted ? Studies using fetal thymic organ cultures have shown that there exists a sharp avidity threshold between positive and negatively selecting ligands [2] , [3] . There is substantial evidence indicating that T are induced by TCR signals that lie below this negative selection threshold , but above that required for selection into the conventional T cell pool [4]–[8] . However , many uncertainties remain . It has been shown that expression of cognate antigen ( which we loosely refer to ‘agonist peptide’ ) in the thymic epithelium is required for the generation of T [9]–[13] , but a recent study showed that T commitment occurs over a wide range of TCR affinities for a ubiquitously expressed self antigen [14] . Further , the partitioning of fates with increasing strength of recognition for self ( deletionTTdeletion ) appears to be questioned by a study in which both expression of an agonist and a weaker partial-agonist could enhance deletion , but only the agonist was able to induce the formation of regulatory T cells [15] , suggesting that either the mapping of avidity to fate is more complex or that qualitatively different signals are required for T and T selection . Many experimental models using TCR transgenic cells ( clonal populations of T cells with identical TCR ) have shown that these cells can develop into both the regulatory and conventional lineages together in the same environment . This observation implies that there is stochasticity in fate determination . This stochasticity can be partitioned conceptually into two sources that are not mutually exclusive . First , there may be heterogeneity at the early double positive stage of development , even within a clonal population , that pre-disposes cells to different fates . This heterogeneity might derive , for example , from differences in expression of factors determining the baseline levels or dynamic range of TCR signalling , or other signalling proteins related to lineage commitment . Second , stochasticity may be present later in the selection process , arising at least in part because each thymocyte encounters an independent sample of self-peptide ligands . Evidence for the latter comes from observations that probabilities of deletion and T generation have been shown to vary with levels of agonist-peptide expression; in-vivo studies in TCR transgenic mice [16]–[18] and in-vitro fetal thymic organ culture [19]–[21] have shown that the efficiency of T selection increases with modest increases in agonist-peptide expression , but drops when expression is high . ( We use the term efficiency here interchangeably with the probability of experiencing a given fate . ) The efficiency of selection into the T lineage also decreases in the presence of increasing numbers of cells of the same specificity [14] , [15] , [22]–[24] . Thus the availability of relevant ligands , either in absolute terms or through competition , can influence fate decisions . The timing of an interaction with a ligand may also influence fate . There is evidence that the sensitivity of thymocytes to TCR stimulation is increased during maturation through the subcellular localisation of signalling molecules such as tyrosine kinase Lck [25]; the inhibition of extracellular signal-regulated kinase ( ERK ) activation and increased expression of inhibitory tyrosine phosphatase SHP-1 [26]; the upregulation of the negative regulator CD5 [27]; and the increased expression of ZAP-70 , a downstream target of TCR signalling [28] . However , the expression of miR-181a , a microRNA that enhances sensitivity to TCR stimulation , is reduced during thymic development [29] , [30] , and TCR signalling in response to low-affinity pMHC ligands is strongest in immature thymocytes [31] . The net effect of changes in TCR signal activating and inhibiting factors is not clear , but it is possible that stimulation with the same ligand will lead to different levels of activation in the same thymocyte at different stages of development . The challenge of synthesising these observations and describing quantitatively how the affinity , number and timing of pMHC contacts shape the developing T cell repertoire invites a mathematical modelling approach . Models of thymic selection have been successful in providing insight into the relationship between diversity of self peptides sampled in the thymus and the cross-reactivity [32] , [33] , alloreactivity [34] , size [35] , [36] and CD4SP/CD8SP ratio [37] of the selected repertoire . Models have also helped us understand the relation of HLA phenotype to viral epitope recognition [38] and the trade-off between MHC and T cell receptor diversity [39] . In this study we use stochastic ( probabilistic ) models to describe previously published in vivo data describing T and T commitment of a transgenic cell population in the presence of varying densities of agonist peptide in the thymus [16] . These data allow us to test and discriminate between models of how developing thymocytes might integrate signals received from pMHC ligands to make lineage decisions . Models of thymic selection must relate the physical interaction between a TCR and a pMHC ligand to the signal interpreted or integrated by the thymocyte . There is evidence to support competing models of TCR-pMHC interactions in which the level of T cell activation is determined by TCR-pMHC dwell times , through the kinetic proofreading model [40] , [41] , TCR occupancy [42]–[44] and overall pMHC ligand affinity [45] . Previous approaches to quantitative modelling of TCR-pMHC interactions can be divided into three broad categories: ( i ) detailed modelling of signal transduction immediately downstream of TCR-pMHC engagement [46] , [47]; ( ii ) kinetic models of binding events using measured rates of TCR-pMHC association and disassociation [48] , [49]; and ( iii ) the use of a ‘string model’ framework in which the strength of an interaction is determined by pairwise interaction energies between peptides and the aligned residues of amino acids on the variable CDR3 loop of randomly generated TCRs [32] , [33] . However , binding kinetic parameters are not available for the full range of endogenous peptides that are encountered during thymic development , and uncertainty remains in the relation between avidity and the signalling thresholds determining fate decisions . Here , we abstract from the mechanistic model of signal strength derived from molecular interactions . Instead we assume a distribution of signal strengths that a given TCR derives from pMHC ligands , in which low-strength signalling events occur with the greatest likelihood and , in line with our knowledge of the specificity of T cell recognition , stronger signalling events occur with decreasing probability . We show that our conclusions are insensitive to the precise form of this distribution . We explore candidate mechanisms of T and T selection using the canonical hypothesis that signals associated with T commitment are stronger than those required for T commitment but are below the threshold for negative selection . We use the data of van Santen et al . [16] to reject a simple model of the selection process in which thymocyte fate is based on testing sequential single TCR-pMHC interactions . Instead , we find the data can be explained with two generalisations of this model in which perceived TCR signal strength correlates to a strict hierarchy of cell fates ( neglectTT negative selection ) . In both models , thymocytes are continuously initiating fate decisions based on measuring the strength of binding to self peptide-MHC ligands in a series of encounters with antigen-presenting cells . In one class of model , the model , cells measure the avidity of each encounter , each of which comprises binding to a sample of multiple self-peptide-MHC ligands simultaneously . The model is motivated by studies implicating the integration of signals from multiple pMHC interactions in the priming of mature T cells by antigen [50]–[53] . In the second class of model , the two-phase model , we examine the consequences of TCR sensitivity of thymocytes varying during development . In this model a cell's interpretation of the signal derived from a given ligand depends on whether it occurs early or late in selection . We show that both models are able to describe the data , and also make predictions that are consistent with studies quantifying the efficiency of T selection with avidity [14] . However , we argue that variable TCR sensitivity is required to explain the effect of partial and full agonist peptide expression in the thymus on T generation and negative selection reported by Cozzo Picca and colleagues [15] .
We use data from van Santen et al . ( 2004 ) [16] in which the frequency of high affinity intra-thymic ligands was manipulated in vivo . Briefly , a mouse line was used that employs the tetracycline inducible system to conditionally express an invariant chain mutant , bearing the T cell epitope from moth cytochrome c ( MCC ) in place of the class-II associated invariant chain peptide ( CLIP ) -encoding region ( TIM ) . TIM was expressed in both cortical and medullary thymic epithelial cells ( cTEC and mTEC ) , and at controllable and graded levels . The mice also contained a transgene encoding a TCR specific for this peptide , such that in the absence of induced TIM , these cells differentiated efficiently into mature CD4 single positive thymocytes . Expression of TIM , measured by TIM RNA transcripts via real-time PCR , influenced both T and T formation in a non-linear fashion ( Figure 1 ) . In Figure 1 we see that ( i ) low frequencies of a strong agonist ( TIM ) do not affect the selection of TCR-specific ( AND ) thymocytes into the conventional T cell pool; ( ii ) moderate increases in agonist expression lead to increasing efficiency of selection of AND cells into T ( ( T ) against ( Relative TIM RNA ) between ; Pearson correlation ) and a concurrent drop in the efficiency of T selection; and ( iii ) high frequencies of a strong agonist lead to the deletion of AND T cells . A very similar trend was observed by Cozzo Picca et al . [17] using TCR transgenic cells specific for an epitope of influenza virus in the presence of different levels of expression of this agonist . Atibalentja et al . [18] also observed this trend following intravenous injection of varying concentrations of hen egg-white lysozyme ( HEL ) , which was rapidly processed and presented in the thymus , resulting in the negative selection of specific TCR transgenic T and an increase in TCR transgenic T at low , but loss at higher , HEL concentrations . Developing thymocytes survey pMHC ligands presented on the surface of thymic epithelial cells . In all models we assume that fate decisions are continually reassessed based on ‘encounters’ , each of which is the sum of interactions with pMHC ( Figure 2 ) , where . Each thymocyte participates in encounters at most , where a thymocyte might undergo negative selection , or initiate development into the T or T lineages , before reaching its -th encounter . We assume that each encounter with one or more pMHC can be divided into four categories determined by its affinity or avidity and the resulting signal through the T cell receptor ( s ) . These are ( i ) a weak or null signal below that required for positive selection; ( ii ) a signal sufficient for selection into the T lineage; ( iii ) a signal that initiates selection into the T lineage; and ( iv ) a strong signal that leads to deletion . We considered two classes of models . In one , the distribution of signal strengths resulting from encounters is constant throughout the selection period - the model . In the other , the two-phase model , we allow for the possibility that this distribution shifts during selection as a result of temporal changes in TCR sensitivity . The key parameters of interest were the encounter size for the model , and the number of encounters in each phase in the two-phase model . Other parameters were estimated simultaneously , but several quantities were taken as inputs to the models because the data from [16] did not allow us to parameterise them directly . These were ( i ) the parameters specifying the relation between relative RNA expression and absolute peptide abundance; ( ii ) the distribution of signal strengths obtained by the AND TCR from randomly sampled self pMHC ligands; and ( iii ) the relation between selection probabilities and absolute cell numbers . First , we explored ranges of parameters defining the mapping function ( equation 6 ) ; . We chose to use this generic sigmoid dose-response curve given our ignorance of the mechanistic relation between RNA expression and peptide-MHC abundance on thymic epithelial cells . However , we were able to partially validate this choice of function , and the region of parameter space that we explored , using data from the study by Obst et al . [58] . They characterised the relation between the degree of activation of adoptively transferred AND T cells and the relative TIM RNA expression on MHC class II-expressing cells , using a similar tetracycline-inducible expression system to that used in [16] . Their readout of immune activation was the fraction of AND cells that had divided 60 h following induction of TIM expression . Assuming this fraction is linearly related to peptide availability we used the data from Obst et al . to estimate the parameters of the mapping function ( equation 6 ) . We found that both the recruited fraction and an alternative measure of immune activation , the estimated per capita rate of recruitment into division , yielded mappings within the envelope of functions generated with our parameter ranges . These mappings also lay well within the 95% uncertainty envelope generated by the best-fitting parameters from our analysis of the data from [16] . For details , see Text S3 , Figure S1 and Table S1 . Second , we assumed the logarithm of the signal strength derived from a single AND-TCR endogenous-pMHC interaction is normally distributed with zero mean and unit variance . The scale of the distribution of signal strengths is arbitrary and its coefficient of variation does not influence our conclusions ( see Results ) . Third , the models provide the probabilities of selection into the T and T lineages and the data are absolute numbers of these populations in the thymus . We relate the numbers to probabilities through a scaling constant derived from the proportion of AND TCR cells that fail negative selection in control mice ( Text S1 ) .
The key features of the data are ( i ) T numbers decline monotonically with agonist expression and ( ii ) modest increases in agonist expression lead to an increase in the absolute number of AND T , with numbers then decreasing at higher levels of TIM expression ( Figure 1 ) . Assuming there is a positive relationship between TIM peptide presentation ( ) and relative TIM RNA expression , equation 4 shows that a model in which fate decisions are re-evaluated after single TCR-pMHC contacts ( ) can describe the T data , which falls progressively with . However , we can see using a graphical argument ( Figure 6 , upper panel ) that the model with constant TCR sensitivity will only be able to capture the trend in T numbers if encounters comprise TCR signals integrated over multiple pMHC bindings ( ) . If the strength of an interaction between a single AND-TCR and agonist TIM ( ) lies within the T-selecting range , we would expect to see a monotonic increase in T numbers with increasing agonist peptide expression; as agonist becomes more abundant , progressively more probability mass is contained within this area , boosting the probability of T selection ( Figure 5 , upper panel; Figure 7A , dotted-blue curve ) . Here , the one-hit model predicts that the absolute increase in T numbers is greater than or equal to the absolute decline in T numbers . Conversely , if is above the threshold for negative selection , , then we would predict a continuous decrease in T as agonist peptide becomes more abundant and increases the probability of deletion ( Figure 5 , upper panel; Figure 7A , dashed red curve ) . Neither of these trends are what is observed and so we rule out these scenarios . Finally , we can exclude the possibility that lies within the T -selecting range; if , increasing TIM expression would then increase the probability of selection into T , which we do not observe . Thus we can reject the simple one-hit model for selection of AND thymocytes . Extending the argument above , to explain the rise and fall of T numbers with agonist peptide expression ( ) requires the probability mass within the T-selecting region to increase then decrease with . This becomes possible when thymocytes read multiple TCR-pMHC bindings simultaneously ( ) . Qualitatively , this is because when , replacing an increasing fraction of endogenous peptides with TIM ( ) right-shifts the distribution of encounter strengths and , in contrast to the case , increases the probability of an encounter within both the T and negative-selection regions ( Figure 3B ) . The probability contained below the T selection threshold falls with , consistent with T numbers falling; the probability of an encounter occurring within the T zone first increases then decreases with , as required to explain the data; and the probability of negative selection continually increases ( Figure 6 , middle panel ) . We explored this quantitatively and sought to identify the parameters of the model from the data . They cannot all be identified uniquely . As described in Methods we took the approach of exploring a range of plausible parameters governing the function mapping RNA expression to endogenous peptide replacement by TIM , and a range of values of the maximum number of encounters , . Remarkably , all values of yielded equivalent descriptions of the data , and the encounter size was highly insensitive to other parameters; it lay between 2 and 5 for all models , with best fitting value , independent of . We also found that a range of mapping functions were able to describe the data equally well ( Table S1 ) . In particular , we predict that at maximum RNA expression , TIM replaces beween 0 . 1% and 12% of endogenous peptides . Representative fits to the data are shown in Figure 7 . Panel A illustrates the failure of the one-hit model , with the best fit obtained by forcing . Panel B shows the fit achieved with the model with a free parameter . The estimate of is also independent of the variance of the TCR-pMHC signal strength distribution . This also derives from the fact that the key quantities are just the probabilities of interactions lying between the different thresholds . However these thresholds become increasingly spaced with ( that is , as the log-normal distribution becomes increasingly fat-tailed ) . The less heavy-tailed the distribution of signal strengths , the smaller is the window of affinity/avidity for triggering T development with respect to the mean signal strength . Small increases in affinity can shift TCR signals from positively to negatively selecting [2] , [3] , and so if signal strength relates linearly to affinity or avidity [14] , our model predicts that the distribution of encounter strengths with self may not be strongly heavy-tailed . Anything between ten and a few hundred pMHC have been shown to be required for T cell activation ( see for example , [60] ) and as few as 3–5 for pMHC recognition by cytotoxic T cell effector function [61] , although with the extent of TCR binding influencing the degree of activation [42] . However , data interpreted using the kinetic proofreading model suggest that multiple interactions with very weak ligands may not lead to activation at the whole cell level ( see , for example , [40] , [41] , [62] ) . Therefore we wanted to test whether the low estimates of are an artefact of the assumption that every TCR-pMHC interaction generates a signal and so an encounter comprising weak TCR-pMHC bindings might still lead to strong signalling . To do this , we extended the model such that only a fraction ( ) of self-peptides are capable of inducing a signal through the AND TCR , and the remaining fraction are classifed as null . This introduces a stochastic element to the number of TCR contributing to the signal from each encounter . We found that increasing the abundance of null ligands increases the estimated TCR engagements per encounter ( Table S1 ) . For example , we estimate the number of proximal TCR-pMHC engagements per encounter ( ) to be between 20–190 if 99% of peptides fail to trigger the TCR , and between 350–1000 when 99 . 9% of peptides are null . Intuitively , the increase in derives from the dilution of the information content of each encounter by the presence of null peptides . For each encounter to be a unit of sufficient information with which fate decisions can be triggered , the sample size must increase in the presence of null interactions . As for the simpler model ( ) the estimate of is also independent of the number of encounters , . Therefore , this extended model predicts that in the AND TCR system the expected number of productive TCR-peptide MHC interaction per encounter remains remarkably small ( of the order 1 ) . This is perhaps unsurprising , as low values of will allow thymocytes to discriminate between ligands with small differences in affinity . Next we explored the implications of a time-varying sensitivity of thymocytes to TCR stimulation during maturation . The two phase model , as described in Methods , extends the one-hit model to include time-varying TCR sensitivity . Its predictions are independent of the direction of variation , but to illustrate we assume an interaction with agonist leads to T commitment during phase A early in development , but causes deletion in phase B when the same peptide is capable of inducing a stronger downstream TCR signal ( Figure 4 ) . Selection into the T lineage is still possible in both phases; what changes between phase A and phase B is a right-shift in the distribution of signal strengths with respect to the selection thresholds . This shift in probabilities within the different fate-determining affinity ranges yields the required trends in T and T production with TIM expression ( Figure 6 , lower panel ) . The details of parameter estimation for this model are in Methods and in Text S2 . The unknowns are , the number of encounters in phase A ( Since , this is the maximum number of pMHC sampled in phase A ) , , the number of encounters in phase B , and , the probability that a randomly sampled pMHC in the low-sensitivity phase A will lead to negative selection . As for the model , a range mapping functions described the data equally well . A representative fit using the two-phase model is shown in Figure 7C . We found a clear inverse relationship between the value of and the total number of encounters , + ( Figure 8A ) . The model predicts that between 1–2% of encounters occur in the lower sensitivity phase ( Figure 8B ) . We have used data in which there is a profound loss of conventional T cells in the presence of relatively low frequency of agonist peptide , while T numbers are maintained and even initially increase with moderate increases in agonist frequency ( Figure 1 ) . These observations suggested the hypothesis that regulatory T cells are intrinsically more resilient to deletion by agonist peptides than conventional T cells [16] . Further , the study by Cozzo Picca et al . [15] showed that a partial agonist can induce deletion of conventional T cells but only an agonist could boost regulatory T cell generation . This led to a hypothesis that agonist peptide may deliver a qualitatively different signal that induces regulatory T cells . We argue that neither of these hypotheses need be invoked . We have shown that both models can explain the first set of observations within a single affinity/avidity framework with different thresholds , without the need to assume differential susceptibilities of T and T to deletion . Further , we can see immediately that the model will not explain the partial/full agonist observations in ref . [15] . Their observation that partial agonist increases the probability of deletion with no increase in T suggests that the presence of the partial agonist shifts the distribution of the sum of interactions far to the right of the wild-type distribution , such that the bulk of the distribution is contained above the negative selection threshold . It follows that strong agonist must push this distribution even further rightwards , and so the probability of signals lying within the T-inducing zone must fall . This is inconsistent with the observed increase in T with agonist strength . In contrast , the simple two-phase model can explain the effect ( Figure 9 ) . Assume that the partial agonist is not strong enough to induce T commitment in phase A when the TCR is relatively insensitive , but in the more sensitive phase B delivers a signal that lies above the negative selection threshold . Then the net effect of introducing a weak agonist is to increase deletion and have little effect on T numbers , as is observed . In contrast , suppose the strong agonist triggers T commitment when the TCR in phase A , but is negatively selecting in phase B ( Figure 9 , right hand columns ) . Then ( i ) expression of the strong agonist will always lead to a fall in conventional T cell numbers , and ( ii ) moderate levels of strong agonist , while increasing the overall probability of negative selection , can drive a net increase in T production by boosting the probability of receiving a T-inducing signal in phase A . Figure 9 illustrates this effect for the case , but the same argument holds for a generalisation of the two-phase model with ( Figure S2 ) .
We focused here on a system in which agonist availability could be manipulated . We can also use the models to make predictions regarding selection efficiency as a function of TCR affinity . Lee et al . [14] quantified the efficiency of T selection in the rat insulin promoter ( RIP ) -mOVA system for a range of TCR clones with varying affinity for OVA . They found T were generated across a broad range of reactivities and found an increase in T selection efficiency with increasing affinity for a self peptide . Both the and variable-sensitivity models are able to explain these observations ( Text S4 and Figure S3 ) , but they make distinct predictions . The model predicts that the probability of the summed-proximal signals at each encounter exceeding the T threshold is lower for weaker clones; but this probability remains constant throughout development . A variable-sensitivity model predicts that T selection efficiency is determined by the duration of the window in which agonist contact triggers T commitment . In an increasing-sensitivity scenario , clones that recognise OVA more weakly will take longer to reach a level of TCR sensitivity that can induce T , and so a prediction of the model is that lower-affinity TCR clones will commit to the T lineage later in development . These models might be distinguished by manipulating thymocytes' TCR sensitivity , for example through altered expression of signalling proteins , at different stages of thymocyte development . The model predicts that the efficiency of T selection would be altered equally for all clones , whereas a model of increasing TCR sensitivity predicts that damping of TCR signalling later in development would most strongly reduce T selection efficiency for clones with low affinity for self peptide . We deliberately did not use model selection criteria to discriminate between models , in part because it is not possible to identify all parameters uniquely . Instead we identified regions of parameter space for each model that provided reasonable descriptions of the data . Importantly , the predicted values of the encounter size for different abundances of null peptides ( ) are insensitive to the remaining parameters ( Table S1 ) . We consider the and two-phase models capable of describing the data equally well because they are able to capture the decline in T and increase then decrease in T with TIM abundance . Both models capture this behaviour provided TIM abundance increases monotonically with RNA expression , which we expect to be the case . Further , model selection criteria are not required to reject the simplest one-hit model , nor to compare the abilities of the and two-phase models to describe the partial agonist observations . One prediction of an avidity-based model is that thymocytes may be deleted if they interact simultaneously with several pMHC at moderate affinities . We believe this prediction is not necessarily unreasonable; such events may be an inevitable byproduct of a selection process that is inherently probabilistic and which can result in cells with identical TCR experiencing divergent fates . The two mechanisms we explore here are not mutually exclusive . We illustrated the two-phase model assuming that decisions are made based on single , rather than summed interactions with pMHC . However it can be generalised to an extended version of the model in which encounters are interpreted differently as TCR sensitivity increases . This model will be able to explain the observations a least as well as the or two phase models , at the cost of extra parameters . Also , we illustrated the impact of increasing TCR sensitivity with a simple model that divided development into two discrete phases , while increases in TCR sensitivity are likely to be continuous . Modelling smooth changes in TCR sensitivity will introduce additional parameters but we expect such a model to yield qualitatively similar results . Importantly , our analysis does not exclude the possibility that T are more resistant to deletion and/or that qualitatively different signals are involved in T and T commitment; we simply show that these mechanisms need not be invoked to explain the observations . There is evidence that positive selection requires multiple low-affinity contacts with self-peptides in the thymus [63]–[65] . In contrast , in our model , a single encounter above the threshold is sufficient for positive selection . We assumed this threshold is very low for the AND thymocytes , which are strongly selecting under normal conditions . These cells will presumably receive essentially continuous positively selecting signals . In the general case , and particularly for weakly signalling TCRs that are near the threshold for death by neglect , the threshold would need to be included in the parameter search and the model would be extended to include a memory of recent interactions; one possibility is a model in which levels of survival or fate-determining proteins are increased by TCR signalling but decay in its absence . There is substantial evidence that increasing the frequency of a given clonal ( single TCR specificity ) population reduces its efficiency of selection and in particular the probability of being directed into the T lineage [14] , [22] , [24] . Our models treat cells as independent entities and do not explicitly incorporate the possibility that competition between thymocytes of similar TCR specificities might influence the availability of selecting ligands . However one mechanism of competition can be represented quite straightforwardly in the models . If the strength of an encounter correlates with its duration , or perhaps increases the probability of internalisation of the pMHC ligand by the thymocyte , the TCR-specific cells will compete for and possibly sequester agonist and other high-avidity pMHC ligands . This will shift the apparent distribution of signal strengths leftwards towards lower avidity ( and more available ) interactions , reducing the probability of acquiring T-selecting signals . This model of competition for higher-avidity pMHC ligands may also explain the observation that the efficiency of T selection can increase with precursor frequency [22] . However it remains an open question whether competition for pMHC plays an important role in selection at physiological precursor frequencies . Signalling through the IL-2 receptor is a requirement for T development [66]–[68] . It is thought that strong TCR signalling below the negative selection threshold may sensitise cells to IL-2 , licensing progression towards the T lineage . Whatever the precise role for IL-2 , it must operate downstream of fate-determining signals if selection is governed by a hierarchy of TCR avidity thresholds . Nevertheless , if IL-2 is limiting it may provide an upper bound on the total rate of production of T , either by redirection of cells to the T lineage or through loss . We argue however that competition for non-specific factors is unlikely to play a significant role in the system we are working with . First , the source of the IL-2 is unclear but we can reasonably suppose that IL-2 production in this system is independent of TIM expression . T numbers increase with TIM at low expression levels , indicating that IL-2 cannot be limiting in this region ( Figure 1 ) . Similarly it cannot be limiting at higher TIM levels as T decrease . It remains possible that a capping of T production through competition for IL-2 may be occuring in a small flat region near the peak in T numbers , but competition for non-specific factors alone cannot explain the key aspects of T development we are attempting to describe . Early neonatal thymectomy experiments suggested that the development of T is delayed compared to conventional T cells [69]–[71] . A key T marker , the transcription factor Foxp3 , is predominantly observed in the mature CD4 single positive stage of thymocyte development [72] . However , there may be a lag between initiation of T development and the stable expression of Foxp3; and it is possible that factors required for T development such as IL-2 [66]–[68] or medullary thymic epithelial cells [73] may only be required later in the maturation process . Thus the timing of T commitment remains unclear . The two models explored here make different predictions regarding this timing . The model suggests that diversion into the regulatory T cell lineage occurs with constant probability per encounter throughout selection . In contrast , the key prediction of the two-phase model is that the development of AND TCR T is triggered most efficiently within a relative short window during which thymocytes are relatively insensitive to TCR signalling . This window is estimated to span roughly 2% of the total pMHC ligands sampled , with the caveat that the two-phase model is an abstraction of what is more likely to be a temporally graded shift in sensitivity . The two-phase model's predictions are identical whether the shorter , less sensitive phase occurs early or late in development . However , expression of the downstream TCR-signalling protein Zap70 increases progressively between the double positive ( DP ) and single positive stages of thymocyte development , and is associated with increasing sensitivity to TCR stimulation [28]; immature DP thymocytes display lower surface expression of TCRs , as compared to mature single positive ( or ) thymocytes , and TCR signalling may be actively inhibited in immature DPs [74] . Thymocytes' signalling environment may also change as they develop . Selection begins in the thymic cortex , where pMHC are encountered on cortical thymic epithelial cells , before cells migrate to the medulla where they encounter pMHC on medullary thymic epithelial cells and dendritic cells . It is thought that levels of co-stimulation and antigen presentation are generally lower in the cortex than in the medulla [75]–[77] , suggesting that there may be an effective increase in TCR signalling during development . Using these observations , the two-phase model predicts that T development begins predominantly , but not exclusively , during a short window at the earliest double positive stage of selection . Clearly , definitive experimental identification of when T development is initiated will substantially increase our understanding of how thymocytes process information . Reliable recognition and discrimination of self and nonself ligands requires both TCR sensitivity and specificity . Specificity will decrease as the number of pMHC integrated per encounter ( ) increases — when is large , many pMHC engagements are integrated at each encounter , and so the thymocyte is then just sampling the mean of the distribution of pMHC affinities , and information is lost . This may explain why the optimum values of in the model are at the lower end of the reported numbers of pMHC engagements required for T cell activation; T cell activation may invole a relatively small number of informative TCR recognition events , together with many brief engagements with null or very low affinity peptide ligands . Our analysis of the model places small lower bounds on the number of non-null TCR engagements per encounter; but the two-phase model explains the data by allowing even a single non-null pMHC recognition event to influence fate . We speculate that varying TCR sensitivity with time in the thymus may allow for increased specificity in self-nonself discrimination . | T cells develop in the thymus , where they are vetted – they must respond weakly to self-antigens , but not so strongly as to risk causing autoimmunity . This selection process involves developing T cells being exposed to a large sample of self-peptides presented on specialised cells in the thymus , and deciding to die or to differentiate into mature T cells of either conventional or regulatory lineages . The rules by which T cells assimilate information from these interactions to make these decisions are not known . In this study we use previously published data to assess and discriminate between different models of thymic selection and find the most support for a model in which T cells vary their sensitivity to self-peptides during their development . This allows fate decisions to be made on the basis of as few as one peptide at a time , which allows for fine specificity in the selection process . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"immunology",
"biology",
"computational",
"biology"
] | 2013 | Models of Self-Peptide Sampling by Developing T Cells Identify Candidate Mechanisms of Thymic Selection |
In early 2015 , a ZIKA Virus ( ZIKV ) infection outbreak was recognized in northeast Brazil , where concerns over its possible links with infant microcephaly have been discussed . Providing a causal link between ZIKV infection and birth defects is still a challenge . MicroRNAs ( miRNAs ) are small noncoding RNAs ( sncRNAs ) that regulate post-transcriptional gene expression by translational repression , and play important roles in viral pathogenesis and brain development . The potential for flavivirus-mediated miRNA signalling dysfunction in brain-tissue development provides a compelling hypothesis to test the perceived link between ZIKV and microcephaly . Here , we applied in silico analyses to provide novel insights to understand how Congenital ZIKA Syndrome symptoms may be related to an imbalance in miRNAs function . Moreover , following World Health Organization ( WHO ) recommendations , we have assembled a database to help target investigations of the possible relationship between ZIKV symptoms and miRNA-mediated human gene expression . We have computationally predicted both miRNAs encoded by ZIKV able to target genes in the human genome and cellular ( human ) miRNAs capable of interacting with ZIKV genomes . Our results represent a step forward in the ZIKV studies , providing new insights to support research in this field and identify potential targets for therapy .
Zika virus ( ZIKV ) is an emerging mosquito-borne flavivirus , first isolated in 1947 from the serum of a pyrexial rhesus monkey caged in the Zika Forest ( Uganda/Africa ) [1] . In 2007 , ZIKV was reported linked to an outbreak of relatively mild disease , characterized by rash , arthralgia , and conjunctivitis on Yap Island , in the western Pacific Ocean [2] . In 2015 , ZIKV circulated in the Americas , probably introduced through Easter Island ( Chile ) by French Polynesians [3] , where concerns over its links with infant microcephaly have been raised . MicroRNAs ( miRNAs ) are small noncoding RNAs ( sncRNAs ) that regulate post-transcriptional gene expression by translational repression . It is estimated that more than 60% of human protein-coding genes are likely to be under the control of miRNAs [4] . Two hypotheses exist as to how miRNAs could influence ZIKV/human-host interaction . First , the virus could transcribe miRNAs that provide benefits associated with cellular and viral gene expression ( e . g . Herpesvirus , Polyomavirus , Ascovirus , Baculovirus , Iridovirus , Adenovirus families ) [5 , 6] . RNA retrovirus miRNAs are transcribed through RNA polymerase III ( pol III ) , instead of pol II . Virus-encoded miRNAs support persistent infections through subtle modulation of gene expression , leading to prevention of host cell death , evasion of the host immune system and regulation of the latent-lytic switch [7] . Second , retrovirus genomes may directly interact with cellular miRNAs to enhance viral replication potential [5] . By recruiting/exploiting cellular miRNAs , an RNA virus can disturb the regulation of host gene expression , which can trigger molecular disease . In order to provide a theoretical background for future experimental verification of these hypotheses , the ZIKV collaborative database ( ZIKV-CDB ) was assembled . This enables , ( i ) searching for predicted ZIKV miRNAs mimicking human miRNAs [searching criteria includes: “Gene name” , “Gene Symbol” or “Ensembl ID”] ( hypothesis 1 ) ; and ( ii ) searching for human miRNAs with possible binding-sites to the ZIKV genomes ( hypothesis 2 ) .
The ZIKA Virus Collaborative Database ( ZIKV-CDB ) was constructed based on two strategies . The first one consists in identifying ZIKA virus ( ZIKV ) microRNA ( miRNA ) molecules that may affect human gene expression . The second strategy consists in identifying human microRNA molecules that may be recruited by the ZIKV genome . Our search included the full set of cDNA sequences of the human genome available on the Ensemble database ( release 83 ) [29] for targets of the predicted ZIKV mature miRNA molecules , using the software miRanda [11] . The mature miRNA sequences were predicted using a pipeline based on three steps . The first step uses the tool RNAfold [9] to compute the minimum free energy and to predict the secondary structures based on Zika virus genome cDNA sequences ( see ZIKV genomes accession number section ) . The second step uses the predicted secondary structures to identify the hairpins formed by miRNA precursors using the HHMMiR workflow [8] . The third step uses the software PHDcleav [10] to identify cleavage sites of the Dicer human enzyme in the hairpin structures to generate the sequences of the mature miRNA . A fasta file containing all predicted precursor and mature sequences of the nine ZIKV-encoded miRNAs is provided in the S1 and S2 Datasets , respectively . The miRNA molecules suppress post-transcriptional gene expression through physical interaction with the messenger RNA ( mRNA ) [30] . To detect miRNA target gene candidates , we used the approach presented by the software miRanda ( 11 ) , which employs the local alignment of the miRNA and mRNA molecules combined with the information of minimum free energy of each nucleotide match of RNA-RNA duplexes . The free energy ( ΔG ) of optimal strand-strand interaction for each match of alignment was determined using the Vienna package [8] . A detailed table containing the miRNA identifiers from ZIKV , target Ensembl transcripts , total score , total energy , maximum score per alignment , maximum energy per alignment , strand , length of the miRNA , length of the target , and the alignment positions is provided as S3 Dataset . Similarly , a detailed table containing the miRNA identifiers from human , target region in the ZIKV genomes , total score , total energy , maximum score per alignment , maximum energy per alignment , strand , length of the miRNA , length of the target , and the alignment positions is provided as S4 Dataset . Kedougou Virus ( NC_012533 and AY632540 ) Bagaza Virus ( NC_012534 and AY632545 ) ; ZIKA Virus isolated from Uganda ( LC002520 , NC_012532 , AY632535 ) , from Central Africa Republic ( KF268949 , KF268948 and KF268950 ) , from Brazil ( KU527068 , KU321639 , KU365778 , KU365779 , KU365777 and KU365780 ) , Martinique ( KU647676 ) , French Polynesia ( KJ776791 ) , Haiti ( KU509998 ) and Puerto Rico ( KU501215 ) . All recovered ZIKV genome sequences were aligned using the software ClustalW7 . Further , the phylogenetic tree was constructed using the online tool Itol: Interactive Tree of Life [12] , applying the Neighbor-joining method , with 100 bootstrap repetitions .
We introduce the ZIKV-CDB , a collaborative database encompassing both , predicted miRNAs encoded by ZIKV genomes that could potentially target the human genome , and cellular ( human ) miRNAs with sequence complementary to ZIKV genomes . This knowledgebase should facilitate researchers when exploring targets that may affect the expression of genes associated with microcephaly and other neurodevelopmental syndromes caused by ZIKV infection . The chosen method for predict ZIKV-encoded miRNAs was based on a previously published benchmark [13] , which shows that among the evaluated tools , miRanda had the highest sensitivity for predicting miRNAs , providing more targets for validation . To increase the effectiveness of this strategy , further analysis using genome sequences of other viruses with experimentally validated virally encoded miRNAs should be explored as positive controls . In contrast to previous reports [14 , 15] , the miRNAs identified here are located in the ZIKV polyprotein coding region . Recently , a study using a similar approach identified miRNAs located in the CDS region of Ebola Virus [16] . Examples of genes predicted to be targeted by miRNAs and previously validated as having a potential link to neurological disorders include the peroxisomal biogenesis factor 26 gene ( PEX26 ) , the fibroblast growth factor 2 ( FGF2 ) , the SET binding factor 1 ( SBF1 ) , the hook microtubule-tethering protein 3 ( Hook3 ) , the pleckstrin homology domain , and the RhoGEF domain containing G4 ( PLEKHG4 ) ( Table 1 ) . All these targets , when aligned to predicted miRNAs , met the minimum criteria of free energy ( minimum energy = 1 ) and score ( 140 ) . PLEKHG4 polymorphisms have been related to spinocerebellar ataxia [17] , a progressive-degenerative genetic disease . Also , Hook3 has been reported to interact with Pericentriolar Material 1 ( PCM1 ) during brain development , and an imbalance in the Hook3-PCM1 interaction can cause premature depletion of the neural progenitor pool in the developing neocortex [18] . Finally , defects in the PEX26 gene can lead to a failure of protein import into the peroxisomal membrane or matrix , being the cause of several neuronal disorders , including Zellweger syndrome ( ZWS ) , and neonatal adrenoleukodystrophy ( NALD ) [19 , 20] . It is important to highlight that none of the predicted miRNAs were associated with every analysed ZIKV genome , nor in all isolates from the recent outbreak in Brazil . Which suggests that these predicted miRNAs are not essential for virus replication , but may improve their replication success [5] . These differences between genomes also may be related to different phenotypes of ZIKV infection , such as microcephaly in infants , Guillain—Barré syndrome [21] , other symptoms similar to those of dengue and chikungunya , or asymptomatic phenotype [22] . Interestingly , several human miRNAs known to exert an influence on the expression of genes with a known functional role in neuronal development were found to have sequence complementarity to regions in the ZIKV genome ( Table 2 ) . One of the human hsa-miR-34a miRNA targets , the Cyclin-Dependent Kinase 6 ( CDK6 ) gene , for instance , was computationally predicted to interact with several ZIKV genomes . CDK6 is associated with the centrosome during mitosis , controlling the cell cycle division phases in neuron production [23] . Mutation in CDK6 can lead to a deficient centrosomes division , which in turns can cause autosomal recessive primary microcephaly ( MCHP ) [19] . There are seven well-know genes encoding centrosomal proteins that are involved in the autosomal recessive primary microcephaly ( MCPH ) [24] , including the CDK5 Regulatory Subunit Associated Protein 2 ( Cdk5rap2 or MCHP3 ) gene . We found a possible binding-site to the hsa-mir-324-3p , a cellular miRNA targeting Cdk5rap2 gene , in the ZIKV genomes . Equally , we found that ZIKV genomic regions can potentially bind the hsa-mir-615-3p and hsa-miR-193b-3p human miRNAs , which target the WD Repeat Domain 62 ( WDR62 or MCHP2 ) , also related to MCHP when mutated . Remarkably , a hsa-mir-21-5p miRNA complementary site was found in the genomes of ZIKV isolated from Brazil , Haiti , Martinique and French Polynesia , but not in those from Africa . This miRNA targets the MCHP4 gene , also linked to microcephaly cases . The geographic and hence historical accumulation of genomic differences may explain the recent rise of microcephaly , and this observation also corroborates the predicted pathway of transmission from Africa , through Oceania , and into Central and South America . To further support this , phylogenetic analysis was performed for all complete ZIKV genomes ( Fig 1 ) , which identified a cluster of strains isolated in the Americas and Oceania ( derived strains ) , and another with the African strains ( ancient strains ) . A third group , including two Flavivirus genotypes closely related to ZIKV , which were added as an out-group . These differences were also supported by phylogenetic analysis of only the predicted miRNAs encoded by each ZIKV strain ( S1 Fig ) . Nine predicted miRNA types were identified , with types 1–4 being shared exclusively by derived strains and 5–8 being exclusive to ancient strains . Type 9 was only found in the genomes of strains from the Central African Republic ( Fig 2 ) . The predicted miRNAs could target a total of 14 , 745 human genes; 9 , 106 are specific to miRNAs from ancient strains , 2 , 840 are specific to those from derived strains , and 2 , 789 are shared . Recently , ZIKV was isolated from the brain tissue of a fetus diagnosed with microcephaly [25] , and two laboratory studies have provided robust evidence that ZIKV infection may cause brain defects in infants by influencing brain cell development [26 , 27] . However , the mechanism by which ZIKV alters neurophysiological development remains unknown , inhibiting the development of therapeutic interventions . Our results suggest a putative influence of miRNAs on the expression of human-genes associated with the symptoms of Congenital ZIKA Syndrome . The ZIKV-CDB provides a useful knowledge base to support research targeted at mitigating the impacts of this emerging health problem [28] . ZIKV-CDB is an open-source and collaboration-based forum for sharing and identifying potential targets . The database can guide experimental investigation to elucidate the possible association between ZIKV infection and neurobiological development in infants . The ZIKV-CDB is going to be further expanded to encompass information related to others sncRNAs , as predicted by other approaches . The database will also be continuously maintained and curated by the Genomics and Computational Biology Group , FIOCRUZ/CPqRR ( http://www . cpqrr . fiocruz . br ) . | The potential interaction between ZIKA Virus ( ZIKV ) infection and infant brain defects still has no explanatory mechanism . The mechanism of action for several other viruses have been described , and are often related to the control or production of sncRNA ( small noncoding RNA ) molecules , which regulate the expression of several human genes . Based on this knowledge , we have predicted miRNAs ( a kind of sncRNA ) encoded by ZIKV genomes , and identified human miRNAs that may be able to interact with human genes related to microcephaly . To help support further investigation , we have created a searchable database for ZIKV miRNAs that may mimic human molecules and possibly interfere with the human gene expression; or to search for human sncRNA molecules that could be recruited by the ZIKV genome , and influence function . This database could aid investigators in elucidating the role played by sncRNAs in human ZIKV infection , and be used to identify novel therapeutic targets . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"and",
"Discussion"
] | [
"medicine",
"and",
"health",
"sciences",
"microcephaly",
"pathology",
"and",
"laboratory",
"medicine",
"gene",
"regulation",
"pathogens",
"microbiology",
"database",
"searching",
"human",
"genomics",
"viruses",
"developmental",
"biology",
"micrornas",
"rna",
"viruses",
"genome",
"analysis",
"morphogenesis",
"thermodynamics",
"research",
"and",
"analysis",
"methods",
"genomics",
"medical",
"microbiology",
"birth",
"defects",
"gene",
"expression",
"microbial",
"pathogens",
"biological",
"databases",
"congenital",
"disorders",
"free",
"energy",
"physics",
"biochemistry",
"rna",
"nucleic",
"acids",
"flaviviruses",
"viral",
"pathogens",
"database",
"and",
"informatics",
"methods",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"computational",
"biology",
"non-coding",
"rna",
"genomic",
"databases",
"organisms",
"zika",
"virus"
] | 2016 | ZIKV – CDB: A Collaborative Database to Guide Research Linking SncRNAs and ZIKA Virus Disease Symptoms |
In the infectious stage of Trypanosoma brucei , an important parasite of humans and livestock , the mitochondrial ( mt ) membrane potential ( Δψm ) is uniquely maintained by the ATP hydrolytic activity and subsequent proton pumping of the essential FoF1-ATPase . Intriguingly , this multiprotein complex contains several trypanosome-specific subunits of unknown function . Here , we demonstrate that one of the largest novel subunits , ATPaseTb2 , is membrane-bound and localizes with monomeric and multimeric assemblies of the FoF1-ATPase . Moreover , RNAi silencing of ATPaseTb2 quickly leads to a significant decrease of the Δψm that manifests as a decreased growth phenotype , indicating that the FoF1-ATPase is impaired . To further explore the function of this protein , we employed a trypanosoma strain that lacks mtDNA ( dyskinetoplastic , Dk ) and thus subunit a , an essential component of the proton pore in the membrane Fo-moiety . These Dk cells generate the Δψm by combining the hydrolytic activity of the matrix-facing F1-ATPase and the electrogenic exchange of ATP4- for ADP3- by the ATP/ADP carrier ( AAC ) . Surprisingly , in addition to the expected presence of F1-ATPase , the monomeric and multimeric FoF1-ATPase complexes were identified . In fact , the immunoprecipitation of a F1-ATPase subunit demonstrated that ATPaseTb2 was a component of these complexes . Furthermore , RNAi studies established that the membrane-bound ATPaseTb2 subunit is essential for maintaining normal growth and the Δψm of Dk cells . Thus , even in the absence of subunit a , a portion of the FoF1-ATPase is assembled in Dk cells .
Trypanosomes are unicellular flagellates from the order Kinetoplastida , which is comprised of some of the most devastating human pathogens in the world . For example , in sub-Saharan Africa , infection from Trypanosoma brucei gambiense and T . b . rhodesiense causes Human African Trypanosomiasis , which is almost always fatal if left untreated [1] . The latest WHO reports estimate that there are 10 , 000 new cases annually in endemic regions . Meanwhile , a third subspecies , T . b . brucei , infects livestock and therefore negatively affects the human population through malnutrition and economic hardships [2 , 3] . T . brucei parasites have a complex life cycle , alternating between the mammalian host and the insect vector , a tse-tse fly . During this environmental switch , the protist undergoes rapid and dramatic changes in cell morphology and metabolism [4–6] . In particular , the single mitochondrion undergoes extensive remodelling , which reflects the adaptability of the parasite to consume different carbon sources based on their availability [4] . The procyclic ( insect ) form ( PF ) of trypanosomes catabolizes amino acids and maintains a well-developed mitochondrion with abundant cristae , Krebs cycle enzymes and a complete oxidative phosphorylation pathway . This pathway includes enzymatic complexes that generate a mitochondrial ( mt ) membrane potential ( Δψm ) that is coupled to ATP synthesis by the FoF1-ATP synthase [5] . In contrast , the bloodstream form ( BF ) of this parasite populates the glucose-rich fluids ( e . g . blood and spinal fluid ) of its vertebrate host , allowing them to utilize just glycolysis for ATP production . This results in a drastically reduced mitochondrion that lacks significant cristae , key enzymes of the Krebs cycle and the cytochrome-containing respiratory complexes that pump protons into the inner mt membrane space [6 , 7] . Despite this reduction , the BF mitochondrion is still an active organelle , holding vital processes e . g . lipid metabolism [8] , ion homeostasis [9] , calcium signalling [10 , 11] , FeS cluster assembly [12] and acetate production for de novo lipid biosynthesis [13] . Importantly , in the absence of proton-pumping respiratory complexes III and IV , the indispensable Δψm is sustained mainly by the hydrolytic activity of the FoF1-ATPase . Thus , this complex possesses an essential , unique and irreplaceable function in BF mitochondria [14] . In other eukaryotes , this reverse activity of the FoF1-ATP synthase is observed only rarely , for very brief moments of time and under very specific conditions ( i . e . during oxygen deprivation or in response to damaged or mutated mt respiratory proteins ) . When the function of the respiratory complexes is compromised , the Δψm falls below a physiological threshold and is restored by the reverse proton pumping activity of the FoF1-ATPase , which is powered by ATP hydrolysis . The hydrolytic activity of the catalytic F1-ATPase is also essential for exceptional cells that lack mtDNA ( ρ° cells ) . These cells do not express several core subunits of the membrane embedded Fo-moiety ( subunits 6 , 8 and 9 in yeast , subunits a and A6L in bovine ) of the FoF1-ATPase , notably those that are components of the proton pore . Thus , the matrix protruding F1-ATPase energizes the inner mt membrane by coupling ATP hydrolysis with the exchange of ADP3- for ATP4- by the ATP/ADP carrier ( AAC ) [15] . The same mechanism for producing the Δψm is utilized by trypanosomes that lack a mt genome , which is called a kinetoplast [16] . These naturally occuring dyskinetoplastic forms ( Dk ) of T . brucei ( e . g . T . b . evansi or T . b . equiperdum ) cause economically significant diseases in horses , camels , and water buffaloes . Remarkably , these parasites are not able to switch to the insect stage and are transmitted mechanically by bloodsucking insects or by coitus [17] . The mtDNA-lacking trypanosomes can also be chemically induced in the laboratory ( e . g . Dk T . brucei EATRO164 ) [18] . Interestingly , each of the Dk cell lines characterized so far , bear one of several different compensatory mutations in the nuclear encoded subunit γ that enable the Δψm to be generated independently of the Fo-moiety [14 , 16 , 19] . In general , the FoF1-ATP synthase complex consists of two functionally distinct enzymatic segments: the hydrophilic F1 catalytic moiety and the membrane-bound Fo pore . Both of these subcomplexes are linked together by the central and peripheral stalks . The central stalk rotates with the c-ring when protons are allowed to pass through the Fo pore , located between the c-ring and subunit a . In contrast to the rotation of the central stalk , the stationary peripheral stalk plays a crucial role in keeping the catalytic F1 headpiece static , thus resisting the rotational torque . The eubacterial F1-moiety consists of the catalytic domain and the central stalk , which are comprised of five subunits in a stoichiometry of α3 , β3 , γ1 , δ1 , ε1 . The Fo-moiety is composed of the oligomeric c10–15 ring and a single subunit a joined together with two copies of subunit b , which extend from the membrane and form the base of the peripheral stalk . The composition of the eukaryotic enzyme has been determined mainly from detailed studies of FoF1-ATP synthase purified from the mitochondria of Bos taurus and Saccharomyces cerevisiae , members of the clade Opisthokonta . These enzymes contain homologous prokaryotic-like core components , but also incorporate additional eukaryotic specific subunits involved in the structure , oligomerization and regulation of the complex [20] . These novel subunits have been assigned to various regions of the FoF1-ATP synthase: i ) subunit ε and inhibitory factor 1 ( IF1 ) bind to the F1-moiety; ii ) the small hydrophobic subunit 8 ( subunit A6L in bovine ) is located in the membrane interacting with subunit a; iii ) the soluble subunit h ( subunit F6 in bovine ) , oligomycin sensitivity-conferring protein ( OSCP ) and hydrophilic subunit d are assigned to the peripheral stalk; iv ) subunits e , f , g , all containing one transmembrane helix , are regarded as accessory peripheral stalk proteins that predominantly reside in the inner membrane where they presumably stabilize the utmost hydrophobic subunit 6 ( or subunit a in bovine ) , thereby assisting the peripheral stalk with its stator function [21–23] . Recently , however , several studies have reported unique structural and functional features of the FoF1-ATP synthase from medically relevant parasites and other organisms that represent the clades of Chromalveolata [24] , Archaeplastida [25] and Excavata [26] . In Plasmodium falciparum , a member of the phylum Apicomplexa , the genomic data indicates the likely presence of all eukaryotic F1 subunits and some of the Fo and stator subunits [27] . Conspicuously absent are the subunits a and b , which are crucial to the function of all known FoF1-ATP synthases . These subunits are also missing in the ATP synthase of P . yoelii and Tetrahymena thermophila , the latter a member of the sister phylum Ciliophora , both of which have been shown experimentally to possess ATP synthase activity [28–30] . This suggest that ciliates and apicomplexan species employ highly divergent or novel subunits to fulfil the functions of the classical subunits a and b . Another example of an unusual FoF1-ATP synthase complex comes from studies of the chlorophycean algea Chlamydomonas reinhardtii and Polytomella sp . that determined the complex contains up to 9 unique subunits ( Asa1-Asa9 ) that either form an innovative peripheral stator or are responsible for complex dimerization [31] . Trypanosoma FoF1-ATP synthase consists of the well conserved F1-moiety comprised of subunits α , β , γ , δ , ε and the trypanosome-specific subunit p18 [26 , 32] , and the less characterized Fo pore and peripheral stalk where only subunits c , a and OSCP were identified at the gene or protein level [26 , 33] . Additionally , the complex contains up to 14 Kinetoplastida-specific subunits that lack homology to any of the previously described subunits . Therefore , their position within the complex and their function is unknown . Interestingly , these recently identified FoF1-ATP synthase subunits specific for T . brucei ( ATPaseTb ) may represent early evolutionary attempts to create the functional and structural components of the eukaryotic accessory subunits for this early divergent species . Previously , two of these novel subunits , ATPaseTb1 ( Tb10 . 70 . 7760 ) and ATPaseTb2 ( Tb927 . 5 . 2930 ) , were shown to be essential for the proper function and structural integrity of FoF1-ATP synthase in the procyclic form of this parasite [26] . Here , we have extended this study and functionally characterized ATPaseTb2 in the disease-relevant stage of T . brucei and Dk T . b evansi . Our results demonstrate that the membrane-bound ATPaseTb2 protein is indispensable for the survival of BF and Dk trypanosomes because it is crucial for maintaining the Δψm in these cells . Furthermore , the unexpected observation of assembled FoF1-ATPase complexes in two different strains of trypanosomes lacking mt DNA suggests that the composition of the Dk FoF1-ATPase complex is more similar to the BF FoF1-ATPase complex than previously thought .
ATPaseTb2 is annotated as a hypothetical protein in the TriTrypDB database ( www . tritrypdb . org ) and its mt localization was predicted by the presence of an N-terminal mt targeting sequence assigned by Mitoprot II v1 . 101 [34] ( probability 0 . 823 ) . Strikingly , this subunit has no detectable homologs outside Kinetoplastida based on a similarity sequence search ( e . g . BLAST ) . To uncover any homologous relationships for protein sequences that do not share high enough sequence identity , but might have similar secondary structure , we employed the HHpred toolkit ( http://toolkit . tuebingen . mpg . de ) , which is based on an HMM-HMM comparison that reveals homologous relationships even if the sequences share less than 20% sequence identity [35] . All of the available Kinetoplastida ATPaseTb2 homologs obtained from the TriTrypDB ( S1A Fig . ) were analyzed using HHpred software . Among the first five structural homologs detected was the bovine FoF1-ATP synthase subunit d , with a probability of 44 . 7 and a P-value of 0 . 0021 . Then a consensus sequence generated from an alignment of this 92 amino acids ( aa ) region from all identified kinetoplastida homologs was resubmitted for HHpred analysis . This search returned a significant hit to the bovine subunit d with a probability of 78 . 1 and a P-value of 0 . 0002 ( S1B Fig . ) . This bovine FoF1-ATP synthase subunit is comprised of 160 aa and resides within the peripheral stalk where it interacts with all the other three components of the peripheral stalk ( b , F6 and OSCP ) . This subunit is predominantly hydrophilic , contains several long alpha-helices and is essential for the enzymatic function [23 , 36 , 37] . The region of similarity to subunit d falls in the middle of the ATPaseTb2 sequence and it is followed by a putative trans-membrane ( TM ) domain ( aa 239–254 ) that was predicted by MEMSAT-SVM software ( http://bioinf . cs . ucl . ac . uk/psipred/ ) [38] with a -0 . 738 prediction score ( Fig . 1 ) . This prediction is intriguing because a typical eukaryotic subunit d is not directly attached to the membrane , but rather it interacts with the soluble region of the membrane-bound subunit b via three coiled-coil interactions [23] . Thus , ATPaseTb2 may represent a novel subunit of the membrane-bound peripheral stalk that possibly depicts an early divergent functional alternative of the stator by creating a chimera of subunits b and d . The mitochondrial localization , as well as the structural and functional association of ATPaseTb2 with the FoF1-ATP synthase in PF T . brucei cells was previously described [26] . Because the activity of this complex significantly differs between the insect and mammalian stage of the parasite , we were interested if the ATPaseTb2 protein is also expressed in trypanosomes residing in the bloodstream of their host . Therefore , we cultured PF427 ( insect stage ) , BF427 ( bloodstream forms ) and the laboratory-induced dyskinetoplastic Dk164 T . brucei cells , in addition to the naturally occurring Dk T . b . evansi . Whole cell lysates from PF427 , BF427 , Dk164 and T . b . evansi cells were fractionated by SDS-PAGE and analyzed by western blot using a specific polyclonal antibody against ATPaseTb2 ( Fig . 2A ) . Interestingly , ATPaseTb2 is expressed in all four cell types; however , its expression is significantly decreased in BF427 and even more reduced in Dk164 and T . b . evansi cells . A similar expression pattern for F1-ATPase subunit β and mt chaperone Hsp70 was also observed between PF , BF and Dk cells ( Fig . 2A ) . These results are in agreement with the proposed reduction of the mitochondria size , function and activity in the various bloodstream forms of T . brucei [39] . To experimentally investigate the subcellular localization of ATPaseTb2 within the bloodstream T . brucei cell , the ATPaseTb2 gene product was C-terminally tagged with a v5-tag and inducibly expressed using tetracycline ( tet ) . Cytosolic and mitochondrial fractions were isolated from a hypotonic lysate of cells over-expressing the ATPaseTb2_v5 protein . The mitochondrial fractions were then treated with digitonin to create mitoplasts that were further incubated with sodium carbonate ( Na2CO3 ) to obtain mt membrane and mt matrix fractions . The subsequent western analyses established the purity of the extracted fractions as the cytosolic enolase , the mt inner membrane carrier protein ( AAC ) and the matrix-localized guide RNA binding protein ( MRP1 ) were confined within their respective subcellular fractions . Notably , ATPaseTb2 was predominantly localized in the mt membrane fraction , while subunit β of the matrix-soluble F1-ATPase was released into the matrix fraction ( Fig . 2B ) , demonstrating that ATPaseTb2 is firmly embedded in the inner mt membrane . To confirm that ATPaseTb2 is a genuine subunit of both the BF and Dk FoF1-ATPase , we employed the same strategy applied previously for PF cells [26] . Therefore , we first created BF cell lines in which either the ATPaseTb2 or the F1 subunit p18 were inducibly expressed with a C-terminal tag . Complexes assembled with a tagged protein were then purified using a one-step IgG affinity purification , followed by treatment with TEV protease to release bound complexes from the IgG beads . The TEV eluates were further analyzed by western blot for the presence of the tagged proteins , known FoF1-ATPase subunits , and AAC ( Fig . 3A ) . Importantly , the presence of the tagged subunits ATPase Tb2 and p18 in the respective cell lines were confirmed using a specific antibody against the c-myc epitope that comprises part of the tag . In addition , the ATPaseTb2_TAP eluate contained FoF1-ATPase subunits β , p18 and ATPaseTb1 . Vice versa , the p18_TAP eluate contained subunits ATPaseTb2 , ATPaseTb1 and subunit β . While AAC has been described to associate with mammalian FoF1-ATPase under certain conditions [40] , it is not a core component of the enzyme and the lack of its signal indicates the stringency of this technique . The TAP tag purification from non-induced cells was used as a negative control to verify that the detected proteins do not bind non-specifically to the charged beads . To analyze the composition of the Dk FoF1-ATPase , p18 tagged complexes were purified and subjected to the same set of antibodies described for the BF complexes ( Fig . 3B ) . These results also depicted the same interactions seen in Fig . 3A . Therefore , the co-purification of ATPaseTb2 with known ATPase subunits p18 and β validates that it is an authentic constituent of the FoF1-ATPase complex in both BF and Dk cells . The sedimentation profile of FoF1-ATP synthase in 10–30% glycerol gradients ( GG ) was previously published and revealed two distinct regions representing the F1-moiety and the FoF1-monomer and multimeric complexes [26] . In order to specify if ATPaseTb2 is a component of the F1-moiety or the FoF1-complex , the glycerol gradient sedimentation profile was determined using mt lysates from all four cell types: PF427 , BF427 , Dk164 and T . b . evansi . In correlation with the literature , a new specific antibody raised against the F1-subunit β depicted both regions , which are defined as ∼10S ( fractions 6–10 ) , and ∼40S ( fractions 14–22 ) ( Fig . 3C ) . A different sedimentation pattern was revealed when a specific antibody recognizing ATPaseTb2 detected only the 40S region ( fractions 14–22 ) , representing the FoF1-monomer and -multimers . These findings suggest that the novel subunit is not a component of the F1-moiety , but rather a member of the fully assembled complex . Furthermore , compelling results hinting at the function of this hypothetical protein were also obtained when the mt lysates of Dk164 and T . b . evansi cells were fractionated on similar gradients . While the majority of the signal for subunit β was predictably detected in fractions 6–10 , representing the F1-ATPase complex , there was also a weak signal detected in higher S values . These same ∼40S complexes were identified using the ATPaseTb2 antibody ( Fig . 3C ) . This indicates that in the absence of subunit a , the F1-moiety is still attached to the mt membrane via a central stalk , a ring of subunit c and the peripheral stalk . Since the Fo subunit a is missing in these cells , the attachment of this complex to the membrane is presumably not as strong as in BF cells . Therefore , upon treatment with detergent to lyse the mitochondria , a majority of F1-ATPase is released and detected in lower S values while a small portion of the FoF1-complexes is preserved and sediments at higher S value fractions . Consequently , our data suggests that previously undetected membrane-bound FoF1-complexes are present in the mitochondria of the Dk trypanosomes . To verify our intriguing results , the FoF1-ATP synthase/ATPase complex was examined by an alternative method involving high resolution clear native electrophoresis ( hrCNE ) followed by western blot analysis . Results obtained using the antibody against subunit β revealed four predominant bands in both the PF427 and BF427 samples , representing various forms of the F1-moiety and FoF1-complexes ( Fig . 3D ) . The lowest band likely corresponds to the F1-ATPase ( subunits α , β , γ , δ , ε ) , while the one above it seemingly represents F1-ATPase with a ring of subunit c , as it has been described for the mammalian complex [41] . Noticeably , in each sample the antibody against ATPaseTb2 immunodetected only two major bands corresponding to the FoF1-monomer and -multimers . When the mt lysates of Dk164 and T . b . evansi cells were fractionated on a hrCNE , subunit β was mainly detected in the lower bands representing the F1-moiety and the FoF1-monomer , while only a weaker band was observed at the size depicting FoF1-multimers . But western blot analysis with the ATPaseTb2 antibody clearly demonstrated the existence of higher assemblies of the FoF1-ATPase complexes ( Fig . 3D ) . Thus , ATPaseTb2 is a membrane-bound subunit of the monomeric and multimeric FoF1-ATP synthase/ATPase complex in all four T . brucei cell types examined . Furthermore , no significant differences in size were observed between the FoF1-monomeric and -multimeric complexes detected in PF427 , BF427 and Dk cells , indicating that subunit a might be the only missing component of the assembled FoF1-ATPase complex in dyskinetoplastic cells . To assess the importance of ATPaseTb2 in the infective stage of T . brucei and to evaluate its functional association with FoF1-ATPase , an ATPaseTb2 RNA interference ( RNAi ) cell line was created . The expression of dsRNA was triggered by the addition of tet to the culture medium and the in vitro growth of the ATPaseTb2 knock-down ( KD ) cells was inspected for seven days ( Fig . 4A ) . Strikingly , a significant growth phenotype was already detected in RNAi induced cells on just the second day of tet addition . Presumably , the powerful selection forces acting on these cells missing a critical protein resulted in the emergence of a subpopulation that was no longer responsive to RNAi induction after only 120 hours , leading to the growth recovery of the culture ( Fig . 4A ) . This phenomenon is often reported for RNAi experiments in BF T . brucei [42] . The targeted KD of ATPaseTb2 was confirmed by western blot analysis of whole cell lysates harvested from an equivalent number of cells for ATPaseTb2 RNAi non-induced and induced cells throughout the RNAi time course ( Fig . 4B ) . Western blot analysis using the antibody against ATPaseTb2 exhibited a reduction of the targeted protein by 73% at day 2 of RNAi induction . A non-specific band detected on the same membrane was used to determine equal loading of the samples . With the efficient knockdown of ATPaseTb2 verified , an in vivo assay was employed to measure the Δψm in the ATPaseTb2 KD cell population ( Fig . 4C ) , since the Δψm is generated by the FoF1-ATPase in BF T . brucei [9] . Flow cytometry analysis was used to determine the changes observed in these cells stained with Mitotracker Red CMX-Ros , whose fluorescent intensity is proportionally dependent on the strength of the Δψm . A substantial decrease of the Δψm in the KD cell population was observed two days ( IND2 ) after RNAi induction , coinciding with the first time point to display a significant growth inhibition . Importantly , this result confirms the vital function of ATPaseTb2 within the FoF1-ATPase . To assess if the decrease in the Δψm is the result of an impaired ATP hydrolytic activity of the enzyme , the total ATPase activity was measured in mt lysates with or without azide , an inhibitor of the catalytic F1-ATPase [43] . Typically , the specific F1-ATPase activity represents ∼ 35–45% of total mt ATPase activity . Our results indicate that neither the total ATPase nor azide-sensitive activities were significantly altered between the non-induced and ATPaseTb2 RNAi-induced cells ( Fig . 5A ) . ATPase activity can also be visualized by in-gel activity staining when mitochondria from these cells are solubilized with dodecyl maltoside and separated on a 2–12% blue native PAGE ( BNE ) [41] . The specific ATPase staining revealed two major bands in non-induced cells , the predominant lower one representing the activity produced from the F1-moiety and the less significant higher band representing the hydrolytic activity of the monomeric FoF1-ATPase ( Fig . 5B ) . It is important to note that solubilizing mitochondria with any detergent can result in unintended structural consequences not found under normal physiological conditions , especially for a large complex comprised of a matrix moiety and a membrane embedded region . Although detergents tend to dissociate the F1-moiety from FoF1-ATPase complexes , it is intriguing that there is still a portion of ATP hydrolysis being produced from the complete FoF1-ATPase in the non-induced BF cells . Conspicuously , this activity is absent when ATPaseTb2 RNAi is induced ( Fig . 5B ) . The absence of activity for the FoF1-monomeric complex might be explained by complex instability . Thus , the structural integrity of the coupled FoF1-complex was further analyzed using hrCNE , which were loaded with digitonin-lysed mitochondria from non-induced and ATPaseTb2 RNAi induced cells . After transferring the separated native protein complexes to a nitrocellulose membrane , a specific antibody against the p18 subunit detected the F1-moiety as well as both monomeric and oligomeric FoF1-complexes ( Fig . 5C ) . Interestingly , we repeatedly observed that the abundance of the monomeric and oligomeric complexes is decreased after RNAi induction and by day 3 these complexes are not detected at all . These results were complemented by the sedimentation profile of FoF1-ATPase complexes on glycerol gradients . Mitochondria were isolated from non-induced RNAi cells and cells induced for two days ( IND2 ) . These were then lysed by 1% Triton X-100 and an equal amount of mt protein from each sample was fractionated on a 10–30% GG . The resolved fractions were then analyzed by western blot and the anti-β and anti-p18 antibodies demonstrated the effect of the ATPaseTb2 KD on the FoF1-ATPase sedimentation profile . As indicated in Fig . 5D and the corresponding scanning densitometry results listed in S2 Fig . , two days of RNAi induction resulted in a diminished amount of β and p18 signals from the higher S-values ( IND2 panels ) , while the GG sedimentation pattern of RNAi non-induced cells was more similar to wild type BF427 ( see Fig . 3C ) . These results are in agreement with the PF RNAi ATPaseTb2 cell lines , in which the stability of the monomeric and multimeric FoF1-ATP synthases was significantly affected upon RNAi [26] . Using a variety of methods employing various detergent types , we have established that the ATPaseTb2 subunit is an essential component of the FoF1-ATPase monomer and higher oligomers in the bloodstream stage of T . brucei . Furthermore , the depletion of ATPaseTb2 diminishes the abundance of membrane-bound FoF1-ATPase complexes , which directly initiates a substantial decrease in the Δψm that manifests as a strong growth phenotype . To further dissect the primary role of the membrane-bound ATPaseTb2 subunit , we investigated the effect of targeted gene silencing in transgenic Dk T . b . evansi , which facilitate the tet inducible expression of dsRNA . These Dk cells were transfected with an ATPaseTb2 RNAi vector and several positive clones were screened for a growth phenotype . As shown in Fig . 6A , the surprising inhibition of cell growth in RNAi induced cells appeared two days after tet induction and the doubling time remained reduced throughout the whole ten day experiment . Furthermore , western analysis of whole cell lysates harvested over the RNAi time course exhibited a significant reduction of ATPaseTb2 in the RNAi induced cells beginning on day 1 ( Fig . 6B ) , while the analysis of a non-specific band served as the loading control . Interestingly , if the only role of the subunit was to stabilize the proton pore , it presumably would not be essential in cells that lack mtDNA and have adapted to the loss of a functioning proton pore . However , these results suggest that ATPaseTb2 is important for FoF1-ATPase complexes in these unique cells and thus this cell line was further exploited to characterize the function of ATPaseTb2 . Since the Δψm in Dk cells is generated using the hydrolytic activity of the F1-moiety coupled with ATP/ADP translocation , the Δψm was examined in ATPaseTb2 non-induced and RNAi induced cell populations . As shown in Fig . 6C , a reduction of the Δψm was detected in the RNAi induced cell population after 2 ( IND2 ) and 3 days ( IND3 ) . Since the reduction of the Δψm already appears at day 2 of tet induction and the growth phenotype does not truly manifest until day 3 , we postulate that the observed decrease of the Δψm is a direct effect of ATPaseTb2 silencing and not a consequence of decreased cell viability . Furthermore , to assess if this decreased Δψm is caused by an impairment of the F1-ATPase hydrolytic activity , the total ATPase activity was measured in mt lysates with or without azide . Similar to BF RNAi ATPaseTb2 cells ( Fig . 5A ) , no changes in total or azide-sensitive activities were observed between the non-induced and ATPaseTb2 RNAi-induced cells ( Fig . 6D ) . The structural integrity of Dk FoF1-ATPase complexes was then investigated using mild non-denaturing hrCNE to fractionate mitochondrial proteins purified from BF427 cells and Dk cells either induced for ATPaseTb2 RNAi or not . A western blot containing BF427 mitochondria as a reference sample was probed with an anti-β antibody and used as marker to visualize each state of the F1-moiety , either isolated by itself or being partnered with the monomeric and multimeric FoF1-complexes . Strikingly , while a majority of the β subunit signal in Dk RNAi non-induced cells was observed in the region of isolated F1-ATPase , we also consistently detected an irrefutable signal corresponding to the FoF1-monomeric complex , presumably assembled intact with the exception of subunit a ( Fig . 6E ) . While there is evidence of mammalian rho cells containing a similar complex [44–46] , this is the first time it has been demonstrated in the comparable Dk trypanosomatids . Furthermore , the induction of ATPaseTb2 RNAi in the Dk cells led to the disappearance of the monomeric complex , leaving only the F1-ATPase to be detected on the western blot ( Fig . 6E ) . Once again , the evidence indicates that ATPaseTb2 is important for the structural integrity of the FoF1-ATPase and its depletion in Dk T . b . evansi cells leads to the disruption of the unique membrane-bound FoF1-ATPase complexes . To corroborate this observation , we purified mitochondria from Dk T . b . evansi non-induced and ATPaseTb2 RNAi-induced cells and lysed them with Triton X-100 . These gently disrupted lysates were then fractionated on glycerol gradients , which revealed that a majority of the signal from subunits β and p18 was located within fractions 6–10 , representing the F1-moiety ( Fig . 6F ) . Nevertheless , weaker bands representing the FoF1-ATPase complexes ( fractions 14–22 ) were again detected in the non-induced samples ( Fig . 6F NON panels ) . Noticeably , these same protein markers in the mt fractionation of ATPaseTb2 RNAi induced cells displayed a significant reduction in fractions 14–22 ( Fig . 6F IND4 panels ) , where the FoF1- monomer/oligomer sediments . This observation is quite surprising since the main role of this enzyme is to create the Δψm , presumably using only the hydrolytic activity of the matrix facing F1-moiety . In an attempt to determine if a portion of the F1-ATPase is membrane associated and that the disruption of this attachment is the reason for the observed decrease in the Δψm and the detected growth phenotype , we decided to visualize these complexes in situ using immunogold labelling followed by electron microscopy . Immunogold labeling with a primary anti-β antibody was performed on ultrasections of Dk T . b . evansi non-induced ( NON ) and ATPaseTb2 RNAi induced cells for 3 ( IND3 ) and 5 ( IND5 ) days . Electron micrographs illustrate that Dk trypanosomes have a simple and reduced mitochondrion ( S3A Fig . ) , lacking both cristae and a typical kinetoplast structure , as previously reported for the dyskinetoplastic T . b . evansi [17] . Whenever these reduced double-membrane organelles could be unequivocally identified , we tallied the number of immunogold beads either in the close proximity of the mt membrane or in the matrix ( S3A Fig . ) . Strikingly , 70% of the gold particles identified were found associated with the mt membrane in the NON images ( S3B Fig . ) . From the 113 images captured from either NON , IND3 or IND5 samples , we performed a Chi-squared analysis on the immunogold beads associated with the mt membrane to determine if their distribution is random ( S3C Fig . ) . With two degrees of freedom , the calculated χ2 = 43 . 5 ( p < 0 . 0001 ) , which signifies that the difference between the observed and expected particles was statistically significant . These values were then plotted to determine the relative labelling index ( RLI = Nobs/ Nexp ) for the mt membrane associated gold particles ( S3D Fig . ) . This analysis reveals that the observed number of gold particles found in close proximity to the mt membrane for the NON electron micrographs is significantly greater than if the particles were randomly distributed ( RLI = 1 ) , while the opposite is true for the IND5 samples . Since we understand the general subjectivity of this method , we performed this experiment in a blinded study and only considered a gold bead to be associated with the mt membrane if it was in the immediate proximity of the membrane . For the very few scored beads that could potentially fall in a membrane-bound grey zone , it should be noted that the physical distance from the subunit β antigen to the gold particle can be up to 15–20 nm , not to mention that subunit β itself is projected from the membrane by subunit γ by as much as 13 nm [47 , 48] . Furthermore , very few immunogold beads were detected outside of defined mt structures and each mt image only contained 1–8 immunogold beads ( average of 2 . 4 beads/mt image ) , suggesting that the labelling was very specific for our abundant mt antigen . It is important to note , that utilizing an assay that does not require the use of any detergents , we demonstrated that a majority of the F1-ATPase is located within close proximity of the mt membrane and this association is disrupted when ATPaseTb2 is depleted .
The T . brucei FoF1-ATPase has obtained unique properties in comparison to its eukaryotic counterparts . The occurrence of novel subunits combined with the lack of typical eukaryotic subunits that compose the Fo membrane-bound moiety and the peripheral stalk is intriguing from an evolutionary viewpoint . Because these atypical subunits lack significant homology to known proteins , questions arise concerning their authenticity as genuine subunits of the complex and what their function and localization within the complex might be . Here , we focused on the functional characterization of one of the trypanosomatid specific subunits , ATPaseTb2 in the bloodstream stages of T . brucei . Notably , we determined that its function is essential for maintaining the normal growth rate of the infectious stage of T . brucei and also for the important veterinary dyskinetoplastic parasite , T . b . evansi . Additional analyses demonstrated that the ATPaseTb2 is membrane embedded and a component of the FoF1-ATPase that is present in both BF and Dk cells . Furthermore , the depletion of this Fo-moiety subunit results in a decreased Δψm and a loss of FoF1-ATPase complexes . Combined with bioinformatic tools that predict a transmembrane domain and identify a low homology to subunit d , we suggest that ATPaseTb2 might be a component of the peripheral stalk of the FoF1-ATPase . The peripheral stalk is indispensable for ATP synthesis as it serves to impede the movement of the catalytic α3β3 headpiece while the proton motive force rotates the c-ring and the connected asymmetrical central stalk in a way that imposes conformational changes in the catalytic nucleotide binding sites of the β subunits . The peripheral stator is also essential when the complex harnesses the energy provided from ATP hydrolysis to drive the rotation of the enzyme in reverse , allowing protons to be pumped across the mt inner membrane to produce the Δψm when physiological conditions dictate . The bovine and yeast peripheral stalk is composed of four subunits , OSCP , b , d , and F6/h ( bovine/yeast nomenclature ) . Subunit b ( ∼20kDa protein ) contains two trans-membrane segments at the N-terminus , while the rest of the protein is hydrophilic—protruding into the matrix and interacting with OSCP and subunit d [23 , 49–51] . The interaction between subunit b and OSCP is stabilized by subunit F6 [22] . The predominantly hydrophilic subunit d interacts with all three mentioned subunits and it has been shown to be essential for the function of the FoF1-complex . The yeast knock-out of subunit d was characterized by the de-attachment of the catalytic F1-ATPase from the protonophoric sector , the loss of detectable oligomycin sensitive ATPase activity and the absence of subunit a in the Fo-moiety [36] . Notably , of these four subunits , the only homolog identified in the T . brucei genome is OSCP and its protein product was identified as a bona fide subunit of the purified FoF1-ATP synthase [26] . Homologs for subunits b , d , and F6 are missing and most likely have been replaced by the novel proteins that associate with the T . brucei FoF1-ATP synthase . Here we demonstrated that the ATPaseTb2 subunit is a membrane-bound protein , containing one predicted transmembrane domain , with a large hydrophilic region extending into the matrix that possesses a low homology to the bovine subunit d . In accordance with other yeast or bovine subunits of the peripheral stalk , the down-regulation of ATPaseTb2 affects the stability of the FoF1-complex and its oligomycin sensitivity ( this study , [26] ) . Considering the relatively large molecular mass of ATPaseTb2 ( 43 kDa ) , it is plausible to speculate that ATPaseTb2 functionally represents the membrane-bound subunit of the peripheral stalk fused with subunit d , offering an early attempt by eukaryotes to add layers of complexity that will allow for greater adaptability . It is noteworthy to mention that a species-specific architectural variant of the peripheral stalk was also proposed for colorless and green algae , where the novel subunits Asa2 , Asa4 and Asa7 fulfil a structural role in forming the peripheral stalk [52 , 53] . Other discrepancies from the typical eukaryotic enzyme model can be found in the ciliate Tetrahymena thermophila , representing the superphylum of Alveolates , where the homolog of the conserved subunit b has so far not been identified in the genome . Instead , three novel proteins detected in this ATP synthase complex have been proposed to substitute for subunit b [28] . Thus , it seems that the composition and overall structural appearance of the FoF1-ATP synthase from organisms representing different lineages other than Opisthokonts , might be more diverse than previously thought and deserve more attention from basic researchers to elucidate alternate evolutionary solutions to a common thread of life . T . brucei is a member of the Excavata clade and represents a desirable model to study the function and regulation of the FoF1-ATP synthase/ase complex since it utilizes the proton motive force to produce ATP in the PF life stage , while in the BF stage it hydrolyzes ATP to pump the protons necessary to create the Δψm . Moreover , dyskinetoplastic T . b . evansi employs the hydrolytic activity of the F1-ATPase and the electrogenic exchange of ADP3- for ATP4- by the AAC as yet another strategy to generate the Δψm ( Fig . 7 ) . Therefore , it was surprising that the ATPaseTb2 protein was detected in all four examined cell types ( PF , BF , Dk 164 and T . b . evansi ) , albeit at a much lower abundance in BF and Dk cells , but that is in full agreement with the reduced mitochondria previously defined for these two forms [39] . Furthermore , when PF and BF mt lysates were resolved by GG and hrCNE , ATPaseTb2 was only detected when co-localized with the monomeric and oligomeric FoF1-ATP synthase/ase complexes . Interestingly , similar results were obtained from the mt lysates of cells lacking mtDNA . Considering that only the F1-ATPase was previously assumed to be important for maintaining the Δψm in the Dk trypanosomes , the presence of these complexes is intriguing . Nevertheless , a proteomics study performed with osteosarcoma 143B ρ° cells revealed that in addition to the F1-ATPase subunits , subunit d of the peripheral stalk was also identified [54] . In a similar project involving fibroblast MRC5 ρ° cells , the oligomycin insensitive complex was purified and determined to contain several Fo ( b and c ) subunits along with a couple of peripheral stalk subunits ( OSCP , d ) . Notably , this complex was loosely associated with the mt membrane [45] . Furthermore , several additional studies have shown that the F1-ATPase in yeast or mammalian cells lacking a mt genome ( i . e . mammalian subunits a/ATP6 , A6L; yeast subunits 6 , 8 and 9 ) is membrane associated , with the attachment most likely occurring via the central and/or peripheral stalk [44 , 46 , 55 , 56] . Our data also suggest that since ATPaseTb2 membrane-bound complexes ( monomeric and multimeric ) do not display significant differences in their native size or sedimentation values , then the mt encoded subunit a might be the only missing subunit from these complexes . Such a uniquely structured complex can be explained by the current model for the assembly of the FoF1-ATPase , in which the Fo subunit a is the last protein incorporated into the enzyme to ensure that it is able to function properly before introducing a complete proton pore . Furthermore , this integration is dependent on the presence of subunit b of the peripheral stalk [57] . RNAi silencing of ATPaseTb2 in BF cells caused a strong growth phenotype concurring with a decreased Δψm . Importantly , ATPaseTb2 also proved to be essential for maintaining the normal growth rate and the Δψm of Dk cells . Combined with the hrCNE and glycerol gradient sedimentation assays that revealed a decrease of the high molecular weight complexes , but not the F1-ATPase , our studies suggest that these membrane-bound enzymes might significantly contribute to the membrane potential in Dk cells . The biological significance of this localization of the F1-ATPase might be to efficiently coordinate the activity of this enzyme with its substrate transporter , both of which are responsible for producing the Δψm in Dk . By restricting the F1-ATPase to the mt membrane , it keeps it within the close vicinity of its biochemically functional partner , the AAC . In this way , it helps ensure that the hydrolysis of ATP results in a relatively concentrated region of substrate for the mt membrane embedded AAC , especially in a mitochondrion lacking defined cristae and the microenvironments they create . Indeed , there is precedence for such a close spatial collaboration to increase efficiency , as an actual physical interaction between these two enzymes has been reported as the ATP synthasome in mammals [40 , 58] . We are currently building a set of tools to perform additional functional assays and imaging techniques that we hope will further resolve the function of the FoF1-ATPase complexes in these pathogenic trypanosomes .
The ATPaseTb2 ( Tb927 . 5 . 2930 ) RNAi constructs targeted an 825 bp fragment of the gene that was PCR amplified from T . brucei strain 427 genomic DNA with the following oligonucleotides: FW—CACGGATCCATGCGCCGTGTATC , REV-CACCTCGAGTTCGGCCCGATC . Utilizing the BamHI and XhoI restriction sites inherent in the primers ( underlined ) , this fragment was cloned into the p2T7-TA blue plasmid [59] ( used to create a BF RNAi cell line ) and the p2T7–177 plasmid [60] ( used to generate a Dk RNAi cell line ) . For the inducible expression of ATPaseTb2 fused with a C-terminal 3xv5 tag , the ATPaseTb2 coding sequence was PCR amplified ( Fw: ACAAAGCTTATGC GCCGTGTATC , Rev: ATAGGATCCGTGATGGGCCTTTTTC ) and cloned into the pT7_v5 vector using HindIII and BamHI restriction enzymes [61] . The pLEW79MHTAP vectors for the tet inducible expression of ATPaseTb2 and Tb927 . 5 . 1710 ( FoF1-ATPase subunit p18 ) TAP-tagged proteins were previously described in [26] . Bloodstream form T . brucei brucei Lister 427 strain , stable acriflavin-induced dyskinetoplastic T . b . brucei EATRO 164 [18] and dyskinetoplastic T . b . evansi Antat 3/3 [62] were all grown in vitro at 37°C and 5% CO2 in HMI-9 media containing 10% FBS . The BF427 single marker cell line [63] and a T . b . evansi cell line [14] , both constitutively expressing the ectopic T7 RNA polymerase and tet repressor ( TetR were used for the tet inducible expression of dsRNA and the TAP- and v5-tagged proteins . Each plasmid was linearized with NotI enzyme and transfected into the appropriate cell line as described previously [63] . Linearized p2T7–177 was targeted to the minichromosome 177 basepair repeat region , while linear pTv5 and pLEW79MHTAP vectors were integrated into the rDNA intergenic spacer region [60 , 61 , 63] . The induction of RNAi and ectopically expressed tagged proteins was triggered by the addition of 1 μg/ml of tet into the media . Cell densities were measured using the Z2 Cell Counter ( Beckman Coulter Inc . ) . Throughout the analyses , cells were maintained in the exponential mid-log growth phase ( between 1x105 to 1x106 cells/ml ) . The TAP protocol was generously provided by L . Jeacock and A . Schnaufer ( personal communication ) and optimized for BF cells . Briefly , 2x108 cells were harvested and lysed for 20 min on ice with 1% Triton X-100 in an IPP150 buffer ( 150 mM NaCl , 0 , 1% NP40 , 10mM Tris-HCl pH 8 . 0 ) containing Complete protease inhibitors ( Roche ) . The lysates were then cleared by centrifugation ( 16 , 000 g , 15 min , 4°C ) . Meanwhile , IgG antibodies were covalently bound to Dynabeads M-270 Epoxy ( Invitrogen ) using a protocol previously described [64] . Charged beads were blocked with 1% BSA and then incubated with the cleared cell lysates ( 4°C for 2 hours , constantly rotating ) . The beads were then washed three times with IPP150 and once with a TEVCB buffer ( 150 mM NaCl , 0 . 1% NP40 , 0 . 5 mM EDTA , 1 mM DTT , 10 mM Tris-HCl , pH 8 . 0 ) . Finally , the bound protein complexes were released by AcTEV protease ( Invitrogen ) cleavage and the eluate was analyzed by SDS-PAGE . Crude mt vesicles were obtained by hypotonic lysis as described in detail earlier [65] . Briefly , cell pellets from ∼2x109 cells were washed with SBG ( 150 mM NaCl , 20 mM glucose , 20 mM NaHPO4 , pH 7 . 9 ) , resuspended in DTE ( 1 mM Tris , 1 mM EDTA , pH 8 . 0 ) and homogenized in a Dounce homogenizer . Alternatively , for a smaller scale hypotonic isolation , cell pellets from ∼1x109 cells were washed in NET buffer ( 150 mM NaCl , 100 mM EDTA , 10 mM Tris , pH 8 . 0 ) , resuspended in DTE and homogenized by passing through a 25G needle . To re-create the physiological isotonic environment , 60% sucrose was immediately added to the lysed cells to reach a final concentration of 250 mM . Samples were spun down at 15 , 000 g , 10 min , 4°C to clear the soluble cytoplasmic material from the lysates . The resulting pellets were resuspended in STM ( 250 mM sucrose , 20 mM Tris pH 8 . 0 , 2 mM MgCl2 ) , supplemented with a final concentration of 3 mM MgCl2 and 0 . 3 mM CaCl2 before incubating with 5 μg/ml DNase I for 1hr on ice . Then an equal volume of STE buffer ( 250 mM sucrose , 20 mM Tris pH 8 . 0 , 2 mM EDTA pH 8 . 0 ) was added and the material was centrifuged at 15 , 000 g , 10 min , 4°C . Pellets enriched with the mt membrane vesicles were washed in STE and kept at -70°C . Protein samples were separated on SDS-PAGE , blotted onto a PVDF membrane ( PALL ) and probed with the appropriate monoclonal ( mAb ) or polyclonal ( pAb ) antibody . This was followed by incubation with a secondary HRP-conjugated anti-rabbit or anti-mouse antibody ( 1:2000 , BioRad ) . Proteins were visualized using the Pierce ECL system ( Genetica/Biorad ) on a ChemiDoc instrument ( BioRad ) . When needed , membranes were stripped at 50°C for 30 min in a stripping buffer ( 62 . 5 mM Tris pH 6 . 8 , 100 mM mercapthoethanol , 2% SDS ) and re-probed . The PageRuler prestained protein standard ( Fermentas ) was used to determine the size of detected bands . Primary antibodies used in this study were: mAb anti-v5 epitope tag ( 1:2000 , Invitrogen ) , mAb anti-mtHsp70 ( 1:2000 ) [66] , pAb anti-AAC ( 1:2000 ) [67] , pAb anti-MRP1 ( 1:1000 ) [68] , pAb anti-APRT ( 1:1000 ) and pAb anti-enolase ( 1:1000 ) [69] . Antibodies against ATPaseTb2 ( 1:1000 ) , ATPaseTb1 ( 1:1000 ) , subunit β ( 1:2000 ) and subunit p18 ( 1:1000 ) were prepared for the purpose of this study . The open reading frames of the respective genes without their predicted mt localization signal were cloned into the E . coli expression plasmid , pSKB3 . The proteins were overexpressed in BL21 E . coli cells and purified under native conditions ( subunit β ) or denatured conditions ( ATPaseTb2 , ATPaseTb1 and p18 ) using a 6 M guanidinium lysis buffer and 8 M urea binding buffer . The denatured proteins were then refolded with a step-wise dialysis procedure that included 0 . 5 M arginine in the refolding buffer . Native and refolded antigens were sent to Pineda ( Antikörper-Service , Germany ) or Davids Biotechnologie ( Regensburg , Germany ) for polyclonal antibody production . The Δψm was determined utilizing the red-fluorescent stain Mitotracker Red CMX-Ros ( Invitrogen ) . Cells in the exponential growth phase were stained with 100 nM of the dye for 30 min at 37°C . Cells were pelleted ( 1 , 300 g , 10 min , RT ) , resuspended in 2 ml of PBS ( pH 7 . 4 ) and immediately analyzed by flow cytometry ( BD FACS Canto II Instrument ) . For each sample , 10 , 000 events were collected . Treatment with the protonophore FCCP ( 20 μM ) was used as a control for mt membrane depolarization . Data were evaluated using BD FACSDiva ( BD Company ) software . ATPase activity was measured with the Sumner assay , which is based on the release of free phosphate when ATP is hydrolyzed by the enzyme as described earlier [14 , 70] . Briefly , crude mt lysates were obtained from 2x108cells by SoTe/digitonin extraction ( 0 . 015% digitonin , 0 . 6 M Sorbitol , 2 mM EDTA , 20 mM Tris-HCl pH 7 . 5 ) . Mt pellets were resuspended in an assay buffer ( 200 mM KCl , 2 mM MgCl2 , Tris-HCl pH 8 . 0 ) and the 20 min reaction was initiated by the addition of ATP to a final concentration of 5 mM . Where indicated , samples were pre-treated with the F1-ATPase specific inhibitor , sodium azide ( 2 mM ) for 10 min at 37°C . The 100 μl enzymatic reactions were deproteinated by the addition of 1 . 9 μl of 70% perchloric acid . After a 30 min incubation on ice , the samples were spun down ( 16 , 000g , 10 min , 4°C ) and 90 μl of the supernatant was incubated for 10 min with 0 , 5 ml of Sumner reagent ( 8 . 8% FeSO4 . 7H2O , 375 mM H2SO4 , 6 . 6% ( NH4 ) Mo7O24 . 4H2O ) [62] . 200 μl was then transferred to a 96 well plate and the absorbance was measured at 610 nm using a Tecan Infinite plate reader ( Infinite M200Pro , Tecan ) . To calibrate the assay , a standard curve was calculated from the absorbance values of linear inorganic phosphate samples ( 0–2 mM ) . Blue native PAGE ( BNE ) of mt lysates , followed by in-gel activity staining was adapted from published protocols [26 , 32] . Briefly , mt vesicles from ∼2x109 cells were resuspended in a mt lysis buffer ( 0 , 75 M amino-n-caproic acid—ACA , 50 mM Bis-Tris , 0 , 5 mM EDTA , pH 7 . 0 ) , lysed with 2% dodecylmaltoside ( DDM ) for 1hr on ice and then cleared by centrifugation ( 16 , 000g for 30 min , 4°C ) . The protein concentration of each mt lysate was determined by a Bradford assay ( BioRad ) , so that 100 μg of total mt protein could be mixed with 1 M ACA and 5% Coomassie brilliant blue G-250 before being loaded on a 2–12% Bis-Tris BNE gel . Immediately after the run ( 3 hr , 100 V , 4°C ) , the gel was incubated overnight in an ATPase reaction buffer ( 35 mM Tris pH 8 . 0 , 270 mM glycine , 19mM MgSO4 , 0 . 3% Pb ( NO3 ) 2 , 11 mM ATP ) . The protocol for hrCNE was adapted from published studies [71 , 72] . In summary , crude mt vesicles from ∼5x108 cells were resuspended in a mt lysis buffer ( 2 mM ACA , 50 mM Imidazole-HCl , 1 mM EDTA , 50 mM NaCl , pH 7 ) and lysed for one hour on ice with 4 mg digitonin/1 mg protein . The samples were spun down at 16 , 000 g for 30 min and the cleared lysate protein concentrations were determined by a Bradford assay . Samples were mixed with a 5x loading dye ( 0 . 1% Ponceau-S , 50% glycerol ) and loaded onto a 3%-12% native gradient gel . After electrophoresis ( 3 hr , 100 V , 4°C ) , the resolved mt lysates were transferred onto a nitrocellulose membrane ( overnight , 20 V , 4°C ) and probed with selected antibodies . Na2CO3 extraction of mt membranes was adapted from a previously published protocol [73] . Mt vesicles from 1x109 cells were isolated by hypotonic lysis as described above . The resulting supernatant from the 25G needle homogenization step was kept as a cytosolic fraction ( CYTO ) . The mt pellet was further treated with digitonin ( 80 μg/ml ) for 15 min on ice to disrupt the mt outer membrane . The material was then cleared by centrifugation ( 12 , 000 g , 20 min , 4°C ) and the pelleted mitoplasts were resuspended in 0 . 1 M Na2CO3 buffer ( pH 11 . 5 ) and incubated for 30 min on ice . A final ultracentrifugation step ( 100 , 000 g , 4°C for 1 hr ) performed in an SW50Ti rotor of a Beckman Instrument resulted in a supernatant comprised of proteins from the mt matrix ( MX ) , including stripped peripheral membrane proteins , and a pellet containing integral proteins isolated from the mt membrane fraction ( M ) . Hypotonically purified mt vesicles from ∼2 . 5x109 cells were resuspended in a GG lysis buffer ( 10 mM Tris , pH 7 . 2 , 10 mM MgCl2 , 200 mM KCl , 1mM DTT ) and lysed with 1% Triton X-100 ( 30 min , on ice ) . The lysates were cleared by a centrifugation step ( 2x 16 , 000 g , 30 min , 4°C ) and the protein concentration was determined by a Bradford assay . Mt cleared lysates were resolved by ultracentrifugation ( Beckman Instrument , SW40 rotor ) at 38 , 000 g for 5 hr on 11ml 10–30% GG , which was poured manually or using the Gradient Station ( Biocomp ) according to the manufacturer’s protocol . The glycerol gradients were then fractionated either manually or with the Gradient station and 500μl fractions were stored at -70°C . Cells ( ∼5x107 ) were pelleted ( 1300 g , 10 min , RT ) , washed in PBS ( pH 7 . 4 ) and immediately fixed in a 4% formaldehyde/ 0 . 1 M phosphate buffer . Samples were then dehydrated at -10°C through a series of seven steps that increased the concentration of ethanol from 30% to100% , pausing at each step for 1 hour . Next , the samples were embedded in LR White Resin ( Electron Microscopy Sciences ) and polymerized ( UV light , 48 hours at -10°C ) . Ultrathin sections were prepared by the Ultracut UCT ultramicrotome ( Leica ) and mounted on nickel grids . Prepared sections were blocked in 5% BSA , labelled with primary anti-β antibody ( 1:10 dilution ) , washed three times with PBS and incubated with protein A conjugated to 10 nm gold particles ( 1:100 dilution , Aurion ) . The immunogold labelled grids were contrasted , carbon coated and examined by the TEM ( JEM-1010 , Jeol ) . Grids , which were immunolabeled with only the protein A conjugated to 10nM gold particles , were used as a negative control . The number of gold particles was statistically evaluated as described earlier [12 , 74] . Using ImageJ software ( NIH , USA ) , a grid consisting of squares ( test points , P ) with constant size ( 16 , 105 μm2 ) was superimposed randomly on the electron micrographs of identified mitochondria . All test squares and immunogold particles within the immediate proximity to the mt membrane were counted separately for each micrograph . In order to statistically evaluate the labelling , all observed gold particles ( Nobs ) and all test points ( P ) from 113 images captured from NON , IND3 and IND5 samples were tallied . Expected numbers of gold particles ( Nexp ) for each sample were calculated as ( total sum Nobs x P ) /total sum P . To determine if the difference between the observed and expected particles was significant and not due to random fluctuations , a Chi-squared analysis χ2 = ( Nobs—Nexp ) 2/ Nexp was performed using GraphPad QuickCalcs calculator ( www . graphpad . com/quickcalcs ) . In addition , the relative labeling index ( RLI ) , where the predicted RLI = 1 for random labelling , was calculated as RLI = Nobs/ Nexp . ATPaseTb2 ( Tb927 . 5 . 2930 ) , ATPaseTb1 ( Tb927 . 10 . 520 ) , TbAAC ( Tb927 . 10 . 14820/30/40 ) , p18 ( Tb927 . 5 . 1710 ) , mtHsp70 ( Tb927 . 6 . 3740 ) , MRP1 ( Tb927 . 11 . 1710 ) , APRT ( Tb927 . 7 . 1780 ) , enolase ( Tb427 . 10 . 2890 ) , subunit β ( Tb927 . 3 . 1380 ) . | The presence of the FoF1-ATP synthase in every aerobic organism suggests that evolution has settled on a basic blueprint for the complex rotary motor capable of synthesizing life’s universal energy currency—ATP . However , compared to yeast and mammalian models of the FoF1-ATP synthase , several recent studies have reported unique structural and functional features of this complex from organisms representing the clades of Chromalveolata , Archaeplastida and Excavata . One of the most striking cases is observed in trypanosomes , important parasites of humans and animals . Notably , the FoF1-ATP synthase/ATPase switches from synthesizing ATP in the insect vector life stage to hydrolyzing ATP in their mammalian hosts to generate the essential mitochondrial membrane potential ( Δψm ) . Moreover , this indispensable FoF1-ATPase contains up to 14 trypanosome-specific subunits . Here we characterize one such novel subunit , ATPaseTb2 . We demonstrate that this subunit is crucial for the survival of the infectious stage of trypanosomes , part of the fully assembled FoF1-complex and it is essential for maintaining the Δψm . Given the enzyme’s irreplaceable function and extraordinary composition , we believe that the FoF1-ATPase is an attractive drug target . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | ATPaseTb2, a Unique Membrane-bound FoF1-ATPase Component, Is Essential in Bloodstream and Dyskinetoplastic Trypanosomes |
Brucellosis and coxiellosis are known to be endemic in ruminant populations throughout Afghanistan , but information about their prevalence and factors that affect prevalence in householders and livestock under diverse husbandry systems and pastoral settings is sparse . We conducted a cross-sectional survey to investigate the seroprevalence of brucellosis and Coxiella burnetii in humans and livestock in six secure districts in Herat from 26th December 2012–17th January 2013 . A total of 204 households with livestock were surveyed in six Kuchi and five sedentary type villages . Blood samples from 1 , 017 humans , 1 , 143 sheep , 876 goats and 344 cattle were tested for brucellosis and Q fever . About one in six households ( 15 . 7% ) had at least one Brucella seropositive person , about one in eight households ( 12 . 3% ) had at least one Brucella seropositive animal and about one in four ( 24 . 5% ) had either seropositive animals or humans . Ninety-seven percent of households had at least one C . burnetii seropositive person and 98 . 5% of households had one or more C . burnetii seropositive animals . Forty- seven householders had serological evidence of exposure to both C . burnetii and Brucella and eight animals were serologically positive for both diseases . Drinking unpasteurised milk ( OR 1 . 6 ) , treating animals for ticks ( OR 1 . 4 ) , milking sheep ( OR 1 . 4 ) , male gender ( OR 1 . 4 ) and seropositivity to Brucella ( OR 4 . 3 ) were identified as risk factors for seropositivity to C . burnetii in householders . Household factors associated with households having either Brucella seropositive animals or humans were Kuchi households ( OR 2 . 5 ) , having ≤4 rooms in the house ( OR 2 . 9 ) and not owning land ( OR 2 . 9 ) . The results from this study provide baseline information for the planning and monitoring of future interventions against these diseases . The implementation of this study greatly improved collaboration , coordination and capability of veterinary and public health professionals from government , NGOs and donor funded projects .
Brucellosis and coxiellosis are known to be endemic in ruminant populations throughout Afghanistan but information about their prevalence and factors that affect prevalence in householders and livestock under diverse husbandry systems and pastoral settings is sparse . Brucellosis in animals results in economic losses due to decreased productivity from abortions and reduced milk yield while the disease in humans can be severely debilitating , often with long- term adverse consequences for health [1] . C . burnetii causes abortions in domestic ruminants and fever , pneumonia , meningo-encephalitis and hepatitis in acute cases of Q fever in humans and endocarditis in chronic cases [2 , 3] . Little is known about the epidemiology of C . burnetii in animals and humans in Afghanistan although a United Nations Food and Agriculture ( FAO ) funded study found serological evidence of its occurrence in Bamyan province in 2011 [4] . Brucellosis and Q fever are likely to be severely under-diagnosed and reported as they are not specifically included in the list of diseases in the donor-funded Essential Package of Hospital Services and Basic Package of Health Services . Until the start of this study there were no facilities for testing for C . burnetii at the central veterinary and public health laboratories in Kabul or elsewhere in the Republic . It is highly likely that both B . melitensis and B . abortus are present in Afghanistan but there is currently no in-country diagnostic capability for their culture or differentiation . The objectives of the study were to
Estimates of prevalence ( at the person and animal level ) were calculated using robust standard errors ( variance linearization ) to account for clustering at the village and household levels . Variance linearization ( also known as robust standard errors clustered on village ) was used to estimate the variance of all prevalences and hence to derive the confidence intervals ( CI ) . Information about the total sample size of the population at risk at each of the levels ( village , household and individual ) was not recorded so neither sampling weights nor finite population correction factors were applied in the analyses . Multivariable models were built to evaluate factors that influenced seropositivity to Brucella and C . burnetii in humans , livestock and households . Data which could conceivably be associated with disease status were first examined with descriptive statistics and tested for associations between the outcome variables as a first screening process and variables significant at p < 0 . 10 were entered into the multivariate models . Multivariable mixed logistic regression models were built with random effects for village and household ( except for the household-level analyses which just included a random effect for village ) . In some models , there was zero village-level variance once the household random effect was incorporated and in these cases the village-level random effect was dropped . Models were built using a manual backwards elimination process . Normality and heteroscedasticity of all random effects were evaluated graphically and unless noted , were considered acceptable . Statistical analyses were performed using Epi Info version 7 . 1 . 2 . 0 , and Stata 13 ( StataCorp LP ) . Graphics were constructed in Excel ( Microsoft ) . The study was approved by the Islamic Republic of Afghanistan Department of Public Health’s Institutional Review Board ( Approval numbers 221665 and 221769 ) . Study participants signed a consent form agreeing to participate prior to the survey and informed signed consent was obtained from all adult study participants and parents of children <18 years old at the time of sampling . Livestock owners provided verbal consent for sampling of their animals . Ethics approval for sampling of the animals for this study was not required for taking blood because it is a normal husbandry procedure and was supervised or conducted by veterinarians . This is in line with Animal Welfare legislation in New Zealand and current practice in the Ministry of Agriculture in Afghanistan .
Despite its conceptual appeal [6] and successful application over the past 20 years for infectious diseases such as bovine spongiform encephalopathy , highly pathogenic H5N1 avian influenza , C . burnetii and leptospirosis , linked public health and veterinary field investigations of important zoonoses for which the reservoir hosts are farmed ruminants have been rarely conducted in low income countries . Serological studies involving humans and their livestock have been reported for brucellosis in Kyrgyzstan [7] , Mongolia [8] and Egypt [9] , for C . burnetii in Egypt [10] and for both brucellosis and Q fever in Chad [11] and the authors are aware of several unpublished investigations for brucellosis in Central Asia and China . Our study was made possible by the terms of reference for the series of in-country studies in South Asia conducted under the Massey University European Commission funded and World Bank administered South Asia project which embodied veterinary and public health collaboration and allowed for investigator’s judgement for collection of information about the three zoonoses of interest which are characterised by multiple allied risk factors for transmission . Our inclusion of brucellosis and Q fever as the diseases of interest approach has obvious marginal cost advantages over study designs which are confined to one disease of interest . Householder and animal Brucella seroprevalences were 5 . 2% ( 95% CI 1 . 8 , 14 . 3 ) and 1 . 3% ( 95%CI 0 . 7 , 2 . 2 ) respectively with 15 . 7% of households having seropositive householders and 12 . 3% seropositive animals . There was evidence of clustering of infection in animals within a household and Kuchi village animals were more likely to be seropositive than those in sedentary villages ( Table 1 ) . The evidence of clustering is presented in the moderate level of household variance ( Table 1 ) The relatively constant prevalence over all age bands in householders ( Fig 1B ) indicates early age exposure to infection and contrasts with a finding of no seropositive animals in the 12–23 month age group of animals ( Fig 3B ) . Humans are susceptible to infection at all ages , whereas with the exception of latent infection , successful infection in animals mainly occurs during pregnancy and most of the animals in the 12–23 month age band would have been either non-pregnant or in early first-parity pregnancy at the time of sampling . Seroprevalences of 1 . 4% in sheep and 1 . 5% in goats are highly indicative of B . melitensis infection for which the predominant reservoir hosts are small ruminants . Serology cannot distinguish Brucella species and the single test positive in cattle could be due to either B . melitensis or B . abortus . The high householder and animal C . burnetii seroprevalences of 63 . 9% ( 95%CI 57 , 70 . 3 ) and 43 . 4% ( 95%CI 34 . 7 , 52 . 5 ) respectively with almost all households ( 97 . 5% and 98 . 5% ) having evidence of exposure to infection in householders and household livestock were unexpected findings . Prevalences varied among villages but not between Kuchi and sedentary villages or between sheep and goats . As with Brucella the prevalence was markedly lower in cattle than in sheep and goats and there was evidence of early age exposure to infection in householders and animals . Participation was voluntary and householders were not randomly selected but given the sample size of five and the average Afghan household size of 7 . 6 and our restriction of ages to between 9 and 60 years of age it was interesting that more female than male householders participated . Gender inequality is an issue for health care in Afghanistan and we suspect that one of the reasons for relatively more females participating was that they welcomed the opportunity afforded by the visits of health workers to their homes . Also females are at home most of the time and some males may have been absent due to work commitments . Given the sample size of five persons per household we consider any bias from possible over-representation of persons who were unwell to be unlikely and of no consequence . The overall seroprevalence estimates for householders and animals provide the best available estimates of risk of exposure to infection in the populations of interest with the caution that the estimates are conditional on uncertainties regarding the operating characteristics of the commercial ELISAs which were used and the prevalence of false positives from other sources of non-specific reactions . The precision of the estimate for householders is less certain than the estimate for animals because it is not known for how long C . burnetii and Brucella titres persist in humans . Sheep and goats remain infected after initial infection with B . melitensis and are serology test positive for life [12] unlike humans where infection , apart from a small proportion of chronic relapsing cases , is usually self-limiting . The nature of C . burnetii infection is probably similar; infections in animals may persist for long periods [13 , 14] and most human infections are self-limiting . We considered household prevalences derived by the study to be the most informative measures of risk of exposure to infection . However , the prevalences of households with evidence of exposure to Brucella infection in animals and householders are biased downwards to an unknown extent because only ten animals were tested in each household and the prevalence of Brucella test positive animals was low . The bias was of no consequence for C . burnetii because of the high proportions of households with test positive householders and animals . Confidence intervals of prevalence estimates were quite wide because multistage sampling with village as the primary sampling unit was done and this was accounted for in the analysis . True random sampling at each level ( village , household and individual ) with participation of 94% of selected households ensured that these are valid estimates . We also remind readers here that the study population consisted of households with animals so the estimates are only valid for that sub-population and not for the general population . Many of the factors examined were village- or household-level factors for which the effective sample sizes were the numbers of villages and households . Despite this limitation and the study’s reliance on serology with no information as to when infection occurred , the multivariable analyses were performed in an attempt to identify at least some of the risk factors for infection in householders . The fact that the only factor identified for Brucella seropositivity in householders was Coxiella seropositivity may have been due to its low prevalence which would have limited the power of the analysis for this infection . However , male gender , milking sheep , drinking raw milk , treating animals for ticks and Brucella seropositivity were identified for C . burnetii ( see Table 4 ) and these are in accord with current understanding of the disease which considers the main routes of infection to be via infected milk and birth products and inhalation of contaminated dust . The nature of C . burnetii infections has been well described in recent comprehensive reviews [15 , 16 , 17] . The organism can survive well for extremely long times in its desiccated spore-like form in dust; cattle sheep and goats are the main reservoirs for infections and ticks may be the natural primary reservoir host[17] . C . burnetii infection is asymptomatic in 50% to 60% of acute infections and its clinical polymorphism makes diagnosis difficult[17] . Abortions may occur in pregnant women and the chronic form which develops in about 5% of infections manifests as endocarditis , hepatitis and pneumonia . Diagnosis in rural communities in Afghanistan is particularly challenging because of poor access to competent and well-equipped health services and laboratories and competition from traditional healers in rural villages . A Disease Early Warning System ( DEWS ) for surveillance , investigation and reporting of infectious diseases in Afghanistan is operated by the surveillance department of the Afghanistan National Public Health Institute within the Ministry of Public Health . Q fever is not on the DEWS list of priority diseases and clinical syndromes but this study has greatly increased awareness of its importance among public health and veterinary authorities . Although C . burnetii infections in animals may result in abortion and shedding in birth products and milk , most infections are asymptomatic and differentiation from other causes of abortion which are known or suspected to be endemic in ruminants in Afghanistan is virtually impossible for reasons similar to those which prevail in the public health sector . Brucella seropositivity ( higher risk ) and Kuchi flocks ( lower risk ) were apparently associated with abortion in animals ( Table 3 ) but no associations were found for C . burnetii seropositivity and abortions . Both infections may cause abortions in animals [16 , 18] but it is likely that a high proportion of animals of breeding age would be immune to C . burnetii because of a high level of exposure before pregnancy unlike conditions for infection which would apply for Brucella . We consider the associations and the difference between Brucella and C . burnetii seropositivity to be interesting and worth reporting with the caution that the links with abortion should only be regarded as associations and not as evidence of causality because the timing of infection with regard to abortion was not possible to determine in a cross-sectional study . The KAP component of the study produced information about the prevalence of risky practices for infection within households and the analysis of the evidence of exposure to Brucella in householders or in their livestock identified high risk households as Kuchi with ≤4 rooms in their house and not owning land ( Table 5 ) . Activities which have been shown to be risky for zoonotic transmission of Brucella in other resource-limited countries include rural residence , contact with infected animals at parturition , consumption of unpasteurised milk and dairy products , high risk occupations such as veterinarians and slaughterhouse workers [19 , 20 , 21 , 22 , 23] and our study recorded disturbingly high prevalences of all of these activities among householders . The study produced useful information which can be incorporated into the design of control strategies and was a catalyst for a nation-wide serological survey of brucellosis in animals and the introduction of Rev1 vaccination of small ruminants by private veterinary units in high prevalence districts . Rev1 vaccination is the most important component of a B . melitensis control program but its success depends on adequate coverage , identification of vaccinates , timely and adequate supply of quality assured vaccine , awareness in rural communities , movement controls and surveillance; all conditions which need to be maintained for many years for sustainability . These conditions are extremely difficult to achieve in the uneasy and unsafe environment which prevails in Afghanistan with almost total reliance on often poorly coordinated donor support from many NGOs and international organisations such as the FAO and the World Health Organisation . Control of C . burnetii infections in humans and animals is difficult even in well-resourced countries [24 , 25] but the outlook for its control in Afghanistan is particularly bleak . The success which C . burnetii enjoys in the Afghanistan environment is testament to the husbandry and livestock management practices which favour its long term survival , exposure at a young age and highly efficient modes of transmission . Awareness campaigns aimed at educating householders and livestock owners about avoidance of risky practices offer some hope and benefits for brucellosis and echinococcosis but are likely to have little impact on C . burnetii . A high risk for Echinococcosis was identified by a high prevalence ( 62% ) of feeding guts from slaughtered sheep or goats to dogs . In summary the study alerted authorities to a hitherto unrecognised high prevalence of C . burnetii infections , acted as a catalyst for Rev1 vaccination of sheep and goats , demonstrated the benefits of a coordinated approach and fostered a better understanding of the nature of infection in different hosts and of the constraints for control faced by government services . A notable feature of the study was the enthusiasm and interest displayed by all of the participants throughout , from heads of government services to field personnel and villagers . The high level of interest displayed by the villagers was illustrated by the very high participation rate ( 93 . 6% ) of the randomly selected households which were invited to participate . Livestock owners regard zoonoses as adversities that affect their livestock and members of their households and do not partition them separately as medical or veterinary problems . Control programs need to take that perception into account and wherever possible avoid vesting ownership separately into veterinary or public health agencies . Education programs targeted to the large numbers of Community Health Centres ( 385 ) , Basic Health Centres ( 813 ) and Veterinary Field Units ( 1 , 200–1 , 500 ) in the Republic have the potential to create awareness about risky practices for zoonoses among livestock owners and householders at village level and thereby reduce the incidence of disease . | Our study alerted authorities to a hitherto unrecognised high prevalence of C . burnetii infections , acted as a catalyst for the introduction of a national vaccination programme for protection of sheep and goats from brucellosis using Rev1 vaccine , demonstrated the benefits of a coordinated approach and fostered a better understanding of the nature of infection in different hosts and of the constraints for control faced by government services . A notable feature of the study was the enthusiasm and interest displayed by all of the participants throughout , from heads of government services to field personnel and villagers . Livestock owners regard zoonoses as adversities that affect their livestock and members of their households and do not partition them separately as medical or veterinary problems . Control programs need to take that perception into account and wherever possible avoid vesting ownership separately into veterinary or public health agencies . | [
"Abstract",
"Introduction",
"Methods",
"Discussion"
] | [] | 2015 | Brucellosis and Coxiella burnetii Infection in Householders and Their Animals in Secure Villages in Herat Province, Afghanistan: A Cross-Sectional Study |
Most diversity in animals and plants results from the modification of already existing structures . Many organ systems , for example , are permanently modified during evolution to create developmental and morphological diversity , but little is known about the evolution of the underlying developmental mechanisms . The theory of developmental systems drift proposes that the development of conserved morphological structures can involve large-scale modifications in their regulatory mechanisms . We test this hypothesis by comparing vulva induction in two genetically tractable nematodes , Caenorhabditis elegans and Pristionchus pacificus . Previous work indicated that the vulva is induced by epidermal growth factor ( EGF ) /RAS and WNT signaling in Caenorhabditis and Pristionchus , respectively . Here , we show that the evolution of vulva induction involves major molecular alterations and that this shift of signaling pathways involves a novel wiring of WNT signaling and the acquisition of novel domains in otherwise conserved receptors in Pristionchus vulva induction . First , Ppa-LIN-17/Frizzled acts as an antagonist of WNT signaling and suppresses the ligand Ppa-EGL-20 by ligand sequestration . Second , Ppa-LIN-18/Ryk transmits WNT signaling and requires inhibitory SH3 domain binding motifs , unknown from Cel-LIN-18/Ryk . Third , Ppa-LIN-18/Ryk signaling involves Axin and β-catenin and Ppa-axl-1/Axin is epistatic to Ppa-lin-18/Ryk . These results confirm developmental system drift as an important theory for the evolution of organ systems and they highlight the significance of protein modularity in signal transduction and the dynamics of signaling networks .
Biological diversity in metazoans is basically generated by two principles , the evolution of novelty and the modification of already existing structures . Many case studies in recent years have shown that the evolution of novelty often depends on the co-option of developmental control genes in new cells and tissues [1] , [2] . However , most diversity in animals and plants is generated by the modification of already existing structures . For example , organ systems are stable over long evolutionary time periods , but they are permanently modified to create a diversity of form and structures [3] . Little is known about the evolution of the molecular mechanisms underlying the specification of such conserved morphological entities . The theory of developmental systems drift proposes that the development of conserved morphological structures can involve large-scale modifications in their regulatory mechanisms [4] , but few case studies have been initiated to test this hypothesis . Such studies require detailed functional comparisons between at least two species with the same body plan and they involve the generation of functional approaches in an evo-devo context , i . e . , the development of genetic and transgenesis tools in nonclassical model systems [5] . We have developed the nematode P . pacificus for such studies allowing detailed analysis of the molecular mechanisms involved in signal transduction . P . pacificus belongs to the family of the Diplogastridae , whereas C . elegans is a member of the Rhabditidae family [6] . While both groups belong to the same nematode clade , the comparison of more than a thousand orthologous genes suggests that the last common ancestor of P . pacificus and C . elegans existed 250–420 million y ago ( Figure 1G ) [7] . To study developmental systems drift we investigate the evolution of vulva induction between C . elegans and P . pacificus . Both organisms are amenable to forward and reverse genetics and transgenesis , allowing detailed functional and mechanistic studies . C . elegans vulva formation serves as one of the major model systems for elucidating the mechanisms of cell fate specification and cell-cell interactions [8] . Therefore , C . elegans vulva formation also represents a framework for functional studies in evo-devo [5] . Three of six equipotent C . elegans vulval precursor cells ( VPCs ) are selected by an inductive signal from the gonadal anchor cell ( AC ) ( Figure 1A and 1B ) [8] . These three cells , P ( 5–7 ) . p adopt a 2°–1°–2° fate pattern and form vulval tissue , whereas the three remaining VPCs , P ( 3 , 4 , 8 ) . p , adopt an epidermal , so-called 3° fate ( Figure 1B and 1C ) . A complex network of signaling processes determines vulval cell fates in C . elegans . The AC secretes an epidermal growth factor ( EGF ) -type ligand that is transmitted in the VPCs by RAS/MAPK signaling . A notch-type lateral signaling process acts downstream or in parallel to EGF/RAS signaling to specify vulval fates . Besides EGF and Notch signaling , Wnt signaling has been implicated in two other aspects of vulva determination . First , Wnt signaling is involved in the regulation of the Hox gene lin-39 , which specifies VPCs early in larval development ( Figure S1 ) . In lin-39/Hox or bar-1/β-catenin mutants , VPCs are not specified and fuse with the hypodermis , adopting the fate of the epidermal cells P ( 1 , 2 , 9–11 ) . p in the anterior and posterior body region ( Figure S1 ) [9] . Later in larval development , another Wnt pathway is involved in the proper execution of the 2° cell lineage ( Figure 1C ) [10] . While P5 . p and P7 . p both adopt a 2° cell fate , a lineage reversal of P7 . p is required for proper vulva morphogenesis . An atypical Wnt signaling pathway , involving the Cel-lin-17/Frizzled , but also the Ryk-type receptor Cel-lin-18 , was shown to regulate P7 . p polarity [10]–[12] . In Cel-lin-17 and Cel-lin-18 mutants , the P7 . p lineage reversal is abolished and P7 . p forms a second vulva-like protrusion in the posterior part of the animal , resulting in a Bivulva phenotype . Thus , Wnt signaling has a key role in the regulation of VPC competence and P7 . p polarity in C . elegans , whereas its role in vulva induction is still debated [11] , [13] . In P . pacificus , the vulva is formed by the same cells as in C . elegans , P ( 5–7 ) . p , which also adopt a 2°–1°–2° fate pattern ( Figure 1D–1F ) [14] . However , vulva induction in P . pacificus relies on Wnt signaling and mutants in the Wnt pathway have induction-vulvaless phenotypes , unlike Wnt pathway mutants in C . elegans ( Figure 1E ) [15] . A Ppa-bar-1/β–catenin deletion mutant is completely induction-vulvaless . In such animals , VPCs do not divide and remain epidermal ( 3° ) , in a manner similar to gonad-ablated animals ( Figures S1 , S2B , and S2C; Table 1A–1C ) . Several WNT ligands and receptors act redundantly in P . pacificus vulva induction: Ppa-mom-2/Wnt; Ppa-lin-18/Ryk double mutants and Ppa-mom-2; Ppa-egl-20; Ppa-lin-18 triple mutants are strongly vulvaless , while single gene mutations are phenotypically normal ( Figure S2D–S2H; Table 1D–1H ) [15] . Ppa-mom-2 and Ppa-lin-44 , another Wnt ligand , are expressed in the AC and the somatic gonad , respectively , representing the gonadal branch of inductive Wnt signal [15] . In contrast , Ppa-egl-20 is expressed in the tail in a similar manner as Cel-egl-20 [15] , [16] , representing the first signaling ligand involved in vulva induction that is expressed outside of the gonad [15] . Wnt signal transduction during P . pacificus vulva induction requires an unusual wiring . In contrast to Ppa-bar-1 , Ppa-egl-20 , Ppa-mom-2 , and Ppa-lin-18 , which promote vulva induction as indicated by the vulvaless phenotype associated with single or double mutants in these genes [15] , Ppa-lin-17/Frizzled has a negative role in vulva formation and mutations in Ppa-lin-17 are multivulva [17] . After ablation of the somatic gonad at hatching , P ( 5–7 ) . p form vulva-like structures in Ppa-lin-17 ( tu33 ) . Also , the Ppa-lin-17 phenotype is suppressed by Ppa-egl-20 , suggesting a novel ligand-receptor interaction ( Figure S2I and S2J; Table 1I and 1J ) [15] . A similar interaction between lin-17 and egl-20 has subsequently been observed during P7 . p polarity specification in C . elegans [12] . However , the molecular mechanisms of the lin-17–egl-20 interaction have not been investigated . Taken together , three hallmarks characterize P . pacificus vulva induction when compared to C . elegans . First , Wnt signaling rather than EGF/RAS signaling induces the formation of this homologous morphological structure . Second , two signaling centers from the somatic gonad and the posterior body region are redundantly involved in vulva induction relying on distinct Wnt ligands . Third , signaling from the posterior body region involves a novel wiring between the Frizzled receptor Ppa-lin-17 and the Wnt ligand Ppa-egl-20 . Here , we elucidate the molecular mechanism of Wnt signaling from the posterior to study developmental systems drift . By using genetic , molecular , and transgenesis studies we show that Ppa-LIN-17 acts by ligand sequestration . The Ryk-type receptor Ppa-LIN-18 transmits Wnt signaling and requires inhibitory SH3 binding motifs , unknown from Cel-LIN-18 . Ppa-LIN-18/Ryk signaling involves the single Axin-like gene in P . pacificus and Ppa-axl-1 mutants represent the strongest multivulva phenotype observed in P . pacificus to date . These studies indicate the acquisition of individual domains in otherwise conserved Frizzled- and Ryk-type receptors during vulva induction in P . pacificus and they highlight the strong dynamics and modularity of signaling networks underlying the evolution of conserved morphological structures , thereby supporting the theory of developmental systems drift .
The key difference between P . pacificus Wnt signaling during vulva induction and other types of Wnt signal transduction is the antagonistic function of the Frizzled-type receptor Ppa-lin-17 [15] . While a mutation in the β-catenin–like gene Ppa-bar-1 is completely vulvaless , Ppa-lin-17/Fz null alleles are multivulva as indicated by the ectopic vulva formation in Ppa-hairy ( tu97 ) ; Ppa-lin-17/Fz ( tu108 ) double mutants ( Table 1P ) [17] . Also , Ppa-egl-20/Wnt suppresses the multivulva phenotype of Ppa-lin-17/Fz , highlighting the unusual wiring of Wnt signaling during P . pacificus vulva induction [15] . To elucidate the molecular mechanism of Ppa-LIN-17/Fz function , we used a genetic approach and screened for mutations in genes that would result in a multivulva phenotype similar to Ppa-lin-17/Fz . We performed screens in three different genetic backgrounds , wild type , Ppa-hairy , and Ppa-ced-3 mutants . In Ppa-hairy and Ppa-ced-3 mutants , the survival of P ( 3 , 4 ) . p results in two epidermal cells that are competent to form vulval tissue ( Table 1K and 1L; Figure S2K and S2L ) [18] , [19] . From these screens we isolated three mutants that fall into two complementation groups . Complementation group 1 has only one allele ( tu383 ) and complementation group 2 consists of the alleles tu98 and tu329 . We first describe the characterization of tu383 . tu383 has a novel vulva phenotype and was isolated in the Ppa-hairy background . In Ppa-hairy ( tu97 ) ; tu383 double mutants , P ( 3–8 ) . p have vulva cell fates and P ( 3 , 4 , 8 ) . p adopt a 2° fate in most mutant animals ( Table 1M ) . After ablation of the somatic gonad at hatching in Ppa-hairy ( tu97 ) ; tu383 double mutants , P ( 3–8 ) . p still adopt vulval cell fates ( Table 1N ) . The tu383 single mutant retains vulva differentiation in P8 . p , whereas P ( 3 , 4 ) . p die of programmed cell death ( Figure 2A and 2B; Table 1O ) . These results indicate that tu383 fulfills both criteria of a multivulva phenotype , ectopic vulva differentiation and gonad-independent vulva formation . We mapped tu383 to a small interval of Chromosome V , which includes Ppa-lin-17/Fz ( Figure 2C ) . No other candidate genes related to Wnt signaling are known in this genetic interval . Therefore , we started DNA-mediated transformation rescue experiments by injecting a 16-kb genomic DNA construct that contains the wild-type copy of Ppa-lin-17/Fz . More than 50% of transgenic animals showed rescue of the tu383 phenotype ( Figure 2D ) . We sequenced the Ppa-lin-17/Fz gene from tu383 mutant animals and found an A to T transition changing the natural stop codon TAA to TAT ( Figure 2E ) . This results in a 17 amino acid extension of the Ppa-LIN-17/Fz protein . Sequencing of the cDNA by reverse transcriptase ( RT ) -PCR confirmed this mutation . We conclude that tu383 represents an unusual allele of Ppa-lin-17/Fz . Ppa-lin-17/Fz reduction-of-function alleles and Ppa-lin-17/Fz ( tu383 ) have distinct vulva phenotypes , affecting P7 . p polarity and P8 . p differentiation , respectively . P7 . p polarity is strongly affected in all Ppa-lin-17/Fz reduction-of-function alleles , but not in Ppa-lin-17/Fz ( tu383 ) ( Table 1I and 1O ) [17] . By contrast , vulva differentiation in P8 . p is only observed in Ppa-lin-17/Fz ( tu383 ) ( Figure 2B ) . Consistent with these observations , trans-heterozygous Ppa-lin-17/Fz ( tu33/tu383 ) animals were phenotypically wild type ( Table 1Q ) . Similarly , 56 of 58 Ppa-lin-17/Fz ( tu383/+ ) animals are phenotypically normal , indicating that the vulva phenotype of Ppa-lin-17/Fz ( tu383 ) mutants does not have a strong dominant effect ( Table 1R ) . In addition , vulva differentiation of P8 . p in Ppa-lin-17/Fz ( tu383 ) is not suppressed by Ppa-egl-20/Wnt ( Table 1S ) . These results indicate that a 17 amino acid extension at the C terminus of Ppa-LIN-17/Fz causes a neomorphic phenotype with vulval patterning defects that are distinct from those of Ppa-lin-17/Fz reduction-of-function alleles . Next , we wanted to know if the extension of Ppa-LIN-17/Fz in the tu383 allele changes the biochemical properties of the protein . Interestingly , the 17 amino acid extension contains a potential SH3 domain-binding motif ( SDBM ) of the sequence PGIP574-577 followed by a potential phosphorylation site THS578-580 ( Figure 2E ) . Another potential SDBM exists in the natural C terminus of Ppa-LIN-17/Fz at amino acid position 531–534 , which is not followed by a phosphorylation site ( Figure 2E ) . To test the functional importance of individual amino acids in the C terminus of Ppa-LIN-17/Fz ( tu383 ) , we developed a transgenic assay with a readout for ectopic vulva differentiation . We used a Ppa-ced-3 ( tu54 ) mutant background ( Figure S2L ) and measured ectopic vulva differentiation in both , P ( 1–4 ) . p and P8 . p ( Figure 2F ) . A transgenic line carrying a full length copy of Ppa-lin-17/Fz ( tu383 ) showed nearly 100% vulva differentiation in P ( 1–4 ) . p and P8 . p ( Figure 2F , dark blue bars ) . In contrast , the transgenic line containing a wild-type copy of Ppa-lin-17/Fz did not show ectopic vulva differentiation in these cells ( Figure 2F , light blue bars ) . Using this assay system , we tested the function of the potential SDBM sites in the intracellular domain of Ppa-LIN-17/Fz ( tu383 ) . First , we generated a transgene with a PGIP574-577AAAA mutation . Transgenic animals showed a reduction in P ( 1–4 ) . p vulva differentiation from 100% to 15% and vulva differentiation in P8 . p is completely abolished ( Figure 2F , green bars ) . Second , we generated a construct that changed THS578–580 into AHA . Animals carrying this transgene showed vulva differentiation below 10% in P ( 1–4 ) . p and P8 . p ( Figure 2F , red bars ) . In contrast , transgenic animals with a mutation of PGIP531–534 to AAAA showed 100% vulval differentiation in P ( 1–4 ) . p and 35% in P8 . p ( Figure 2F , orange bars ) . These experiments suggest that the multivulva phenotype of Ppa-lin-17/Fz ( tu383 ) depends on the PGIPTHS motif in the 17 amino acid extension of the Ppa-LIN-17/Fz receptor . We speculate that this protein extension provides an ectopic SDBM that interferes with other protein interactions in VPCs . One potential mechanism would be that PGIPTHS574-580 mimics a peptide of another protein in the Wnt signaling network , i . e . , the receptor for Ppa-egl-20/Wnt , and competes for binding to a regulatory protein that acts downstream in the Wnt pathway . Next , we searched for potential SDBMs in other P . pacificus Frizzled-type receptors and Ppa-LIN-18/Ryk . While none of the other P . pacificus Frizzled-type receptors contains an SDBM , there are three such motifs in the intracellular domain of Ppa-LIN-18 ( Figure 3A ) , the Ryk-type receptor that is known to be involved in vulva induction ( Figure S2F–S2H; Table 1F–1H ) [15] . Sequence comparison between Ppa-LIN-18/Ryk and Cel-LIN-18/ryk reveals high sequence similarity in both , the extracellular and intracellular parts of the LIN-18/Ryk proteins ( Figure 3A , right ) . The intracellular domain consists of a kinase domain , which in C . elegans is not required for P7 . p polarity [10] . While the P . pacificus and C . elegans LIN-18/Ryk proteins show a high overall sequence similarity , Ppa-LIN-18/Ryk contains three potential SDBMs , whereas only one sequence with weak similarity to this motif is found in Cel-LIN-18/Ryk ( Figure 3A , right ) . To study the function of individual domains of Ppa-LIN-18/Ryk , we developed a transgenic assay with a functional reporter system . In Ppa-lin-18/Ryk; Ppa-mom-2/Wnt double mutants only 2% of the worms form a normal vulva with a 2°–1°–2° pattern , whereas 39% of the animals are vulvaless ( Figures 3B and S2G; Table 1G ) . Transgenic animals that carry a full-length wild-type copy of Ppa-lin-18/Ryk can partially rescue the Ppa-lin-18/Ryk; Ppa-mom-2/Wnt double mutant phenotype . In two independent transgenic lines , 25% and 45% of animals form a normal vulva ( Figures 3C , 3D , and S3 ) . Similarly , a full-length Ppa-LIN-18/Ryk fused to an HA tag at the carboxyl terminus showed 29% rescue of the double mutant phenotype ( Figures 3D and S3 ) . Next , we generated constructs containing only the extracellular domain and the transmembrane domain ( “ECD-TM” construct ) , or the signal peptide , the transmembrane domain and the intracellular domain ( “Signal peptide-TM-C-term” construct ) . None of these transgenic lines rescued the Ppa-lin-18/Ryk; Ppa-mom-2/Wnt double mutant phenotype , suggesting that in contrast to C . elegans , the extracellular domain of Ppa-LIN-18/Ryk on its own does not have rescuing activity ( Figures 3D and S3 ) . These results provide the first evidence that Ppa-LIN-18/Ryk requires C-terminal domain in vulva induction and has a distinct function than in C . elegans . To determine if the three potential SDBMs in the intracellular domain of Ppa-LIN-18/Ryk are involved in the regulation of Wnt signaling , we mutated the SDBM encoding fragments of Ppa-lin-18/Ryk and tested for rescuing activity of the resulting transgenes . Transgenic lines carrying mutations of the first or third SDBM ( “ECD-TM-C-term 1st SDBM265-268AAAA” and “ECD-TM-C-term 3rd SDBM552-559AAFDLAA” ) showed a strong increase in rescuing activity ( Figures 3D and S3 ) . Specifically , four of five transgenic lines with mutations at SDBM265-268 showed more than 64% rescue . Similarly , two out of three lines with mutations at SDBM552-559 showed more than 50% rescue ( Figures 3D and S3 ) . In contrast , mutations of the second site , SDBM368-371 , resulted in normal vulval patterning in only 15% of transgenic animals ( Figures 3D and S3 ) . These results suggest that SDBM sites 1 and 3 of Ppa-LIN-18/Ryk have an inhibitory function , which might be involved in binding to an inhibitor that prevents signaling in the absence of Wnt ligands . We speculate that the removal of SDBMs 1 or 3 in Ppa-LIN-18/Ryk results in an enhancement of rescuing activity , because in the absence of these SDBMs the inhibitor can no longer bind to Ppa-LIN-18/Ryk . Under the assumption that the ectopic SDBM site PGIPTHS574-580 of Ppa-LIN-17/Fz ( tu383 ) mimics the SDBMs of Ppa-LIN-18/Ryk , Ppa-LIN-17/Fz ( tu383 ) attracts the inhibitor by competitive binding from Ppa-LIN-18/Ryk , thereby causing ligand independent signal transduction . The experiments described above indicate that the SDBMs are important for Ppa-LIN-18/Ryk function and they are consistent with the hypothesis that Ppa-LIN-18/Ryk and Ppa-LIN-17/Fz ( tu383 ) competitively bind inhibitory SH3 domain proteins via their SDBMs . To study if Ppa-LIN-18/Ryk and Ppa-LIN-17/Fz ( tu383 ) have similar binding properties , we replaced the ectopic SDBM PGIPTHS574-580 of Ppa-LIN-17/Fz ( tu383 ) with the individual SDBMs of Ppa-LIN-18/Ryk ( Figure 4A ) and tested the resulting constructs in the transgenic assays for ectopic vulva differentiation in P ( 1–4 ) . p and P8 . p using the Ppa-ced-3 ( tu54 ) mutant background ( Figure 4 ) . Transgenic animals carrying construct in which the ectopic SDBM of Ppa-LIN-17/Fz ( tu383 ) was replaced by the first or third SDBM of Ppa-LIN-18/Ryk , showed strong vulva induction in P ( 1–4 ) . p and P8 . p ( Figure 4B and 4C ) . In contrast , transgenic animals carrying a Ppa-LIN-17/Fz ( tu383 ) substitution with the second SDBM of Ppa-LIN-18/Ryk , showed only 25% vulva induction in P ( 1–4 ) . p , which is similar to the result of transgenes containing a Ppa-LIN-17/Fz ( tu383 ) PGIP574-577AAAA mutation ( Figure 4B and 4C ) . These results indicate that the ectopic SDBM of Ppa-LIN-17/Fz ( tu383 ) mimics the first and third SDBM sites of Ppa-LIN-18/Ryk , suggesting that Ppa-LIN-17/Fz ( tu383 ) and Ppa-LIN-18/Ryk competitively bind to a common SH3 domain containing inhibitory protein . If Ppa-LIN-18/Ryk and Ppa-LIN-17/Fz ( tu383 ) compete for binding , a Ppa-lin-18/Ryk ( tu359 ) mutant should be epistatic to Ppa-lin-17/Fz ( tu383 ) . Indeed , Ppa-lin-18/Ryk ( tu359 ) suppresses the multivulva phenotype of Ppa-lin-17/Fz ( tu383 ) and 33 of 45 ( 73% ) Ppa-lin-18/Ryk ( tu359 ) ; Ppa-lin-17/Fz ( tu383 ) double mutants are phenotypically wild type ( Table 1T ) , indicating that Ppa-lin-17/Fz ( tu383 ) requires functional Ppa-LIN-18/Ryk for signaling . Taken together , these experiments suggest that the SDBMs in Ppa-LIN-18/Ryk and Ppa-LIN-17/Fz ( tu383 ) have similar SH3 binding properties and that Ppa-lin-18/Ryk ( tu359 ) suppresses Ppa-lin-17/Fz ( tu383 ) . Next , we characterized the remaining multivulva locus ( complementation group 2 , see above ) , which represents the strongest multivulva phenotype obtained in P . pacificus to date . Two alleles , tu98 and tu329 , have been isolated from large-scale mutagenesis experiments . tu98 and tu329 mutant animals have two ventral epidermal cell lineage defects , both of which indicate that this gene represents a negative regulator of vulva formation . First , P8 . p adopts a vulva fate in 85% and 100% in tu98 and tu329 mutant animals , respectively ( Figure 5A and 5B ) . Second , P ( 9–11 ) . p survive and adopt vulval cell fate in most mutant animals ( Figure 5A and 5B ) . Furthermore , P ( 5–11 ) . p can adopt vulval fates after ablation of the somatic gonad at hatching , indicating that the multivulva phenotype is gonad independent ( Figure 5A ) . Similarly , in a Ppa-ced-3 ( tu54 ) ; tu98 double mutant all epidermal cells from P1 . p to P11 . p can undergo vulva differentiation ( Figure 5A and 5C ) . We mapped the locus to an interval of several hundred kb between the molecular markers S240 and S327 in the center of Chromosome V ( Figure 5D ) . One potential candidate in that region is a gene with high sequence similarity to axin , a known negative regulator of Wnt signaling in C . elegans [20] , [21] . While C . elegans contains two axin-like genes , Cel-pry-1 and Cel-axl-1 , the current draft of the P . pacificus genome has only one axin-like gene , which is called Ppa-axl-1 because it shows highest sequence similarity to Cel-axl-1 ( Figure 5E ) . Overall , sequence similarity is restricted to the known functional domains , the “Regulator of G Protein Signaling” ( RGS ) and the “Dishevelled-Axin” ( DIX ) domain ( Figure 5E ) . To determine if this complementation group corresponds to Ppa-axl-1 , we performed rescue experiments . A 7-kb genomic construct containing the complete coding region as well as 2 . 8 kb of upstream sequence fully rescued the phenotype of tu98 . When we sequenced the Ppa-axl-1 gene in the two alleles tu98 and tu329 , we found mutations in the RGS domain in both of them . tu329 contains a mutation resulting in an amino acid alteration from Leu to Phe at position 115 and the trimethylpsoralen-ultraviolet–induced allele tu98 contains two mutations in the RGS domain resulting in amino acid alterations at position 47 ( Ile to Asn ) and position 50 ( Phe to Tyr ) ( Figure 5E ) . All mutations were confirmed by sequencing cDNA fragments . Together , these results indicate that loss-of-function alleles in Ppa-axl-1 result in the strongest known multivulva phenotype in P . pacificus . Next , we studied the genetic interactions between Ppa-lin-18/Ryk , Ppa-axl-1 , and Ppa-bar-1/β-catenin . First , we generated a Ppa-axl-1 ( tu98 ) ; Ppa-mom-2/Wnt ( tu263 ) ; Ppa-lin-18/Ryk ( tu359 ) triple mutant to see if Ppa-axl-1 ( tu98 ) can suppress the vulvaless phenotype of Ppa-mom-2/Wnt ( tu263 ) ; Ppa-lin-18/Ryk ( tu359 ) double mutants . Indeed , a Ppa-axl-1 ( tu98 ) ; Ppa-mom-2/Wnt ( tu263 ) ; Ppa-lin-18/Ryk ( tu359 ) triple mutant is multivulva indicating that Ppa-axl-1 is epistatic to Ppa-lin-18/Ryk ( Table 1U ) . Next , we determined the role of Ppa-bar-1/ß-catenin in relation to the other factors . Ppa-axl-1 ( tu98 ) ; Ppa-bar-1/β-catenin ( tu362 ) and Ppa-bar-1/ß-catenin ( tu362 ) ; Ppa-lin-17/Fz ( tu383 ) double mutants are completely vulvaless , indicating that the unusual Wnt network acting during P . pacificus vulva induction is Ppa-BAR-1/β-catenin dependent , with Ppa-BAR-1/ß-catenin representing the most downstream component identified to date ( Table 1V and 1W ) . The experiments described above indicate an antagonism of Ppa-LIN-17/Fz and Ppa-LIN-18/Ryk during Ppa-EGL-20/Wnt signaling from the posterior body region . Two testable predictions derive from these observations . First , the antagonistic role and the multivulva phenotype of Ppa-lin-17/Fz mutants suggest that Ppa-LIN-17/Fz works by ligand sequestration of Ppa-EGL-20/Wnt . According to this model , Ppa-LIN-17/Fz balances the amount of Ppa-EGL-20/Wnt from the posterior body region . In a Ppa-lin-17/Fz null mutant ectopic Ppa-EGL-20/Wnt signal will reach the VPCs causing vulva formation in the absence of the signal from the second signaling center , the somatic gonad ( Figure 6D ) . Second , Ppa-lin-17/Fz should be coexpressed with Ppa-egl-20/Wnt in the posterior body region . To test the ligand sequestration model ( Figure 6D ) , we expressed Ppa-egl-20/Wnt under two constitutive promoters and determined if Ppa-egl-20/Wnt overexpression was sufficient to cause ectopic vulva formation . We expressed Ppa-egl-20/Wnt under the control of the Drosophila heat shock promoter Dm-hsp-70 and the Ppa-mig-1/Fz promoter , respectively , and studied the effect of Ppa-EGL-20/Wnt overexpression in Ppa-ced-3; Ppa-egl-20/Wnt double mutant animals . The Ppa-egl-20/Wnt background eliminates endogenous Wnt activity and the cell-death defective Ppa-ced-3 background results in the survival of P ( 1–4 , 9–11 ) . p ( Figure S2L ) . We analyzed ectopic vulva differentiation in P ( 1–4 ) . p , the cells furthest away from the endogenous source of Ppa-EGL-20/Wnt signaling in the posterior body region . In Dm-hsp-70::Ppa-egl-20/Wnt transgenic animals , vulva differentiation in P ( 1–4 ) . p was on average observed in 25% of heat shocked animals , whereas Ppa-mig-1::Ppa-egl-20/Wnt transgenic animals showed 12% vulva differentiation in these cells ( Table 2 ) . In contrast , nontransgenic Ppa-ced-3; Ppa-egl-20/Wnt mutant animals have no ectopic vulva differentiation ( Table 2 ) . Thus , Ppa-egl-20/Wnt is sufficient for vulva induction , supporting the ligand sequestration model for Ppa-LIN-17/Fz ( Figure 6D ) . To study the expression pattern of Ppa-lin-17/Fz we generated a Ppa-lin-17/Fz::RFP ( NLS ) construct containing 1 . 5-kb upstream sequence of Ppa-lin-17/Fz . We found Ppa-lin-17/Fz expression in the posterior body region surrounding the rectal area in all larval stages ( Figure 6A–6C ) . Coexpression of Ppa-lin-17/Fz::RFP with Ppa-egl-20/Wnt::GFP indicates that the two genes have an overlapping expression pattern during larval development that is , however , not completely identical ( Figure 6A–6C ) . Low level Ppa-lin-17/Fz::RFP expression is also seen in more anterior regions including the VPCs during the J3 larval stage ( unpublished data ) . All described expression patterns were highly reproducible and were observed in several independent transgenic lines . This pattern is consistent with ligand sequestration and the antagonistic interaction of Ppa-EGL-20/Wnt and Ppa-LIN-17/Fz ( Figure 6D ) . However , given the Ppa-lin-17/Fz expression in the posterior body region and the VPCs , we cannot distinguish if ligand sequestration occurs in the posterior body region , the central body region , or both . To address this question , we expressed Ppa-lin-17/Fz under the Ppa-egl-20/Wnt promoter . A Ppa-egl-20/Wnt::Ppa-lin-17/Fz transgene that is only expressed in the posterior region showed rescue of the Ppa-lin-17/Fz ( tu108 ) multivulva phenotype in a Ppa-ced-3 ( tu104 ) mutant background ( Figure S4 ) . These experiments suggest that ligand sequestration is largely restricted to the posterior tail region .
Transcription factors and signaling pathways are highly conserved throughout the animal kingdom [22] , [23] . Nonetheless , morphological and developmental structures show a near endless diversity . How this diversity is generated in the context of conserved developmental control genes remains largely elusive . We try to tackle this problem by studying the evolution of the developmental mechanisms underlying the specification of a conserved developmental and morphological structure , the nematode vulva . While previous studies clearly indicated that the vulva is a conserved organ that is formed from homologous precursor cells , little was known about the exact molecular mechanisms at work [24] . Here , we determined the molecular mechanisms of Wnt signaling from the posterior body region during P . pacificus vulva induction and demonstrate the importance of modularity and dynamics of signaling networks for the evolution of developmental mechanisms . We have obtained evidence for a novel wiring in a Wnt signaling network with a ligand sequestration function of LIN-17/Fz and the coupling of the LIN-18/Ryk receptor to Axin and BAR-1/β-catenin regulation . Four major conclusions can be drawn from these results . First , differences in vulva induction between C . elegans and P . pacificus are not restricted to the differential use of EGF/RAS versus Wnt signaling . The shift of signaling pathways involves a novel wiring of Wnt signaling . Second , signaling pathways that are conserved throughout the animal kingdom are highly dynamic and alterations of individual protein interactions can be under strong influence of small protein domains . Therefore , the presence or absence of SDBMs in nematode LIN-18/Ryk receptors highlights the significance of protein modularity for evolution . Also , evolutionary conserved proteins differ between species in that LIN-17/Fz acts as antagonist in P . pacificus vulva induction but as agonist in most Wnt signaling pathways . This supports a third conclusion , that mutational changes in protein-coding regions of core developmental control genes do occur and are of evolutionary importance . The significance of mutations in coding regions that change the structure of proteins versus mutations in cis-regulatory elements that alter the expression pattern of a given gene has been the subject of a recent debate in evo-devo [25] , [26] . Several case studies provide evidence for both types of mutation patterns , but these studies centered on adaptive changes between closely related species [27] , [28] . The comparison of vulva induction between C . elegans and P . pacificus provides an example for the evolutionary mutation spectrum of core developmental pathways . Because both nematodes are amenable to forward and reverse genetics , functional and mechanistic studies are possible that made it feasible to identify functional changes in the coding region of developmental control genes [5] . Our study therefore , provides support for the claim that the discussed prevalence of mutations in cis-regulatory elements might result from an ascertainment bias [27] . Finally , the comparison of Wnt signaling in animal development reveals a diversity of molecular mechanisms that is to a certain extent independent of phylogeny . For example , the coupling of the LIN-18/Ryk receptor to β-catenin regulation in P . pacificus is more similar to what has been described in mammals than to C . elegans or Drosophila Ryk function [29]–[31] . Similar findings were made previously when studying the role of GROUCHO proteins and their interaction with basic helix-loop-helix transcription factors in P . pacificus and C . elegans [19] . In this study we report what to our knowledge are three novelties of Wnt signaling in nematode development . First , we have discovered a new mechanism for LIN-17/Fz function in Wnt signaling . We hypothesize that Ppa-LIN-17/Fz balances the amount of Ppa-EGL-20/Wnt leading to ligand sequestration ( Figures 6D and 7A ) . In a Ppa-lin-17/Fz mutant , Ppa-EGL-20/Wnt is not properly sequestered resulting in gonad-independent vulva differentiation and a multivulva phenotype ( Figure 6D ) . Ligand sequestration is most likely restricted to the posterior body region close to the source of Ppa-EGL-20/Wnt ligand , as suggested by the observed rescuing capability of a Ppa-egl-20/Wnt-driven Ppa-lin-17/Fz transgene in a Ppa-lin-17 ( tu108 ) ; Ppa-ced-3 ( tu104 ) background ( Figure S4 ) . In C . elegans , a related regulation of EGL-20/Wnt activity was described by the Ror receptor tyrosine kinase CAM-1 [32] . Cel-EGL-20/Wnt regulates the anterior migration of the hermaphrodite-specific neuron ( HSN ) and the right Q neuroblast . Overexpression of Cel-egl-20/Wnt , as well as cam-1 mutants result in too far anterior migration , indicating that Cel-CAM-1 negatively regulates Cel-EGL-20/Wnt . In Drosophila , the Frizzled-type receptor Dfz2 has also been reported to antagonize Wnt signaling [33] . Second , the analysis of the neomorphic Ppa-lin-17/Fz ( tu383 ) allele provides insight into a novel aspect of the molecular mechanism of Wnt signaling . The mutation of the natural stop codon and the resulting 17 amino acid extension in Ppa-LIN-17/Fz ( tu383 ) led to the identification of a potential SDBM and three similar SDBMs in Ppa-LIN-18/Ryk . We propose that Ppa-EGL-20/Wnt signaling is transduced by Ppa-LIN-18/Ryk and the SDBM sites in Ppa-LIN-18/Ryk act as negative regulators ( Figure 7A ) . Through the binding of an inhibitor or an inhibitory complex , these SDBM sites prevent ligand-independent Wnt signal transduction . In the neomorphic Ppa-lin-17/Fz ( tu383 ) allele , the inhibitor is attracted by the ectopic SDBM site in Ppa-LIN-17/Fz ( tu383 ) and the release of the inhibitor from Ppa-LIN-18/Ryk results in ligand-independent vulva differentiation and a multivulva phenotype ( Figure 7B ) . This model is supported by the fact that Ppa-lin-18/Ryk ( tu359 ) suppresses the multivulva phenotype of Ppa-lin-17/Fz ( tu383 ) . However , the effect of tu383 and its 17 amino acid extension is specific to Ppa-lin-17/Fz as indicated by the finding that the fusion of the ectopic SDBM site to another Fz-type receptor , Ppa-MIG-1/Fz , does not result in a similar multivulva phenotype as Ppa-lin-17/Fz ( tu383 ) ( unpublished data ) . To our knowledge , this study provides the first evidence for the involvement of SDBMs in Wnt signaling . We speculate that the inhibitor or inhibitory complex in P . pacificus Wnt signaling involves SH3 domain proteins , the molecular nature of which awaits future analysis . There are 38 SH3 domain proteins in the P . pacificus genome including representatives of all major genes known from other model organisms ( Table S1 ) [34] . Given the fact that the mutagenesis screens for vulva defective mutants in P . pacificus have been saturated and no uncloned multivulva locus remains , it is well possible that redundancy masks the function of individual genes in this process . It is interesting to note that in contrast to Wnt signaling , SH3 domain proteins are known to play a key role in EGF/RAS signaling . In C . elegans vulva induction , the GRB2 homolog sem-5 is a positive regulator of vulva formation [35] , whereas sli-1 , a member of the Cbl ubiquitin ligase family , is a negative regulator of LET-23/EGFR [36] . One intriguing possibility therefore would be that Ppa-LIN-18/Ryk interacts with any of these gene products . Third , the functional characterization of Ppa-axl-1 shows that the novel wiring of Ppa-EGL-20/Wnt , Ppa-LIN-17/Fz , and Ppa-LIN-18/Ryk requires Axin and acts through the β-catenin–type protein Ppa-BAR-1 . No in vivo genetic evidence has been described previously between LIN-18/Ryk , Axin , and β-catenin . Data from C . elegans and mammalian cell culture provide evidence that Ryk is required for TCF-driven transcription [37] , [38] . Similarly , a recent study described the isolation of a pry-1/Axin mutation in C . briggsae [39] . While Cbr-pry-1/Axin mutants provide evidence for the conservation and diversification of Wnt signaling in Caenorhabditis , Cbr-pry-1/Axin was shown to function upstream of Cbr-bar-1|β-catenin and Cbr-pop-1 [39] . Taken together , the comparison of Wnt pathway components and their functions between P . pacificus and C . elegans reveals an enormous diversity in the regulatory linkage of Wnt signal transduction and in the composition of the regulatory networks that control conserved morphological entities . Therefore , the evolution of vulva development provides a prime example for developmental systems drift [4] . The results described in this study highlight the strong dynamics and modularity of signaling networks in animal development and can therefore be used to extend the concept of developmental systems drift further . The evolution of individual protein domains , and not necessarily complete genes or proteins , can result in novel protein interactions and thus , new regulatory linkages . These findings result in the intriguing idea that protein domains might be the prime subjects of natural selection . This idea would be consistent with the general notion that developmental control genes are highly conserved in sequence throughout the animal kingdom . This conservation however , is often restricted to the DNA binding domains of transcription factors or already well-characterized domains of intracellular proteins . Small peptides , like the SDBM described here , can and most likely , do evolve relatively fast . However , functionally important peptides might often remain unnoticed , as they cannot be simply deduced from computational analysis . Only a detailed genetic and biochemical characterization can identify the functional significance of such domains . The independent evolution of small protein domains in otherwise conserved proteins increases the “evolutionary freedom” of signaling pathways and developmental networks . The introduction of a new interaction partner into an already existing pathway , but also the cross-connectivity of signaling pathways in developmental networks can be facilitated by the evolution of novel protein domains . Thus , the acquisition of novel protein domains might allow the de novo evolution of functional modules in the context of conserved regulatory control genes .
All worms were kept on nematode growth medium ( NGM ) agar plates with the Escherichia coli strain OP50 as food [14] . P . pacificus PS312 ( California ) was used for all genetic analysis . For cell ablation experiments , animals were picked into M9 buffer placed on a pad of 5% agar in water containing 10 mM sodium azide as anesthetic . All ablation experiments were carried out 0–1 h after hatching ( 20°C ) . The following mutants were used: Ppa-bar-1/β-catenin ( tu362 ) , Ppa-egl-20/Wnt ( tu382 ) , Ppa-mom-2/Wnt ( tu363 ) , Ppa-lin-18/Ryk ( tu359 ) , Ppa-mom-2/Wnt ( tu363 ) ; Ppa-lin-18/Ryk ( tu359 ) , Ppa-mom-2/Wnt ( tu363 ) ; Ppa-lin-18/Ryk ( tu359 ) ; Ppa-egl-20/Wnt ( tu364 ) , Ppa-lin-17/Fz ( tu33 ) ; Ppa-egl-20/Wnt ( tu382 ) ( all originally described in [15] ) , Ppa-hairy ( tu97 ) [19] , Ppa-lin-17/Fz ( tu33 ) [17] . The following single and double mutants are newly described in this study: Ppa-lin-17/Fz ( tu383 ) , Ppa-hairy ( tu97 ) ; Ppa-lin-17/Fz ( tu383 ) , Ppa-lin-17/Fz ( tu383/tu33 ) , Ppa-lin-17/Fz ( tu383 ) ; Ppa-egl-20/Wnt ( tu382 ) , Ppa-lin-17/Fz ( tu383 ) ; Ppa-lin-18/Ryk ( tu359 ) , Ppa-axl-1 ( tu98 ) ; Ppa-mom-2/Wnt ( tu363 ) ; Ppa-lin-18/Ryk ( tu359 ) , Ppa-lin-17/Fz ( tu383 ) ; Ppa-bar-1/β-catenin ( tu362 ) , Ppa-axl-1 ( tu98 ) ; Ppa-bar-1/β-catenin ( tu362 ) . Mixed stage animals were washed off the plates in M9 buffer , after which ethyl methanesulphonate ( EMS ) was added to a final concentration of 50 mM for 4 h at 20°C . After washing in M9 five times , worms were spotted onto the surface of NG plates . After 1 h , excess liquid had been absorbed and individual motile L4 hermaphrodites were picked individually to plates . In the F2 generation , egg-laying defective mutants were isolated and their progeny were reanalyzed for vulva defects using Nomarski microscopy . For genetic mapping , mutant hermaphrodites in the California background were crossed with males of the Washington strain [40] . To extract genomic DNA , F2 mutant animals were picked into single tubes each containing 2 . 5 µl of lysis buffer ( 50 mM KCl; 10 mM Tris-HCl [pH 8 . 3]; 2 . 5 mM MgCl2; 0 . 45% NP-40; 0 . 45% Tween; 0 . 01% gelatin; 5 µg/ml proteinase K ) and incubated for 1 h at 65°C , followed by inactivation of the proteinase K at 95°C for 10 min . To assign linkage of a mutation to a certain chromosome , two representative SSCP markers per chromosome were tested against 42 Washington-backcrossed mutant animals . For SSCP detection , PCR samples were diluted 1∶1 in denaturing solution ( 95% formamide , 0 . 1% xylene cyanol , 0 . 1% bromophenol blue ) , denatured at 95°C for 5 min , and loaded onto a GeneGel Excel prepoured 6% acrylamide gel ( PharmaciaBiotech ) . Extrachromosomal arrays were generated by coinjecting a transgene with the marker Ppa-egl-20/Wnt: TurboRFP ( 10 ng/µl ) and genomic carrier DNA ( 60 ng/µl ) into wild-type P . pacificus , Ppa-lin-17/Fz ( tu383 ) , ced-3 ( tu54 ) , ced-3 ( tu54 ) ; egl-20/Wnt ( tu382 ) , or Ppa-lin-18/Ryk ( tu359 ) ; Ppa-mom-2/Wnt ( tu363 ) hermaphrodites as described previously [41] . Extrachromosomal arrays for rescue of Ppa-lin-17/Fz ( tu383 ) were made by injecting 5 ng/µl of a Ppa-lin-17/Fz wild-type construct into Ppa-lin-17/Fz ( tu383 ) mutants . Rescue of tu98 and tu329 was made by injecting 0 . 5 ng/µl of a 7-kb Ppa-axl-1 PCR construct together with tu98 genomic carrier DNA . Ppa-egl-20:TurboRFP expression was used to identify transgenic animals . For all constructs , multiple independent lines were generated . Expression constructs showed highly reproducible expression patterns between lines and animals . A total list of transgenes used in this study is provided as Table S2 . | Diversity of biological form in animals can be generated by the modification of already existing developmental and morphological structures . One major challenge in evolutionary biology is to identify the molecular and genetic changes associated with such morphological modifications . A decade ago , the theory of developmental systems drift was proposed arguing that large-scale changes in regulatory mechanisms can underlie the development of conserved morphological structures . Our work supports this hypothesis by comparing the development of the egg-laying organ between Caenorhabditis elegans and Pristionchus pacificus , two nematode species that have significantly different mechanisms of vulva induction despite their morphological similarity . Our studies in P . pacificus reveal major molecular alterations of signaling pathways that involve first , a novel wiring and second , the acquisition of novel protein domains in otherwise conserved receptors in WNT signaling . We show that all Wnt signaling molecules analyzed are conserved in sequence , but crucial receptor molecules have acquired novel small peptides that allow new regulatory linkages . The independent evolution of small protein domains in otherwise conserved proteins increases the evolutionary freedom of signaling pathways and developmental networks . Thus , our analysis of a developmental process that follows developmental system drift highlights the significance of protein modularity in signal transduction . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"biology"
] | 2011 | Antagonism of LIN-17/Frizzled and LIN-18/Ryk in Nematode Vulva Induction Reveals Evolutionary Alterations in Core Developmental Pathways |
Hsp70 chaperones are well known for their important functions in maintaining protein homeostasis during thermal stress conditions . In many bacteria the Hsp70 homolog DnaK is also required for growth in the absence of stress . The molecular reasons underlying Hsp70 essentiality remain in most cases unclear . Here , we demonstrate that DnaK is essential in the α-proteobacterium Caulobacter crescentus due to its regulatory function in gene expression . Using a suppressor screen we identified mutations that allow growth in the absence of DnaK . All mutations reduced the activity of the heat shock sigma factor σ32 , demonstrating that the DnaK-dependent inactivation of σ32 is a growth requirement . While most mutations occurred in the rpoH gene encoding σ32 , we also identified mutations affecting σ32 activity or stability in trans , providing important new insight into the regulatory mechanisms controlling σ32 activity . Most notably , we describe a mutation in the ATP dependent protease HslUV that induces rapid degradation of σ32 , and a mutation leading to increased levels of the house keeping σ70 that outcompete σ32 for binding to the RNA polymerase . We demonstrate that σ32 inhibits growth and that its unrestrained activity leads to an extensive reprogramming of global gene expression , resulting in upregulation of repair and maintenance functions and downregulation of the growth-promoting functions of protein translation , DNA replication and certain metabolic processes . While this re-allocation from proliferative to maintenance functions could provide an advantage during heat stress , it leads to growth defects under favorable conditions . We conclude that Caulobacter has co-opted the DnaK chaperone system as an essential regulator of gene expression under conditions when its folding activity is dispensable .
Hsp70 proteins play a key role in thermal stress adaptation by assisting the folding and refolding of client proteins , thereby preventing and reversing the accumulation of toxic protein aggregates [1] . In addition to this chaperoning activity these proteins can play critical roles in regulating gene expression in response to stress [2] . In Escherichia coli and other proteobacteria , the bacterial Hsp70 protein DnaK inhibits the heat shock sigma factor σ32 by binding to and targeting it for degradation by the membrane-bound protease FtsH until increasing demand for its folding function titrates the chaperone away to misfolded proteins [3–5] . Liberation of σ32 from DnaK in response to unfolded protein stress allows the sigma factor to associate with the RNA polymerase ( RNAP ) and induce expression of molecular chaperones and cellular proteases , among other physiological changes . These adaptive processes are collectively referred to as the heat shock response . Importantly , σ32 also induces the expression of the dnaK gene itself . This σ32-dependent upregulation of DnaK levels leads to re-sequestration of σ32 and a shut-down of the heat shock response through an autoregulatory loop [6–8] . DnaK function has been extensively studied in E . coli . Deletion of dnaK in this organism leads to cell death under heat shock conditions [9] . At intermediate and low temperatures ΔdnaK E . coli cells are viable but show growth defects [9–11] . That DnaK is dispensable under these conditions has been explained by the high degree of functional redundancy of the chaperone network and the finding that few DnaK clients are essential proteins [9 , 12–15] . In contrast to the situation in E . coli , DnaK is essential for growth not only at elevated temperature but also under optimal conditions in other bacterial species [16–20] . In Mycobacterium smegmatis , it was shown that DnaK has a non-redundant role in the folding of essential proteins and in generally maintaining protein homeostasis at both high and low temperatures and is therefore required even in the absence of stress [19] . The reasons for DnaK essentiality remain unclear in other bacteria . Like M . smegmatis , the α-proteobacterium Caulobacter crescentus , which is a model organism for studying bacterial cell cycle regulation , strictly depends on DnaK at all temperature conditions [20 , 21] . Recent work in this organism identified a previously suggested [22] role for DnaK in controlling DNA replication [23] . It was demonstrated that depletion of DnaK induces degradation of the replication initiator DnaA through upregulation of the protease Lon as part of the heat shock response . Consequently , DnaA is cleared from DnaK-depleted cells and DNA replication initiation inhibited . Here , we show that in C . crescentus DnaK not only impacts DNA replication but has an additional important role in maintaining cell growth under favorable conditions . Unexpectedly , depletion of DnaK does not lead to major protein aggregation in C . crescentus under non-stress conditions . By isolating mutations that permit growth of DnaK-depleted cells we demonstrate that in the absence of stress the sole essential function of DnaK is to inhibit σ32 . We further find that efficient repression of σ32-dependent transcription is a growth requirement and that σ32 activation under conditions when DnaK is not available reprograms gene expression on a global scale leading to a re-organization of cellular activities from proliferative to cytoprotective modes .
Depletion of the chaperone DnaK with its co-chaperone DnaJ ( hereafter DnaKJ ) results in growth arrest of C . crescentus and an inability to form colonies at all temperatures tested ( Fig 1A ) . This phenotype is not growth medium-specific as no differences were observed between incubation on complex and minimal media ( S1 Fig ) . The phenotype of cells depleted either for the main chaperone DnaK alone or its co-chaperone DnaJ mimicked the phenotype of the double depletion demonstrating that both DnaK and DnaJ are essential under the conditions tested ( S1 Fig ) . Previous work demonstrated that DnaKJ depletion in C . crescentus leads to a strong reduction in abundance of the replication initiator DnaA culminating in a G1-arrest [23] which we hypothesized could explain the essentiality of DnaK both at optimal and lower temperatures . However , restoration of DNA replication by ectopically overproducing DnaA does not restore colony formation in DnaKJ-depleted cells ( Fig 1A ) . This suggests that DnaKJ depletion impairs growth independently of its effect on DnaA . To test if the observed DNA-replication-independent growth defect in cells lacking DnaKJ was due to a collapse of general protein homeostasis ( proteostasis ) , we analyzed global protein aggregation by isolating the insoluble detergent-resistant fraction from cellular lysates ( Fig 1B ) . At 30°C , the standard and optimal growth temperature of C . crescentus , DnaKJ depletion did not induce strong overall protein aggregation compared to exponentially growing non-depleted cells , both in the presence and in the absence of the protease Lon , the primary protease degrading misfolded proteins in E . coli [24] . Instead , only a small subset of protein bands appeared or increased in abundance in the aggregate fraction of cells depleted of DnaKJ at 30°C , which could correspond to a limited amount of aggregating DnaK substrates and interacting heat shock proteins that are known to be strongly upregulated after loss of DnaKJ [6 , 25 , 26] . Exposing DnaKJ depleted cells to 45°C for two hours caused clearly stronger protein aggregation compared to the wild type , demonstrating the importance of DnaK's chaperone function at high temperature . Together , these data show that DnaKJ depletion under non-stress conditions does not induce drastic protein aggregation , indicating that a proteostasis collapse is not the reason for the observed growth defect in DnaKJ-depleted cells . We searched for mutations that can restore growth in the absence of DnaKJ using a suppressor screen ( see Material and Methods for a detailed description of the screening procedure ) . Mutations identified in the screen mapped to genes encoding different components of the transcriptional apparatus as well as a subunit of an ATP-dependent protease . We found five mutations in rpoH , the gene encoding σ32 , one mutation in rpoB , which codes for the β-subunit of the RNAP , one mutation upstream of rpoD encoding the house keeping sigma subunit σ70 of the RNAP and one mutation in the hslU gene , which encodes the ATPase subunit of the protease HslUV ( S2 Fig ) . The mutations in rpoH and rpoB were missense mutations leading to amino acid replacements . The mutation upstream of rpoD was an 11 bp-deletion , affecting the -10 region of one of the promoters . The mutation in hslU was a 9 bp-deletion removing three amino acids ( Δ201-203HKT ) in the substrate recognition domain of HslU ( S2 Fig ) . To confirm that the suppression of growth defects observed in the clones isolated in our suppressor screen are linked to these mutations , we re-introduced them into the parental DnaKJ-depletion strain ( SG400 ) harboring no other genetic modifications or plasmids . Re-introduction of each of these mutations restored growth in the absence of DnaKJ ( Fig 2A ) . Based on this result , we used the strains with the re-introduced suppressor mutations in all following experiments . We assessed colony formation of these strains at different temperatures and observed that the degree of suppression was temperature-dependent and varied between the different mutants . Most notably , the rpoH ( V170A ) allele suppressed loss of DnaKJ well at 22°C and 30°C , but failed to restore colony formation at 15°C ( Fig 2A ) . Furthermore , under non-depleting conditions this mutation prevented colony formation at 37°C , at which the parental strain and all other suppressor mutants readily form colonies ( S3 Fig ) . All together , we isolated several mutations capable of restoring growth at the physiological temperature of 30°C in the absence of DnaKJ . Previously identified loss-of-function point mutations in rpoH have been shown to weaken heat shock gene induction [23 , 27] . Consistent with these data , all suppressor mutations we identified that mapped to rpoH led to reduced σ32-dependent gene expression following DnaKJ depletion as monitored by LacZ-assays using a PdnaKJ-lacZ reporter plasmid [28] or , in strains that exhibited difficulties maintaining the plasmid , by quantifying mRNA levels of the σ32-controlled hslU gene ( Fig 2B ) . Western blot analysis after nine hours of DnaKJ-depletion revealed no significant differences between σ32-levels in the wild type background and strains harboring the different rpoH alleles ( S4 Fig ) , suggesting these mutations reduce transcriptional activity rather than protein stability . In accordance with this , quantitative chromatin immunoprecipitation ( qChIP ) experiments with two of the σ32 mutants , σ32 ( Q202R ) and σ32 ( D252G ) , showed a decrease in σ32 occupancy of the hslVU operon promoter after six hours of DnaKJ-depletion ( Fig 2C ) . We wondered if complete absence of σ32 could also suppress the lack of DnaKJ , and deleted the rpoH gene from the DnaKJ depletion strain . Under non-depleting conditions this strain formed colonies at intermediate ( 30°C and 22°C ) and low temperatures ( 15°C ) , but not at 37°C , similar to the strain harboring the V170A mutation in rpoH ( S3 Fig ) . Spot assays showed that at 15°C the rpoH deletion strain indeed suppressed the lethal effect of the DnaKJ depletion ( Fig 2A ) . However , unlike the rpoH suppressor mutations , deletion of rpoH did not allow growth at 30°C or 22°C , even though the block in DNA replication after loss of DnaKJ was alleviated ( S5 Fig ) . Given the complete absence of heat shock response induction in this strain ( Fig 2B ) , we reasoned that ΔrpoH cells are not able to compensate for the lack of DnaKJ chaperoning function at 30°C and 22°C . Indeed , we observed a strong increase in recovered protein aggregates after depletion of DnaKJ in the ΔrpoH background ( Fig 2D ) . By contrast , the DnaKJ depletion strain containing the rpoH ( Q202R ) mutation , which shows intermediate σ32-dependent promoter activity ( Fig 2B and 2C ) , behaved similarly to DnaKJ-depleted cells with wild type σ32 and did not show significant protein aggregation ( Fig 2D ) . In summary , our data demonstrate that partial and , at low temperatures , complete loss of σ32 activity can bypass the lethality of DnaKJ depletion . These data suggest that DnaKJ’s regulatory function on σ32 is essential for survival in the absence of stress , while its chaperoning function is redundant and can be compensated by the upregulation of other HSPs . We also examined σ32-dependent gene expression in the strains harboring mutations in hslU , rpoD or rpoB either by LacZ-assays or qRT-PCR . Interestingly , as for cells harboring mutations in rpoH , in these strains σ32-dependent transcription was also notably reduced after 6 and 9 hours of DnaKJ depletion ( Fig 2B ) , suggesting that all suppressor mutations identified in the screen act through the same pathway . How does a mutation in hslU affect σ32-dependent gene expression ? To test if growth restoration in DnaKJ-depleted cells harboring the hslU ( Δ201-203HKT ) allele was due to loss of function of the protease HslUV , we analyzed growth and PdnaKJ-lacZ expression following DnaKJ depletion in a ΔhslU strain lacking the HslUV unfoldase subunit HslU . However , unlike the hslU ( Δ201-203HKT ) suppressor mutation , this ΔhslU mutation neither reduced σ32-dependent gene expression nor restored growth in DnaKJ-depleted cells ( Fig 2A and 2B ) . To test the possibility that the hslU ( Δ201-203HKT ) mutation affects σ32 degradation in DnaKJ-depleted cells , we measured σ32 steady-state levels and stability . During DnaKJ depletion σ32 levels significantly increase over the course of depletion , and deletion of the entire hslU gene has no effect on this observed increase . However , during DnaK depletion in the hslU ( Δ201-203HKT ) mutant , σ32 levels only slightly increase and remain low throughout ( Fig 3A ) . Consistent with these data , after 8 h of DnaKJ depletion the half-life of σ32 was strongly decreased in the hslU ( Δ201-203HKT ) mutant ( t1/2 ≈ 23 . 5 min ) , compared to the pronounced stabilization ( t1/2 > 360 min ) in the parental DnaKJ-depletion strain ( Fig 3B ) . We next asked if overexpression of hslU ( Δ201-203HKT ) would also lead to reduced σ32 levels and restoration of colony formation in DnaKJ depleted cells . Western blots and spot assays confirmed reduction of σ32 levels and restored colony formation ( Fig 3C and 3D ) . Furthermore , σ32 levels were reduced beyond detection when co-overexpressing hslU ( Δ201-203HKT ) along with hslV , encoding the proteolytic subunit of the protease complex . This stronger reduction could account for the fact that these cells could not grow at 30°C in the absence of DnaKJ but we cannot yet explain why they do not grow at 15°C . It is well established that FtsH is the protease normally degrading σ32 [3 , 4] , therefore to investigate whether FtsH was required for the hslU ( Δ201-203HKT ) -dependent destabilization of σ32 , we overexpressed hslVU ( Δ201-203HKT ) in cells depleted for FtsH . On its own , FtsH depletion leads to a strong stabilization of σ32 , as expected ( Fig 3E ) . In contrast , overexpression of hslVU ( Δ201-203HKT ) completely counteracted this stabilization ( Fig 3E ) , demonstrating that hslU ( Δ201-203HKT ) destabilizes σ32 independently of FtsH . Furthermore , although the hslU ( Δ201-203HKT ) mutant readily destabilizes σ32 , overexpression of wild type hslU alone or in conjunction with hslV had no effect on σ32 levels and could not restore growth in the absence of DnaKJ ( Fig 3C–3E , S7 Fig ) . Based on these data and the fact that the Δ201-203HKT mutation is located in the substrate recognition region of HslU ( S2 Fig ) , we propose that this mutation is a gain-of-function mutation rendering HslUV capable of degrading σ32 . Importantly , we neither observed increased degradation of the two unstable proteins CtrA and FtsZ nor the stable protein RpoB in the hslU ( Δ201-203HKT ) mutant strain ( S6 Fig ) . Therefore , while these data indicate that the HslVU ( Δ201-203HKT ) mutant does not affect the majority of proteins , at this point we cannot say whether the effect of the mutation is restricted to σ32 or if the degradation of other proteins is affected as well . Next , we addressed the molecular basis of how the 11 bp deletion in the promoter region of rpoD restores growth in DnaKJ-depleted cells . To this end , we tested if the mutation alters σ70 levels during the course of DnaKJ depletion ( Fig 4A ) . Initially , the amount of σ70 was indistinguishable between the mutant and the parental strain . However , during the course of depletion , σ70-levels decreased to ~60% in the wild type background while they increased to ~150% in the rpoD promoter mutant . To test whether upregulation of σ70 is sufficient to restore growth of DnaKJ-depleted cells , we ectopically overexpressed rpoD in the DnaKJ depletion strain . Indeed , during DnaKJ depletion and rpoD overexpression colonies formed at low and intermediate temperatures ( Fig 4B ) . Based on previously proposed sigma competition models [29 , 30] , we suggest that increased levels of σ70 may outcompete σ32 for binding to the RNAP core . In accordance with this model , the suppressor mutation in the rpoD promoter led to reduced σ32 occupancy of the σ32-dependent hslVU promoter compared to the parental DnaKJ depletion strain , without significantly affecting σ32 levels ( Fig 2C , S4 Fig ) . Further , the rpoB ( P1272S ) suppressor mutation we identified is located in a region known to be required for interaction with sigma factors ( S2 Fig ) [31] . Thus , this mutation might impair σ32 association with the RNAP core , a hypothesis supported by the observation that hslVU promoter occupancy by σ32 was reduced in cells harboring the rpoB ( P1272S ) allele ( Fig 2C ) . Altogether our data clearly demonstrate that attenuation of σ32 activity , whether by mutation or deletion of the rpoH gene , by increasing σ32 degradation , by outcompeting σ32 for binding to the RNAP , or by reducing the affinity of the RNAP towards σ32 , can bypass the growth defect of DnaKJ depleted cells ( Fig 4C ) . Our data strongly suggest that DnaKJ is essential to prevent inappropriate induction of σ32-dependent genes or high levels of the sigma factor itself , which could be detrimental to the proper functioning of the cell . To test this hypothesis more directly , we next analyzed the effects of rpoH overexpression on the growth rates . Overexpression of rpoH led to a noticeable decrease in growth rate , particularly in a strain containing a transposon insertion in the dnaKJ promoter ( PdnaK::Tn ) ( Fig 5A ) , which is deficient in upregulating dnaKJ upon σ32 activation leading to a more persistent induction of the heat shock response [23] ( Fig 5B ) . The growth defect was further amplified when overproducing a σ32 variant containing the amino acid substitution V56A , which corresponds to the I54A substitution in E . coli that allows σ32 to escape FtsH-dependent degradation [32 , 33] . Although overexpression of rpoH ( V56A ) clearly induced heat shock gene expression more strongly than wild type rpoH , we did not observe stabilization of the mutant variant ( S8 Fig ) , suggesting that in C . crescentus this mutation increases σ32 activity by another mechanism . The growth inhibition induced by σ32 ( V56A ) overproduction was clearly reduced by introducing the Q202R suppressor mutation into σ32 ( V56A ) . Similarly , disruption of σ32 ( V56A ) transcriptional activation by exchanging the three amino acids Y106 , W109 and W110 responsible for promoter melting [34–37] against alanines ( YWW::AAA ) abolished heat shock response induction and completely restored normal growth ( Fig 5A and 5B ) . Measurement of σ32 steady state levels after vanillate-dependent induction confirmed that all variants were expressed at similar levels ( S8 Fig ) . Finally , the deleterious effect of elevated σ32 ( V56A ) levels could also be counteracted by overproducing σ70 ( Fig 5C ) , supporting the hypothesis that σ70 can rescue DnaKJ-depleted cells by outcompeting σ32 for binding to the RNAP , thus reducing σ32-dependent transcription . In summary , these data show that σ32 inhibits growth in a dose-dependent manner and does so , albeit to a lower degree , also in cells harboring DnaKJ . To better understand the inhibitory effect of σ32 on growth we monitored transcriptome changes induced by σ32 activation through RNA-sequencing . The transcriptional profile of DnaKJ-depleted cells harboring wild type rpoH correlated well with σ32 ( V56A ) overproducing cells ( R2 = 0 . 77 ) ( Fig 6A ) . By contrast , in DnaKJ-depleted cells containing the ΔrpoH deletion this correlation was largely lost ( R2 = 0 . 11 ) , confirming that transcriptional changes after DnaKJ depletion are mainly due to increased σ32 activity . In the DnaKJ depletion strain containing the rpoH ( D252G ) mutation , gene expression changes were similar to the parental DnaKJ depletion strain , however less pronounced ( Fig 6A and 6C; S9 Fig ) . Genes that were >2-fold up- or downregulated both upon DnaKJ depletion and upon σ32 ( V56A ) overexpression and that showed a significantly milder or no response ( > 2-fold ) in the ΔrpoH background were classified as σ32-regulated genes . We found in total 338 up- and 82 downregulated genes , belonging to this group ( Fig 6B and 6C ) . Integration of our transcriptomics data with existing C . crescentus σ32-ChIPseq data [38] revealed that 181 of the 338 upregulated genes harbor an upstream ChIP-seq signal for σ32-enrichment or are in an operon with such a gene ( Fig 6C ) . These 181 directly upregulated genes include most heat shock genes encoding highly conserved proteases and chaperones , but also many metabolic genes as well as genes involved in small molecule membrane transport ( S9 Fig ) . Likewise , the group of downregulated genes included genes involved in metabolism and membrane transport , but also several genes encoding cell cycle regulators and factors involved in cellular development ( S9 Fig ) . Hence , σ32 induction leads to wide transcriptomic changes that go far beyond the upregulation of chaperones and proteases . To gain a more complete view on the cellular effects mediated by σ32 induction following DnaKJ depletion , we compared the proteome in DnaKJ-depleted cells to the non-depleted control by quantitative proteomics using a tandem mass tag ( TMT ) isobaric labeling approach . We identified 346 proteins that were at least 1 . 5-fold upregulated ( p ≤ 0 . 05 ) ( S9 Fig ) . For 203 of these proteins the corresponding transcripts were also upregulated in a σ32-dependent manner ( Fig 6B and 6C ) . In the case of the 447 downregulated proteins ( > 1 . 5-fold , p ≤ 0 . 05 ) , less than 12% of the corresponding genes were transcriptionally downregulated . To investigate energetic investment into the different functional categories of the proteome in DnaKJ-depleted and non-depleted cells , we compared the relative abundances of all analyzed proteins using proteomaps [39] . Depletion of DnaKJ induced large-scale changes of proteome fractions dedicated to maintenance , metabolism and proliferative processes . More specifically , the mass fraction of the proteome dedicated to folding , export and degradation increased by 49% due to the strong upregulation of chaperones and proteases , particularly due to the highly abundant GroEL/ES chaperone ( Fig 7A and 7B , S10 Fig ) . By contrast , the majority of differentially regulated abundant proteins involved in protein translation were downregulated . The downregulation of this group of proteins likely accounts for the decrease in total protein synthesis that we have previously observed in DnaKJ depleted cells [23] . In particular , levels of the highly abundant translation factor EF-Tu ( encoded by tuf ) , and the small ribosomal protein RpsF were notably reduced ( by 45% or 44% , respectively ) ( Fig 7A ) . Importantly , compromised activity of σ32 in the rpoH ( D252G ) and rpoH ( Q202R ) backgrounds restored EF-Tu levels in DnaKJ-depleted cells ( Fig 7B ) , demonstrating that the observed downregulation depends on σ32 . Taken together our results demonstrate that induction of σ32 in DnaKJ-depleted cells induces wide proteome changes and a reorganization of cellular functions . While the major proteome fractions dedicated to repair and maintenance functions were upregulated , major subproteomes required for the growth-promoting functions of protein translation , DNA replication and for certain metabolic processes were downregulated .
In this study we have analyzed in detail the essential function of DnaK in an α-proteobacterium in which the chaperone is strictly required for growth at all temperatures . Our study demonstrates that the reason for DnaK's essentiality under non-stress conditions is its role in repressing σ32 activity and in this way counteracting extensive reprogramming of global gene expression , with growth-inhibitory consequences ( Fig 7C ) . The ancestral function of DnaK as a folding catalyst is required only at elevated temperatures . Under optimal conditions it is dispensable and its loss can be compensated by other chaperones and proteases . Our data highlight that gene essentiality can have multiple causes that differ depending on the growth condition; in the case of DnaK , two essential functions switch in a temperature-dependent manner . Our conclusions regarding DnaK's essentiality are based on suppressor mutations that bypass the lethality of DnaK depletion and that all act by attenuating σ32 activity ( Fig 4C ) . Besides cis-acting mutations in the rpoH gene that reduce σ32 activity , we also identified mutations in rpoD , rpoB and hslU that reduce σ32 activity in trans , providing important new insight into the regulatory pathways controlling σ32 activity . Most notably , a mutation in hslU led to increased σ32 degradation , and a mutation in the rpoD promoter resulting in higher levels of the housekeeping σ70 negatively affected σ32-dependent gene expression . HslUV is the most recently identified ATP-dependent bacterial protease [40] . Despite much biochemical and structural insight , the physiological roles of this protease remain poorly understood and only few confirmed substrates are known in E . coli . In Caulobacter , which has advanced as an important model for regulated proteolysis [41] , no substrates have yet been identified for this protease . Past work proposed that HslUV together with the other cytosolic proteases might degrade σ32 in E . coli [42] . A follow-up study showed that purified E . coli HslUV degrades E . coli σ32 in vitro at elevated temperatures [43] . In our experiments , we observed that the hslU ( Δ201-203HKT ) mutation caused rapid FtsH-independent degradation of σ32 . By contrast , we did not observe degradation of σ32 by wild type HslUV , indicating that removal of the three amino acids HKT in the substrate recognition region of HslU may alter substrate selectivity or alternatively , bypass the need for an accessory factor that is required to deliver σ32 to HslUV and might be present only under certain conditions . Future work will characterize the discovered HslU mutation in more detail and elucidate the role of wild type HslUV in σ32 degradation . Our data also demonstrate that elevated levels of the housekeeping σ70 can negatively affect σ32-dependent gene expression , demonstrating that the two sigma factors are able to influence each other . In agreement with sigma-competition models [29 , 30] , our data suggest that σ70 outcompetes σ32 for binding to the RNAP and in this way reduces σ32-dependent transcription . Previous work showed that under heat stress conditions σ70 is aggregation prone [44] , which might further relieve competitive pressure on σ32 when necessary . While our data demonstrate that increased σ70 levels can inhibit σ32 , it is possible that conversely high levels of active σ32 during heat shock or DnaK depletion inhibit σ70 from associating with the RNAP . A σ32-induced reduction of σ70-dependent transcription could contribute to the global gene expression changes that we observed in our RNA-sequencing experiments . Mutations reducing σ32 activity have also been found to improve growth at low temperatures in a dnaK deletion mutant in E . coli [27] . In an analogous way , growth defects at physiological temperatures after loss of the cytoplasmic Hsp70s SSA1 and SSA2 in yeast can be suppressed by mutations that attenuate the functionality of the heat shock transcription factor HSF1 [45] . The emergence of regulatory interactions between Hsp70 homologs and heat shock response effectors during evolution has thus added a new layer of cellular dependency on a chaperone system whose main ancestral function as a folding catalyst is dispensable at optimal temperatures . Noticeably , in Bacillus subtilis , a member of the firmicutes , which does not possess σ32 , deletion of DnaK has almost no observable consequences in the absence of stress [46] . It remains to be studied whether the deleterious effects of DnaK loss in other proteobacteria are linked to increased σ32 activity and global gene expression changes and how the functions of DnaK are utilized for survival in other bacterial groups . Our data demonstrate that liberated σ32 activity shifts cellular resources from proliferative functions to cytoprotective functions . While this re-allocation leads to detrimental growth defects under favorable conditions , it could represent an effective way to protect stress sensitive processes and to re-allocate resources to active stress adaptation . In particular , nascent proteins are aggregation prone since incomplete polypeptides cannot yet fold into their native conformation and are threatened by unproductive interactions with other chains emerging from neighboring ribosomes [47] . Slowing down growth and global protein synthesis can provide advantageous conditions for protein folding when the chaperoning capacity is limited [13 , 48] . Furthermore , slowly growing or dormant cells are often resistant to certain stress conditions like antibiotic treatment [49] .
C . crescentus cultures were grown in the complex medium PYE or the minimal medium M2 with the following supplements if required: 0 . 3% xylose , 0 . 2% glucose , 3% sucrose , 1 mM IPTG or 500 mM vanillate . Antibiotics were added in concentrations as described earlier [23] . Liquid cultures were routinely grown while shaking at 200 rpm at the desired temperatures and regularly diluted to assure growth in the exponential phase . To deplete proteins of interest , cells were washed two times with PYE by centrifugation ( 6000 g , 4 min ) before resuspension in medium devoid of the inducer . Transductions were performed using ϕCr30 as described previously [50] . E . coli was grown for cloning purposes in LB supplemented with antibiotics as necessary at 37°C . We performed the suppressor screen with strain ML2009 [23] a derivative of the xylose-dependent DnaKJ depletion strain SG400 [21] , which contains a chloramphenicol resistance-conferring high copy plasmid carrying PvanA-dnaA . Shifting the cells from a xylose-containing to a glucose-containing medium results in a block of dnaKJ expression while addition of vanillate allows for the inducible overproduction of the DNA replication initiator DnaA . Overly high levels of DnaA are lethal and can be counteracted by increased degradation through the protease Lon which is upregulated as part of the heat shock response after depletion of DnaKJ [23] . By overproducing DnaA in DnaKJ-depleted cells , we intended to produce selective pressure against restoration of DnaKJ levels through mutations in the xylose-dependent promoter or xylR which would fully abolish heat shock gene induction and thus decrease degradation of DnaA . ML2009 cells were depleted of DnaKJ for five hours without overproduction of DnaA in PYE liquid medium containing glucose and chloramphenicol . 100 μL of log-phase culture ( OD600 of 0 . 1–0 . 4 ) were plated on PYE agar plates containing glucose/chloramphenicol/vanillate to repress DnaKJ synthesis and to induce overexpression of dnaA . Colonies appearing after three , four and five days of incubation at 30°C were restreaked on PYE plates supplemented with either xylose/chloramphenicol or glucose/chloramphenicol/vanillate . Cells grown on the latter medium were restreaked on plates containing xylose/chloramphenicol/vanillate to check for continued lethality of dnaA overexpression in the presence of DnaKJ and to confirm the absence of mutations reducing DnaA functionality . Clones incapable of growing on xylose/chloramphenicol/vanillate were cultured overnight in liquid PYE medium supplemented with either xylose/chloramphenicol or glucose/chloramphenicol/vanillate . Cells grown under DnaKJ producing ( PYE supplemented with xylose/chloramphenicol ) as well as DnaKJ-depleting and DnaA-overproducing conditions ( PYE supplemented with glucose/vanillate/chloramphenicol ) were frozen and tested for the absence of DnaK by western blot . The majority of clones which arose after three or four days of incubation on the initial plate had restored DnaK levels to varying degrees , indicating that these clones harbored mutations in the xylose-inducible promoter . By contrast , in most clones that appeared after five days , DnaK levels were as low as in the parental strain following 24 hours of depletion , indicating that these clones have acquired mutations that allow them to grow in the absence of DnaK . Suppressors in rpoH were identified by conventional sequencing of the rpoH gene . The other suppressor clones were verified for the absence of mutations in the gene encoding the XylR repressor or in Pxyl upstream of the dnaKJ operon and then subjected to whole genome sequencing . In addition to mutations in rpoH , rpoB , rpoD and hslU , we also identified five mutations in other genes ( S2 Table ) . Because of genetic instability and difficulties to re-introduce these mutations into the DnaKJ depletion strain , these clones were excluded from further analysis . Whole genome sequencing was performed by GENEWIZ , South Plainfield , NJ using an Illumina Nextera XT DNA prep workflow and sequencing on MiSeq with 2x150bp configuration . Mapping to the Caulobacter crescentus NA1000 ( NC_011916 ) reference sequence was performed for each of the samples , after which detection of SNPs/INDELs was conducted . Cells analyzed by phase-contrast microscopy were fixed with 1% formaldehyde immediately after sampling and stored at 4°C . Prior to analysis the cells were mounted on PYE 1% agarose pads . Phase contrast images were obtained using a Ti eclipse inverted research microscope with a 100x/1 . 45 NA objective ( Nikon ) . The images were processed with Fiji ( ImageJ ) . Flow cytometry samples were prepared as previously described [51] and analyzed with the BD LSRFortessa ( BD Biosciences ) flow cytometer . For each sample 30000 cells were analyzed . The experiments were performed in biological replicates and representative results are shown . The obtained data were analyzed and processed with FlowJo . Sample preparation and western blot analysis was performed as previously described [52] . Equal loading of total protein and the quality of the transfer was assessed through protein visualization using the TCE in-gel method [53] . The following primary antibodies were used in appropriate dilutions: α-DnaK ( this study ) , α-DnaA [54] , α-GroEL ( kindly provided by S . L . Gomes ) , α-CtrA ( kindly provided by M . Laub ) , α-FtsZ ( kindly provided by M . Thanbichler ) , α-σ32 ( clone 3NR3 , BioLegend Cat . No . 663402 ( used in experiment for Fig 3B ) and an antibody kindly provided by F . Narberhaus ( used in all other experiments ) , α-σ70 ( clone 2G10 , BioLegend Cat . No . 663203 ) , α-RpoB ( clone 8RB13 , Biolegend Cat . No . 663903 ) and α-EF-Tu ( clone 900 , Hycult Cat . No . HM6010 ) . The primary antibodies were detected using secondary HRP-conjugated antibodies . SuperSignal Femto West ( Thermo Scientific ) served as detection reagent . The blots were scanned using the Chemidoc ( Bio-Rad ) system . Relative band intensities were quantified with the Bio-Rad Image Lab software and images were processed with Fiji ( ImageJ ) . His6-DnaK was expressed from the pML375-dnaK plasmid and purified by Ni-NTA affinity chromatography as described previously [55] . The purified His6-DnaK was used to generate rabbit polyclonal antisera against DnaK ( Davids Biotechnologie GmbH ) . Cell cultures with an OD600 corresponding to the exponential phase of wild type cells ( OD600 0 . 1–0 . 4 ) in PYE were diluted to OD600 0 . 05 and serially diluted 1:10 before 2 μL of each dilution were spotted on PYE rich or M2 minimal medium agar plates . Cultures were diluted to an OD600 of 0 . 025 before being added as 100 μL volumes into sterile 96 well plates . Cells were cultured shaking at 30°C and measured every 10 min using either an infinite 200 Pro ( Tecan ) ( linear shaking amplitude 3 . 5 mm ) or a SpectraMax i3x ( Molecular Devices ) ( linear shake intensity set to high ) plate reader . The in vivo aggregation protocol was adapted from previously published procedures [19 , 56 , 57] . About 40 OD600 units of exponential cells were rapidly cooled in an ice-water bath and pelleted ( 6000 g , 10 min , 4°C ) . All following steps were performed at 4°C and buffers used during or after cell lysis were supplemented with cOmplete ULTRA protease inhibitor cocktail ( Sigma-Aldrich ) . The pellets were washed once in buffer A ( 50 mM Tris/HCl pH8 . 0 , 150 mM NaCl ) and resuspended in 4 mL buffer A supplemented with 12 U/mL benzonase ( Merck Millipore Cat . No . 70746 ) . Cells were lysed by two passages at 1000 psi in a FRENCH Press using a Manual-Fill Mini-Cell ( FA-003 ) ( Thermo Spectronic ) . Lysates were centrifuged ( 5000 g , 10 min ) two times in order to remove intact cells . The protein concentration of the cleared lysate was measured by Bradford assay and 3 mL were centrifuged ( 20000 g , 20 min ) to pellet the insoluble protein fraction . The supernatant was discarded , the pellets were resuspended in 300 μL buffer A and subsequently sonicated in a Bioruptor ( Diagenode ) ( set to high , 1 cycle for 30 s , 4°C ) followed by a centrifugation ( 20000 g , 20 min ) for washing . This procedure was repeated two more times using buffer A supplemented with 1% ( v/v ) Triton X-100 with an incubation of 1 h on ice with regular vortexing prior to the sonication , before a last wash step identical to the first one was performed . The protein pellet was resuspended in 100 μL buffer A by sonication ( set to high , 1 cycle for 30 s , 4°C ) , supplemented with 5x SDS loading buffer and heated to 95°C for 10 min . Protein fractions were normalized to each other according to the concentration measurements of the Bradford assay by addition of 1x SDS loading dye . Cells containing the PdnaKJ-lacZ ( pDel1 ) reporter were harvested from exponential phase ( OD600 0 . 1–0 . 4 ) PYE ( supplemented with tetracycline for maintenance of the reporter plasmid ) liquid cultures grown shaking at 30°C . β-galactosidase activity was measured using the standard protocol [58] . RNA was isolated from log phase bacterial cultures grown shaking at 30°C using the RNeasy mini kit ( Qiagen ) . Equal amounts of RNA were reverse transcribed into cDNA using the iScript cDNA synthesis kit ( Bio-rad ) . The obtained cDNA served as template in a real-time PCR reaction using the iTaq universal SYBR Green Supermix ( Bio-rad ) . For the analysis of 16S rRNA or hslU mRNA levels in each sample , the primer pairs RT_16SFor/RT_16SRev and OFS399/OFS400 were used , respectively . The analysis was performed using a StepOnePlus real-time PCR system ( AppliedBiosystems , Foster City , CA ) in standard run mode . For each reaction a dissociation curve was run after completion to rule out detection of primer dimerization or amplification artifacts . Gene expression profiles of hslU were normalized against 16S rRNA levels as endogenous control . Relative expression levels were calculated employing the comparative Ct method . Samples were prepared according to a previously published protocol [59] with the modifications to the protocol and the subsequent qRT-PCR as described in Heinrich et al . 2016 [52] . For immunoprecipitation of σ32-DNA complexes an α-σ32 antibody ( kindly provided by F . Narberhaus ) was applied in a 1:400 dilution . For measuring promoter occupancy by σ32 , qRT-PCR assays were performed with primers OFS984/OFS985 amplifying a 100 bp region encompassing the heat shock-inducible promoter of the hslVU operon , corresponding to a σ32-binding site according to Haakonsen et al . 2015 [38] . In vivo protein degradation rate measurements were performed as previously described [23] . Chloramphenicol was used at a concentration of 100 μg/mL to shut down protein synthesis . RNA was extracted from cell pellets using the RNeasy mini kit ( Qiagen ) . RNA-sequencing was performed by GENEWIZ , South Plainfield , NJ . Fold changes were calculated as the ratio between the normalized expression values of the respective DnaKJ depletion samples or the σ32 ( V56A ) overexpression and the non-depleted Pxyl-dnaKJ reference sample . Transcripts were considered differentially regulated after DnaKJ depletion in the WT background when they were at least twofold up- or downregulated after six and nine hours of depletion ( S3 Table ) . Gene expression data are available at the Gene Expression Omnibus repository: GSE102372 . For the identification of putative direct transcriptional targets of σ32 we used published ChIP-seq data from Haakonsen et al . 2015 [38] ( GSM1906341 ) . Peaks were defined where the read density remained above 10 , assigned to the position where their maximum value was found ( +/- 10 bp due to binning of the original data ) , and summed to determine the peak area . We then searched for ORFs with a translation start site within 200 bp up- and 60 bp downstream of the peak maximum . ORFs were considered to be in an operon if specified by Schrader et al . [60] in which case all peaks potentially affecting transcription of the ORF were considered . For quantitative proteomics experiments 100 mL cultures of SG400 cells grown in PYE xylose or in PYE glucose for 12 h to an OD600 between 0 . 1 and 0 . 4 were rapidly cooled in an ice water bath . The cultures were pelleted ( 6000 g , 10 min , 4°C ) and washed once with ice-cold ddH2O . Each condition was cultured as independent biological triplicates . The cell pellets were kept at -80°C before analysis . Protein digestion , TMT10 plex isobaric labeling and the mass spectrometrical analysis was performed by the Clinical Proteomics Mass Spectrometry facility , Karolinska Institute/Karolinska University Hospital/Science for Life Laboratory . Sample protein abundances were normalized to portions of 100% for each sample and each protein quantity mean of the experimental sample was divided by the mean of the control sample . The statistical significance of the log2-fold changes was assessed using p-values obtained based on Student’s t-test ( S3 Table ) . Proteins were considered differentially regulated after DnaKJ depletion when they were at least 1 . 5-fold up- or downregulated . Proteomaps were generated as previously described [39 , 61] . Functional categorizations were manually assigned to the proteins identified through mass spectrometry using KEGG pathway maps , gene ontology ( GO—biological process ) as basis and by literature search ( S3 Table ) . | Molecular chaperones of the Hsp70 family belong to the most conserved cellular machineries throughout the tree of life . These proteins play key roles in maintaining protein homeostasis , especially under heat stress conditions . In diverse bacteria the Hsp70 homolog DnaK is essential for growth even in the absence of stress . However , the molecular mechanisms underlying the essential nature of DnaK have in most cases not been studied . We found in the α-proteobacterium Caulobacter crescentus that the function of DnaK as a folding catalyst is dispensable in the absence of stress . Instead , its sole essential function under such conditions is to inhibit the activity of the heat shock sigma factor σ32 . Our findings highlight that some bacteria have co-opted chaperones as essential regulators of gene expression under conditions when their folding activity is not required . Furthermore , our work illustrates that essential genes can perform different essential functions in discrete growth conditions . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"cellular",
"stress",
"responses",
"caulobacter",
"enzymes",
"cell",
"processes",
"enzymology",
"mutation",
"prokaryotic",
"models",
"dna",
"replication",
"experimental",
"organism",
"systems",
"gene",
"types",
"dna",
"molecular",
"biology",
"techniques",
"bacteria",
"research",
"and",
"analysis",
"methods",
"heat",
"shock",
"response",
"proteins",
"caulobacter",
"crescentus",
"gene",
"expression",
"hyperexpression",
"techniques",
"molecular",
"biology",
"molecular",
"biology",
"assays",
"and",
"analysis",
"techniques",
"gene",
"expression",
"and",
"vector",
"techniques",
"biochemistry",
"cell",
"biology",
"nucleic",
"acids",
"genetics",
"biology",
"and",
"life",
"sciences",
"proteases",
"suppressor",
"genes",
"organisms"
] | 2017 | An essential regulatory function of the DnaK chaperone dictates the decision between proliferation and maintenance in Caulobacter crescentus |
Integrating large-scale functional genomic data has significantly accelerated our understanding of gene functions . However , no algorithm has been developed to differentiate functions for isoforms of the same gene using high-throughput genomic data . This is because standard supervised learning requires ‘ground-truth’ functional annotations , which are lacking at the isoform level . To address this challenge , we developed a generic framework that interrogates public RNA-seq data at the transcript level to differentiate functions for alternatively spliced isoforms . For a specific function , our algorithm identifies the ‘responsible’ isoform ( s ) of a gene and generates classifying models at the isoform level instead of at the gene level . Through cross-validation , we demonstrated that our algorithm is effective in assigning functions to genes , especially the ones with multiple isoforms , and robust to gene expression levels and removal of homologous gene pairs . We identified genes in the mouse whose isoforms are predicted to have disparate functionalities and experimentally validated the ‘responsible’ isoforms using data from mammary tissue . With protein structure modeling and experimental evidence , we further validated the predicted isoform functional differences for the genes Cdkn2a and Anxa6 . Our generic framework is the first to predict and differentiate functions for alternatively spliced isoforms , instead of genes , using genomic data . It is extendable to any base machine learner and other species with alternatively spliced isoforms , and shifts the current gene-centered function prediction to isoform-level predictions .
Determining the functions of proteins is a central goal of genetics , fundamental for understanding the molecular basis of diverse genetic diseases [1]–[7] . During the past few decades , significant efforts have been made to integrate and develop diverse machine learning algorithms for gene function prediction through large-scale genomic data integration [8]–[12] , such as Support Vector Machines , Bayesian classifications and Artificial Neural Networks . Despite differences in implementation and performance of the specific algorithms , the essence of these methods for gene function prediction is ‘supervised learning’ , in which a model of features derived from functional genomic data ( such as microarray , protein-protein physical interactions , gene-gene genetic interactions ) is constructed to delineate a defined set of ‘positives’ ( genes annotated to the function under consideration ) and ‘negatives’ ( genes without the function ) . These algorithms have significantly accelerated our understanding of gene functions . A major intellectual limitation of the current function prediction paradigm is that it considers a gene as a single entity without differentiating the functional diversity of alternatively spliced isoforms . Alternative splicing is a major source of protein molecular function diversity and regulatory diversity . In humans , 95% of multi-exon genes undergo alternative splicing , generating proteins of potential different functions [13]–[21] . For example , TRPM3 , which encodes a type of cation-selective channels in human , can be alternatively spliced into two variants targeting different ions [22]–[24] . Other splice variants have been reported with distinctly opposite functions . For example , the splice variants of BCLX are antiapoptotic and pro-apoptotic , respectively [25] . Similarly , CASP3-L variant is pro-apoptotic while CASP3-S is antiapoptotic [26] . Differences in function are sometimes reflected on the regulatory level: two alternatively spliced transcripts of OSR2 have opposite transcriptional activities , activation and repression [26] . Attempts to capture such differential functions are currently limited to low-throughput experimental approaches and protein domain analysis . However , based on the protein domain annotation we downloaded in Dec , 2012 , only 34% of the isoform pairs ( of the same gene ) in the NCBI database have different domains . The majority of the isoforms of the same gene differ in subtle ways which are not reflected by protein domains . We expect that computational methods developed to differentiate isoform functions through integrating functional genomic data will assist a deeper , high-resolution understanding of gene functions . The key challenge facing isoform prediction is the lack of a systematic catalog of isoform-level function annotations and large-scale genomic data resolved at the isoform level . The latter is resolved by the unprecedented amount of transcriptomic data generated by next-generation sequencing [27]–[29] . RNA-seq data available in public databases [30] now surpass the number of microarray data that have been previously used to infer functions at gene levels [31] . Algorithms have been developed to assign isoform-level expression values [28] , [32]–[39] . They provide a resource for isoform-level features that can be used as input data to infer isoform functions . However , the fundamental challenge remains: a genome-wide set of ‘ground-truth’ annotations of functions at the isoform level is still lacking . Isoform functions have been computationally inferred through domains , binding regions and individual binding sites [40]–[46] . In widely used databases such as Gene Ontology [47] , [48] and KEGG [49] , biological functions are defined at the gene level . Under these circumstances , to predict whether a gene is related to a specific function , supervised learning algorithms will be deployed to derive a model from the genomic data we collect . For example , to predict ‘mitochondria biogenesis’-related genes , we need a ‘gold standard’ set of genes known to be related to mitochondria biogenesis , and find out how these genes differ from other genes in their expression patterns . Without an existing and comprehensive set of annotated isoforms , standard ‘supervised learning’ approaches are not applicable to predicting different functions for splice isoforms . We developed a generic framework that improves gene function prediction and provides information for isoform-level functions . Our framework intends to locate the ‘responsible’ set of isoform ( s ) of annotated genes of a specific function , through iteratively correcting the composition of this set to maximize its ‘discriminativity’ against the negatives . For example , for mitochondria biogenesis , we have a set of genes G+1 , G+2 , … , G+n annotated to this function ( positive genes ) . Each gene is considered as a bag of alternatively spliced isoforms . We then try to find out which isoform ( s ) of this positive set of genes can be selected to maximize the difference between them and the negative isoforms , i . e . , the Gik+ , where k is the gene index , and i is the isoform index . The isoforms in the Gik+ set are iteratively updated to maximize the similarity within them . The model derived from the Gik+ set ( instead of Gk+ ) is then used to classify whether an isoform has the biological function , mitochondria biogenesis , in this example . From this perspective , our problem is a multiple instance learning task [50]–[53] , in which each example considered positive with respect to certain property contains multiple discrete elements , of which at least one of them must be positive . We used an iterative algorithm to approximate the solution to the above optimization problem . We found that our algorithm can successfully capture the differential isoform functions as evidenced by prediction accuracy on multi-isoform genes as well as literature and experimental validation on isoforms that are drastically different in their assigned functions . Computationally predicting isoform functions or differentiating functions for isoforms of the same genes in a genome-wide manner using high-throughput genomic data has not been done prior to our work . This framework is also generic . It can be integrated into any basic machine learning algorithm such as logistic regression , random forests or deep learning , and can be readily extended to predict isoform functions in other organisms as well as other properties of isoforms such as phenotypes and interaction networks . Our prediction results are available to the user community online at http://guanlab . ccmb . med . umich . edu/isoPred/ .
We aim at predicting isoform functions while functional annotations have been based on genes . The model therefore needs to be learned at the isoform level rather than at the gene level ( see Figure S1 for a comparison between a traditional gene function learning problem and our isoform function prediction problem ) . Our core idea is to model the common patterns of a subset of isoforms across genes associated with a particular function , with the requirement that at least one of the isoforms of each positive gene must retain the common feature pattern . This is a multiple instance problem , the aim of which is to identify the hidden labels of the isoforms of the positively annotated genes and use these hidden labels to construct classification models to label additional isoforms . We used maximum margin-based classification as base-learner in this iterative process [54] , due to the success of SVM in the protein function prediction domain [9] . Such a framework could be integrated into any other machine learner , such as deep learning or Bayesian classification . To elucidate our algorithm , we will use the following definitions throughout the paper: Our algorithm aims to identify a subset of isoforms of the positive genes that maximizes the difference between them and the negative isoforms . Identifying the best combination of isoforms from the positive genes is difficult . The ideal solution requires excessive computational time . We therefore approximated the solution with an iterative algorithm ( Figure 1 ) . For the training set , in the first iteration , we assign every isoform of a positive gene to be positive . We then establish a classifier , with which we can go back to classify the training set . This classifier will assign some isoforms of the positive gene to be positive ( the ‘witnesses’ ) , and some to be negative , but at least one isoform of a positive gene must be positive . This subset of ‘witnesses’ is iteratively updated to maximize the inter-class distance . Using this assignment , a new classifier is established , which could again be used to classify the training set . We iterate this process to update the ‘witnesses’ . This procedure is repeated until convergence is achieved ( Figure 1 ) . Because of the properties and relationships between genes and isoforms , this Multiple Instance Learning ( MIL ) framework is suitable for our problem . MIL assumes that there are one or more positive instances in a positive bag , and , if we can identify these positive instances in positive bags , we can expect to have a better classifier by excluding remaining instances from the positive class . Biologically , if a gene is annotated to a function , one or more of its isoforms will be selected as witness ( es ) as the ones that are related to this function , and other isoforms will be marked as non-functional . By excluding these non-functional isoforms , the classifier is expected to be more accurate . To generate alternatively spliced transcript-level features , we collected transcriptomic data from public RNA-seq experiments ( Dataset S1 ) . These RNA-seq datasets cover a wide sampling of tissues and different experimental conditions , such as liver , brain , muscle and testis . The ENCODE RNA-seq data covering ovary , mammary gland , stomach , kidney , liver , lung , spleen , colon and heart , are also included in our data . Co-expression patterns in these data can be highly informative for co-functionality . We assigned isoform-level expression values for each of these experiments using the state-of-the-art tools [32] , [34] , [35] ( Figure 1 ) . Public RNA-seq datasets came from different experimental protocols and such information is not always recorded in databases in a standardized way . We filtered these datasets based on their quality and coverage ( see Methods ) . This data collection serves as our genomic data input . Essentially , each isoform was described by a vector of values specifying its normalized expression value in various RNA-seq samples . Gene Ontology is arranged in a hierarchy , where complex relationships exist between different terms . To test different variations of MIL and tune the parameters of our algorithms , we focused on a list of biological process terms that have been voted by biologists to be able to describe and cover different biological processes that are experimentally testable [55] . This list included 99 terms , with GO term size 20–300 . Two important parameters ( other than the standard SVM parameters ) in this algorithm are the proportion of positive isoforms to be labeled as positive in each iteration , and whether we should label the rest of isoforms in the positive bag negative or discard them . Therefore , we tested two basic formulations of the algorithm . The first approach tries to impute all non-witness instances in positive bags as negative instances and then considers the problem as a supervised learning problem . The second approach tries to identify a single witness from each positive bag which is responsible for the positive label . Then , a classifier is built based on these witnesses only , while other instances are dropped . SVM formulations of these two approaches are , respectively , mi-SVM and MI-SVM [51] . For the first approach , we envisioned that different ratio of instances can be retained as ‘witnesses’ and tested three different cutoffs ( Figure 2A–B ) . Ideally , testing of isoform function prediction should use isoform-level gold-standard functional annotation . However , such comprehensive functional annotation does not exist in any database ( if they did exist , regular supervised classification would be sufficient to predict isoform functions ) . Therefore , we first evaluated the performance of our algorithms at the gene level . The probability of each gene to be associated to a function is assigned with the maximum value of all its instances , under the assumption that the eventual gene function is carried out by at least one of its isoforms . For all methods and parameters tested , the algorithm converged within several iterations . Additionally , we found that different thresholds or methods resulted in relatively stable performance on the gene level . mi-SVM with 75% of all negative scores as the cutoff for defining ‘witnesses’ resulted in the highest AUC of 0 . 73 ( Figure 2C ) . Therefore , we used this method for inferring isoform functions and all the evaluation and validation below is based on this threshold . The performance obtained in the previous section is a mix between single-isoform genes and multiple-isoform genes . We hypothesize that , if our framework does bring in discriminative power at the isoform level , multi-isoform genes should be predicted with better accuracy than single-isoform genes for the same set of GO terms evaluated . Single-isoform genes include both real ones and those that are missed in the database [58] . In fact , although it has been estimated that 95% of the multi-exon genes in human have multiple isoforms [13] , only 13% of the genes are documented with validated multiple isoforms in NCBI . For these genes , the performance is expected to be poorer since the input features for each gene approximate the average of all its isoforms and our algorithm is not applicable to these genes in differentiating isoform functions . Therefore , we separately evaluated the performance of single-isoform genes and multiple-isoform genes . Two-fold cross-validation was carried out to evaluate our models to ensure there are sufficient genes in both the training and test sets for multi- and single-isoform genes . To make the comparison feasible , negatives in each group were randomly chosen to ensure that the ratio of positives to negatives is the same for multi- and single-isoform genes for the same GO term . In doing so , the AUC , AUPRC , precision at 1% recall and precision at 10% recall of each GO term are recalculated for both sets separately . These results are again organized into 5 groups , based on the number of positive genes in the test set for each GO term ( Figure 4 ) . We found consistently better performance for multi-isoform genes than single-isoform genes for the same GO term evaluated ( Figure 4; complete evaluation results are included in Dataset S3 ) . For multi-isoform genes , we acquired median AUCs of 0 . 68 , 0 . 70 , 0 . 70 , 0 . 73 and 0 . 76 for the five groups of multi-isoform genes , compared to median AUCs of the same groups of GO terms for single-isoform genes , 0 . 62 , 0 . 63 , 0 . 65 , 0 . 66 and 0 . 68 , which correspond to 57% , 56% , 34% , 47% and 40% improvements against the baseline ( 0 . 5 ) , respectively . Similar better performance is seen for AUPRC values . We observed an increase from 0 . 002 to 0 . 006 ( 162% improvement , baseline is 0 . 0009 ) , from 0 . 003 to 0 . 009 ( 168% improvement , baseline is 0 . 0015 ) , from 0 . 005 to 0 . 012 ( 152% improvement , baseline is 0 . 0021 ) , from 0 . 008 to 0 . 022 ( 156% improvement , baseline is 0 . 0034 ) , from 0 . 016 to 0 . 042 ( 154% improvement , baseline is 0 . 0072 ) . In fact , 71% of the 1591 GO terms with more than 20 positive genes and less than 300 positive genes have gained better performance for multi-isoform genes . This implies that our approach is effective in capturing the functionality of multi-isoform genes , which is robust regardless of the metrics used to quantify performance . For example , for GO terms GO:0000279 ( M phase ) , GO:0006952 ( defense response ) , GO:0006936 ( muscle contraction ) , GO:0007507 ( heart development ) , GO:0002252 ( immune effector process ) , GO:0003002 ( regionalization ) , precision at 10% recall increased from 0 . 121 to 0 . 571 ( baseline 0 . 0069 ) , from 0 . 051 to 0 . 500 ( baseline is 0 . 0103 ) , from 0 . 052 to 0 . 364 ( baseline is 0 . 0038 ) , from 0 . 045 to 0 . 389 ( baseline is 0 . 0097 ) , from 0 . 042 to 0 . 750 ( baseline is 0 . 0048 ) , from 0 . 074 to 0 . 320 ( baseline is 0 . 0089 ) , respectively . ( Figure S3 ) . Such precision improvement is robust to accuracy measurement across the entire precision-recall spectrum and consistent across a wide sampling of GO terms . For the five groups of GO terms , median of precision at 1% recall increased from 0 . 005 to 0 . 014 ( baseline is 0 . 0009 ) , from 0 . 008 to 0 . 031 ( baseline is 0 . 0015 ) , from 0 . 011 to 0 . 045 ( baseline is 0 . 0021 ) , from 0 . 016 to 0 . 083 ( baseline is 0 . 0034 ) , from 0 . 028 to 0 . 154 ( baseline is 0 . 0072 ) ( Figure 5 ) . This result indicates that our algorithm is more effective in predicting functions for genes with multiple isoforms than single-isoform genes , which is likely caused by the power of our algorithm in differentiating isoform functions for the same genes . We further considered two factors to validate the robustness of our algorithm . First , because the estimated expression levels are less reliable for genes that have low expression values , we tested whether our algorithm can predict the functions for low-expressing isoforms . We partitioned genes into three groups of equal numbers - the high , the medium and the low groups - based on their expression level averaged across all experimental conditions and evaluated prediction results separately for these groups to see if performance is affected by the overall expression level of genes . We showed that prediction performance is fairly robust to the genes' expression levels ( Figure 6A ) . Although the genes in the highest expression group do show better performance ( AUC = 0 . 79 ) , AUCs for the medium ( 0 . 69 ) and low ( 0 . 69 ) groups are comparable to the global accuracy ( Figure 6A ) , indicating that our method is applicable to genes that have relatively low expression levels . Secondly , we tested whether our good performance comes from the homolog gene pairs of which one member is grouped into the training set and the other member is grouped into the test set . This can potentially cause overfitting by information leak , which is a common phenomenon in all functional prediction algorithms , especially those based on sequence information . In order to test robustness of our method with respect to having homolog pairs in test and training sets , we partitioned all genes into equal size test and training sets by putting all genes in a homolog group as defined in Ensembl [59] together and re-evaluated the performance . We found that the performance is comparable to the case where paralog genes can occur in training and test sets ( Figure 6B ) . This suggests that our method utilizing RNA-seq data integration is also robust to the information leak from homologs . To generate the final isoform function prediction , we applied bootstrap bagging to assign scores for each isoform ( Figure 1 , http://guanlab . ccmb . med . umich . edu/isoPred/ for complete prediction results ) . The median AUC for the bootstrap result across all GO terms is the same as the cross-validation result , indicating the robust performance of bootstrapping . Essentially , genes are sampled with replacement to construct a training set , and the rest of the examples form an “out-of-bag” set . This process is iterated , and the eventual predictions are drawn from the median across all “out-of-bag” sets . Bootstrap bagging is suitable for our isoform function prediction task , where the numbers of positive and negative examples are highly imbalanced [56] . The robustness of bootstrap bagging has been tested for positive example set sizes ranging from less than 20 to more than 200 [56] , [60] , which is close to the GO term sizes for which we provide predictions in this study . Because each GO term has a different background probability for a gene to be associated with it , we calculated the fold change against background for each gene to be associated to a GO term . Indeed , it is pervasive that isoforms of the same gene are assigned with different confidence levels for the same function under consideration ( see Table S1 for a selection of examples and http://guanlab . ccmb . med . umich . edu/isoPred/ for complete prediction results ) . Because biological functions of transcripts are eventually delivered at the protein level , we hypothesized that the functional isoform ( s ) of a gene must be expressed at the protein level in the normal physiological condition . We therefore used splice variant protein expression data in normal mammary tissue to validate our predictions . Using our previously developed protocol [61] , we identified genes of which only one isoform is strongly expressed ( but the other isoforms are not detected ) . In this gene list , we focused on the ones that have a known specific function ( i . e . , the function has less than 300 genes annotated to them ) , but their isoforms are predicted with drastically different confidence levels to carry out this function . In total this resulted in 15 isoform groups . We could then check whether the predicted ‘responsible’ isoform ( s ) correspond to the expressed isoforms in normal breast tissues . We found a strong match between the expressed splice variant and the isoform ( s ) predicted to be responsible for the known specific function of the gene ( Table 1 and Dataset S4 ) . All the expressed isoforms are predicted with the highest value in at least one of the known functions , indicating the consistency between the predicted functional isoforms and the expressed isoform . For example , we predicted that NM_172745 . 3 of Tufm is responsible for its function translation ( 34 fold over background probability ) , while the other isoform , NM_001163713 . 1 , is much less likely to be functional ( 3 fold over background probability ) . Indeed , only NM_172745 . 3 is expressed in normal breast tissue . Among the six alternatively spliced isoforms of Tardbp , we correctly predicted that NM_145556 . 4 is the one responsible for its function in RNA splicing and stabilization; this variant is the only one identified in the proteomic sample . In fact , of the 9 functions annotated for Tardbp , NM_145556 . 4 is predicted with the highest value eight times . The majority of the exceptions occur for tissue or developmental-stage-specific GO terms , such as GO:0007507 heart development , GO:0035051 cardiac cell differentiation and GO: 0006936 muscle contraction . The predicted functional isoform is not the one identified in our proteomic sample , most likely because our sample is tissue-specific and normal and these functions might be carried out by other isoforms in other tissues or conditions . However , overall , the predicted functional isoforms are consistent with the ones we identified in our proteomic data , indicating that our algorithm can correctly identify the functional splice variants in normal conditions . Our algorithm is more than just identifying the ‘functional’ isoform of a gene . The power of our algorithm to predict the functional disparity between isoforms is further illustrated by isoforms that carry out different aspects of gene functions . It is relatively common that only one isoform is predicted to be responsible for one particular function of a gene , as described in the previous section . In this section , we focus on analyzing specific examples in which isoforms of a single gene are assigned with unrelated biological functions . CDKN2a is the only known example where alternative splicing results in different reading frames ( Figure 7A ) . Two genes ( XBP1 , GNAS1 ) produce alternate reading frames , but start with single transcript and therefore do not fall into the realm of alternative splicing . We predicted that the two isoforms of CDKN2a would carry distinct functions for the gene ( Figure 7B ) . NM_001040654 . 1 is predicted to be involved in apoptotic nuclear changes with a probability 74 times the background probability , while the probability for NM_009877 . 2 approximates background . On the other hand , NM_009877 . 2 is predicted to be involved in positive regulation of the transmembrane receptor protein serine/threonine kinase signaling pathway ( 3 times background ) , while NM_001040654 . 1 is not . Because crystallized protein structures are available for both proteins in the human but not in the mouse , we used I-TASSER , the state-of-the-art protein structure prediction algorithm [62] , to model the 3-D protein structures of the two isoforms ( Figure 7C–D ) . Although the translated products of NM_001040654 . 1 ( 168 aa ) and NM_009877 . 2 ( 169 aa ) are almost the same lengths , the transcript sequences are in different open reading frames . This resulted in five ankyrin repeats in the NM_001040654 . 1 ( Figure 7C ) , compared to a cyclin-dependent kinase inhibitor N-terminus domain in NM_009877 . 2 ( Figure 7D ) . The drastically different 3-D structures support the potential disparate functions predicted by our algorithm . The function predictions for NM_009877 . 2 and NM_001040654 . 1 , which we made by mining only large-scale public RNA-seq datasets , are consistent with the two distinct biological roles of the two isoforms . NM_009877 . 2 is an inhibitor of CDK4 kinase , a member of the Ser/Thr protein kinase family , directly supporting its role in GO:0071900 , regulation of protein serine/threonine kinase activity . NM_001040654 . 1 encodes an alternate open reading frame ( ARF ) that generates a protein structurally unrelated to NM_009877 . 2 . This protein enhances p53-dependent transactivation and apoptosis [63] , supporting its role in apoptotic nuclear changes . Interestingly , although coding for structurally dissimilar proteins , both isoforms share a common functionality in cell cycle G1 control [64] . This shared functionality is correctly predicted by our algorithm; for regulation of G1/S transition of mitotic cell cycle ( GO:2000045 ) : NM_001040654 . 1 has a probability 3 . 5 times the background probability , and NM_009877 . 2 has a probability 2 . 4 times that of background . The CDKN2a example involves isoforms of drastically different protein domains . We used the isoforms of ANXA6 ( NM_013472 . 4 and NM_001110211 . 1 ) to validate our model in predicting isoforms of very similar structure . The only difference between the two isoforms at the protein sequence level is the presence of six residues in the longer NM_013472 . 4 ( ‘VAAEIL’ , 525–530 ) which are missing in NM_001110211 . 1 . The three-dimensional structures of the translated sequences of these isoforms were published by some authors of this paper [65] . Although the global topology of the I-TASSER models for the two isoforms of ANXA6 is almost identical ( with RMSD = 0 . 38 Å and TM-score = 0 . 99 ) , there is an obvious structural variation identified by TM-align [66] . The positions of Thr-535 and Ser-537 in NM_013472 . 4 compared to NM_001110211 . 1 make NM_013472 . 4 more likely to undergo phosphorylation [65] . The fold changes on GO-terms related to phosphorylation by our function prediction algorithm supported the conclusion from the structural comparisons . The fold change for peptidyl-serine phosphorylation for NM_013472 . 4 was 3 . 5 compared to 1 . 9 for NM_001110211 . 1 . Re-searching the mass-spectrometric data with phosphorylation on Serine or Threonine ( Phospho ( S ) and Phospho ( T ) ) as potential residue modification , yielded a peptide ‘DQAQEDAQVAAEILEIADTPSGDKTSLETR’ ( found only in NM_013472 . 4 ) with 3281 . 506 daltons as the mh ( calculated peptide mass plus a proton ) indicating potential phosphorylation [67] . In contrast , the peptide that matched the spliced region in NM_001110211 . 1 ( residues ‘VAAEIL’ are missing in this peptide ) did not show any phosphorylation . These observations further supported our predictions . In addition , the overall function enrichment showed the smaller isoform , NM_001110211 . 1 as involved in biological processes related to cell adhesion and cell migration , whereas the longer form is predicted to be involved in localization . It is important to emphasize the fact that our computational predictions based on RNA-seq data alone were able to pick up the differences between these two isoforms with 99% sequence identity and predict distinct functions for the isoforms . These findings suggest that our approach can solve the pressing need of isoform function differentiation , which would be invaluable for a better understanding of the diversity of functions created by alternative splicing of a limited set of genes .
Gene functions are delivered through alternatively spliced transcript isoforms that encode proteins of different functions . It is highly beneficial that the investigation of functions is carried out at the isoform level . From this point of view , the standard gene function prediction paradigm has a major drawback in that it considers a gene as one single entity without differentiating its isoforms . The availability of transcript-level expression data from RNA-seq provides a rich resource for addressing this drawback . However , algorithmically , any supervised learning algorithm developed for gene function prediction cannot be directly applied to isoform function prediction because of the lack of isoform-level , ‘ground-truth’ functional annotations . To address this challenge , we developed an iterative algorithm that predicts functions at the individual isoform level by conceptualizing a gene as a ‘bag’ of isoforms of potentially different functions . Our key idea is to iteratively extract the common pattern of a subset of isoforms across the positive genes of the function under investigation , aiming at maximizing the coherence within this subset of isoforms and the discriminative power against the other ‘negative’ genes ( genes not related to the specific function under consideration ) . Through experimental validation , we demonstrated that our approach in combination with publicly available RNA-seq data is capable of differentiating isoform functions , promising better and deeper understanding of gene functions . RNA-seq data are the richest resource for genome-wide , isoform-level data so far . But the basic concept is extendable to other large-scale datasets providing isoform-level information , such as protein domain data and post-translational regulation datasets . These datasets are not included in this study due to their strong overlap with the ‘Gold-Standard’ gene ontology annotations , which might lead to a circularity problem in evaluating our algorithms . Furthermore , our study only focuses on the base learner SVM . However , our approach is highly extendable to other modeling methods , such as logistical regression and random forests . Our study is limited to the incomplete isoform catalog maintained by NCBI , but it can be readily updated whenever the genome annotation of isoforms is updated . Additionally , alternatively spliced isoforms often show tissue-specific expression and functions [23] , [27] , [68]–[71] . Our generic algorithm does not yet take the tissue-specific functionality into consideration . We expect that more accurate and biologically meaningful isoform function prediction could be achieved if tissue specificity were taken into account . As a result , our validation carried out in breast tissue is only used to validate the ‘generic’ functions of the isoforms . Recent studies found that the same principal isoform is often present in different tissues [72]–[75] . We expect that tissue-specific functions can be validated in corresponding tissues when these tissue-specific predictions can be made . Our study is further limited by the current technology to assign isoform-level expression values , as well as the differential capability between platforms for capturing isoform-specific expression . We used Cufflinks in the Tuxedo suite [32] , one of the state-of-the-art algorithms , to estimate the isoform-level read counts and achieved good performance . However , if more advanced algorithms are developed , our algorithm could directly utilize the estimates from those algorithms and generate isoform function predictions . Our approach represents a novel and generic strategy to look at gene functions at a higher resolution . Cross-validation , literature , and experimental analysis of proteomic data provided evidence that our algorithm is powerful in differentiating isoform functions . Broadly speaking , the genomic data integration field typically relies on the supervised learning concept , which cannot generate predictions for spliced isoforms , e . g . , predicting gene-disease association and gene regulatory networks . We envision that similar concepts will be developed for generating isoform-level models for these prediction tasks .
We downloaded 811 RNA-seq experiments for the mouse from the NCBI sequence read archive ( SRA ) database as of May 1 , 2012 [30] . These datasets represent different conditions and tissues . Heterogeneity of the datasets allowed us to look at the isoform expression variations across different conditions and tissues . Because datasets are heterogeneous in terms of library preparation procedures and sequencing platform ( Dataset S1 ) , we adopted the following processing and filtering pipeline to ensure that all datasets included in the final predictions have sufficient coverage . NCBI build 37 . 2 reference genome was downloaded from the TopHat homepage , and BOWTIE2 [76] index files were created by bowtie2-build software . For each RNA-seq dataset , short reads were aligned against the NCBI Mus musculus reference genome ( Build 37 . 2 ) using TopHat v2 . 0 . 051 [32] , [34] . The reference GTF annotation file from NCBI ( build 37 . 2 , downloaded from TopHat homepage ) was given to TopHat and the no-novel-juncs option was used . With this option , TopHat creates a database of splice junctions indicated in the supplied GTF file and maps the previously unmapped reads against the database of these junctions to create an estimate of isoform expression levels . Then we used Cufflinks v2 . 0 . 0 [32] to measure the relative abundances of the transcripts using normalized RNA-seq fragment counts [32] , [35] . The unit of expression levels is Fragments Per Kilobase of exon per Million fragments mapped ( FPKM ) . To ensure data quality and overall coverage , we removed those experiments with less than 10 million reads or with less than 50% reads being successfully mapped to the genome . The above procedure resulted in 365 experiments used in our study . Genes detected in less than half of the experiments were removed . After filtering these poorly-covered genes , there are 19209 genes left with a total of 24274 isoforms . Distribution of the number of isoforms per gene is included in Figure S4 . FPKM values were log2-transformed; missing values were approximated with a value of “-15” . We constructed gold standard gene functions using the Gene Ontology ( GO ) database [48] , [77] . For each biological process term , we treated the genes which are annotated to that GO term and any of its descendent terms as positives and others as negatives . We maintained all GO evidence codes . The sources of these annotations include ( 1 ) hand annotation from primary literature , ( 2 ) electronic annotation based on gene name and symbols , ( 3 ) annotation from SwissProt keywords and ( 4 ) Enzyme Commision ( EC ) numbers [78] . These annotation sources are reasonably accurate for our analysis . As stated in the Results section , the isoform function prediction problem is a Multiple Instance Learning problem . Since the introduction of MIL by [79] for the drug activity prediction , several methods to solve this problem have been proposed in the literature . In [80] , the authors introduced the concept of Diversity Density , whose aim is to find a point in feature space that has a high Diverse Density . This means high density of instances from positive bags and low density of instances from negative bags . Additionally , Ray and Page [81] proposed the method multiple-instance regression . Their algorithm assumed that each bag has a witness instance and treated it as a missing value; then the EM ( Expectation-Maximization ) method was used to learn the witness instances and do the regression simultaneously . Ramon et al . [82] utilized the Neural Network technique on Multiple Instance Learning and proposed the Multiple Instance Neural Network . Finally , Andrews et al . [51] applied the Support Vector Machines to Multiple Instance Learning . In this paper , we chose to use a method that utilizes the Support Vector Machine . Two different approaches have been proposed to solve the MIL problem using SVM . The first tries to impute all non-witness instances in positive bags as negative examples and then considers the problem as a supervised learning problem . The second tries to identify a single witness from each positive bag which is responsible for positive label . Then , a classifier is built based on these witnesses only , while other instances are dropped out of the classification process . SVM formulations of these two approaches are labeled mi-SVM and MI-SVM by Andrews et al . [51]; we implemented and tested several alternatives of these algorithms . Without loss of generality , we assume that the ith gene with m isoforms is denoted as . The corresponding class label yi is set to 1 if the gene is annotated to the function under consideration and 0 if not . For each gene , we hypothesize that if a gene is annotated to a function , at least one of its isoforms should be annotated to the function; if a gene is a negative example used in training , none of its isoforms can be annotated to the function , i . e . , ( 1 ) Then , the mi-SVM formulation of MIL can be written as follows ( 2 ) In the standard classification setting , the labels yij of the isoform xij would be given; however , in equation ( 2 ) labels of isoforms that belong to a positive gene are treated as unobserved hidden integer variables . Therefore , the soft-margin criterion is maximized jointly over hyperplanes and over all possible label assignments . The algorithm is looking for a separating hyperplane such that all isoforms of negative genes are in the negative half-space , whereas there is at least one isoform from every positive gene in the positive half-space . Meantime , the margin is maximized with respect to the selected labels . Alternatively , in the MI-SVM formulation , the definition of margins is extended to bag-level . Margin of a bag with respect to a separating hyperplane can be defined as the maximum margin of its instances . In the case of positive bags , bag margin is defined by the most positive instance , whereas , for negative bags , the bag margin is defined by the least negative instance . Using this definition of bag margin , the MI-SVM formulation can be written as follows: ( 3 ) where yi takes the form and s ( i ) is the selector variable that denotes the isoform selected as witness from each positive gene . Note that , for the mi-SVM formulation , every instance in a positive bag has an effect on the margin maximization equation , whereas , in the MI-SVM formulation , only one instance per positive bag is taken into account , because this instance alone will determine the margin of the bag . Finding the optimum solution to ( 2 ) and ( 3 ) is a combinatorial optimization problem , which cannot be found efficiently with the state-of-the-art tools . Therefore , we approximated the solution by using the following optimization heuristics which is proposed by Andrews et al . [51] . Both formulations of MIL explained above can be considered as mixed-integer problems . In the mi-SVM formulation , instance margin is maximized over hidden labels of instances in positive bags , whereas , in the MI-SVM formulation , bag margin is maximized over selector variable , which selects a single witness from every positive bag . Optimization heuristic uses the fact that , given these integer variables , the problem reduces to a quadratic programming problem which can be solved exactly . The optimization heuristic includes two steps: ( i ) for a set of given integer variables ( i . e . hidden labels in mi-SVM and selector variable in MI-SVM ) , solve the soft-margin maximization problem and find the optimal separating hyper-plane , ( ii ) for a given separating hyper-plane , update all integer variables so that they maximize the objective locally . These two steps are run iteratively until integer variables are not updated anymore in step ( ii ) . The following workflow explains this optimization heuristic further in detail for both formulations . For every function , we need to assign each isoform a score no matter whether the gene that the isoform belongs to has an annotation or not . We therefore used bootstrap bagging to estimate the probability that an isoform is associated with a specific biological process . Essentially all genes are sampled with replacement ( 0 . 632 bootstrap ) to construct a training set . The scores of the held-out set are recorded and the process is iterated 30 times . For each isoform , the final score is assigned with the median across all iterations . Functions of splice variants must eventually be delivered at that protein level . To test whether the predicted differential functions are correct , we compiled the data from LC-MS/MS of normal mammary tissue [83] . The original study reported that normal tissues were harvested from 5 normal mice , processed into tissue lysates and pooled . The pooled sample was digested by trypsin for mass spectrometric analysis . The mzXML files were searched against our modified ECgene database for alternative splice variant analysis using X ! Tandem [84] . All prediction results are stored in MySQL databases and delivered through a searchable website: http://guanlab . ccmb . med . umich . edu/isoPred . | In mammalian genomes , a single gene can be alternatively spliced into multiple isoforms which greatly increase the functional diversity of the genome . In the human , more than 95% of multi-exon genes undergo alternative splicing . It is hard to computationally differentiate the functions for the splice isoforms of the same gene , because they are almost always annotated with the same functions and share similar sequences . In this paper , we developed a generic framework to identify the ‘responsible’ isoform ( s ) for each function that the gene carries out , and therefore predict functional assignment on the isoform level instead of on the gene level . Within this generic framework , we implemented and evaluated several related algorithms for isoform function prediction . We tested these algorithms through both computational evaluation and experimental validation of the predicted ‘responsible’ isoform ( s ) and the predicted disparate functions of the isoforms of Cdkn2a and of Anxa6 . Our algorithm represents the first effort to predict and differentiate isoforms through large-scale genomic data integration . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [] | 2013 | Systematically Differentiating Functions for Alternatively Spliced Isoforms through Integrating RNA-seq Data |
Hierarchical organization—the recursive composition of sub-modules—is ubiquitous in biological networks , including neural , metabolic , ecological , and genetic regulatory networks , and in human-made systems , such as large organizations and the Internet . To date , most research on hierarchy in networks has been limited to quantifying this property . However , an open , important question in evolutionary biology is why hierarchical organization evolves in the first place . It has recently been shown that modularity evolves because of the presence of a cost for network connections . Here we investigate whether such connection costs also tend to cause a hierarchical organization of such modules . In computational simulations , we find that networks without a connection cost do not evolve to be hierarchical , even when the task has a hierarchical structure . However , with a connection cost , networks evolve to be both modular and hierarchical , and these networks exhibit higher overall performance and evolvability ( i . e . faster adaptation to new environments ) . Additional analyses confirm that hierarchy independently improves adaptability after controlling for modularity . Overall , our results suggest that the same force–the cost of connections–promotes the evolution of both hierarchy and modularity , and that these properties are important drivers of network performance and adaptability . In addition to shedding light on the emergence of hierarchy across the many domains in which it appears , these findings will also accelerate future research into evolving more complex , intelligent computational brains in the fields of artificial intelligence and robotics .
Hierarchy is an important organizational property in many biological and man-made systems , ranging from neural [1 , 2] , ecological [3] , metabolic [4] , and genetic regulatory networks [5] , to the organization of companies [6] , cities [7] , societies [8] , and the Internet [9 , 10] . There are many types of hierarchy [11–13] , but the one most relevant for biological networks [14 , 15] , especially neural networks [1 , 2 , 16] , refers to a recursive organization of modules [10 , 13] . Modules are defined as highly connected clusters of entities that are only sparsely connected to entities in other clusters [17 , 19 , 20] . Such hierarchy has long been recognized as a ubiquitous and beneficial design principle of both natural and man-made systems [14] . For example , in complex biological systems , the hierarchical composition of modules is thought to confer greater robustness and adaptability [1 , 2 , 16 , 21] , whereas in engineered designs , a hierarchical organization of simple structures accelerates the design , production , and redesign of artifacts [19 , 22 , 23] . While most studies of hierarchy focus on producing methods to quantify it [4 , 9 , 11 , 14 , 24–28] , a few have instead examined why hierarchy emerges in various systems . In some domains , the emergence of hierarchy is well understood; e . g . , in complex systems , such as social networks , ecosystems , and road networks , the emergence of hierarchy can be explained by resource constraints or by local decisions and interactions [3 , 15 , 29–31] . But , in biological systems , where the evolution of hierarchy is shaped by natural selection , why hierarchy evolves , and whether its evolution is due to direct or indirect selection , is an open and interesting question [3 , 15 , 32] . Non-adaptive theories state that the hierarchy in some , but not all , types of biological networks may emerge as a by-product of random processes [29] . Most adaptive explanations claim that hierarchy is directly selected for because it confers evolvability [33 , 34] , which is the ability of populations to quickly adapt to novel environments [35] . Yet in computational experiments that simulate natural evolution , hierarchy rarely , if ever , evolves on its own [36–38] , suggesting that alternate explanations are required to explain the evolutionary origins of hierarchy . Moreover , even if hierarchy , once present , is directly selected for because of the evolvability it confers , explanations are still required for how that hierarchy emerges in the first place . In this paper we investigate one such hypothesis: the existence of costs for network connections creates indirect selection for the evolution of hierarchy . This hypothesis is based on two lines of reasoning . The first is that hierarchy requires a recursive composition of modules [10] , and the second is that hierarchy includes sparsity . A recent study demonstrated that both modularity and sparsity evolve because of the presence of a cost for network connections [17] . Connection costs may therefore promote both modularity and sparsity , and thus may also promote the evolution of hierarchy . It is realistic to incorporate connection costs into biological network models because it is known that there are costs to create connections , maintain them , and transmit information along them [39–41] . Additionally , evidence supports the existence of a selection pressure in biological networks to reduce the net cost of connections . While it remains an open question how strong such selection is [41] , multiple studies have shown that biological neural networks , which are hierarchical [1 , 2] , have been organized to reduce their amount of wiring by having fewer long connections and by locating neurons optimally to reduce the wiring between them [40 , 42–44] . A relationship between hierarchy and connection costs can also be observed in a variety of different man-made systems . For example , very large scale integrated circuits ( VLSIs ) , which are designed to minimize wiring , are hierarchically organized [16] . In organizations such as militaries and companies , a hierarchical communication model has been shown to be an ideal configuration when there is a cost for communication links between organization members [45] . That connection costs promote hierarchy in human-made systems suggests that the same might be true in evolved systems . Here we test that hypothesis in computational simulations of evolution and our experiments confirm that hierarchy does indeed evolve when there is a cost for network connections ( Fig 1 ) . We also investigate the hypothesis that hierarchy confers evolvability , which has long been argued [1 , 2 , 16 , 46] , but has not previously been extensively tested [16] . Our experiments confirm that hierarchical networks , evolved in response to connection costs , exhibit an enhanced ability to adapt . Experimentally investigating the evolution of hierarchy in biological networks is impractical , because natural evolution is slow and it is not currently possible to vary the cost of biological connections . Therefore , we conduct experiments in computational simulations of evolving networks . Computational simulations of evolution have shed substantial light on open , important questions in evolutionary biology [47–49] , including the evolution of modularity [17 , 18 , 20 , 50–52] , a structural property closely related to hierarchy . In such simulations , randomly generated individuals recombine , mutate , and reproduce based on a fitness function that evaluates each individual according to how well it performs a task . The task can be analogous to efficiently metabolizing resources or performing a required behavior . This process of evolution cycles for a predetermined number of generations . We evolved computational abstractions of animal brains called artificial neural networks ( ANNs ) [53 , 54] to solve hierarchical Boolean logic problems ( Fig 2A ) . In addition to abstracting animal brains , ANNs have also been used as abstractions of gene regulatory networks [55] . They abstract both because they sense their environment through inputs and produce outputs , which can either be interpreted as regulating genes or moving muscles ( Methods ) . In our experiments , we evolve the ANNs with or without a cost for network connections . Specifically , the experimental treatment selects for maximizing performance and minimizing connection costs ( performance and connection cost , P&CC ) , whereas the control treatment selects for performance only ( performance alone , PA ) . Following previous work on the evolution of modularity [17 , 51] , in our default treatments the evolving networks are layered , feed-forward ( i . e . connections are allowed only between neurons in one layer and neurons in the next layer ) , and have eight inputs and a single output ( Methods ) . During evaluation , each network is tested on all possible ( 256 ) input patterns of zeros and ones , and the network’s output is checked against a hierarchical Boolean logic function provided with the same input ( Fig 2A and Table 1 ) . An ANN output ≥0 is considered True and an output <0 is considered False . A network’s performance ( fitness ) is its percent of correct answers over all input patterns .
On the main experimental problem ( Fig 2A ) , the addition of a connection cost leads to the evolution of significantly more hierarchical networks ( Fig 2B and 2G ) . Confirming previous findings on different problems [17 , 38] , the addition of a connection cost also significantly increases modularity ( Fig 2C and 2G ) and reduces the number of generations required to evolve a solution ( Fig 2D ) . Importantly , while final performance levels for the performance and connection cost ( P&CC ) treatment are similar to those of the performance alone ( PA ) treatment , there is a qualitative difference in how the networks solve the problem . P&CC networks exhibit functional hierarchy in that they solve the overall problem by recursively combining solutions to sub-problems ( Fig 2F ) , whereas the PA networks tend to combine all input information in the first layer and then process it in a monolithic fashion ( Fig 2E ) . Such functional hierarchy can be quantified as the percent of sub-problems a network solves ( e . g . the AND and XOR gates in Fig 2A ) . A sub-problem is considered solved if , for all possible network inputs , there exists a threshold such that a neuron in a network exists that outputs an above-threshold value whenever the answer to the sub-problem is True , and a sub-threshold value when the answer is False , or vice-versa ( Methods ) . This measure reveals that evolved P&CC networks solve significantly more sub-problems than their PA counterparts ( Fig 2G , p < 2 . 6 × 10−16 via Fisher’s exact test ) . To further investigate how the ability to solve sub-problems is related to hierarchy and modularity , we plotted the percent of sub-problems solved vs . both hierarchy and modularity ( Fig 3 ) . The plots show a significant , strong , positive correlation between the ability to solve sub-problems and both hierarchy and modularity . It has been hypothesized that one advantage of network hierarchy is that it confers evolvablity [1 , 2 , 16] . We test this hypothesis by first evolving networks to solve one problem ( the base environment ) and then evolving those networks to solve a different problem ( the target environment ) . To isolate evolvability , we keep initial performance equal by taking the first 30 runs of each treatment ( PA and P&CC ) that evolve a perfectly-performing network for the base environment [17] . Each of these 30 networks then seeds 30 runs in the target environment ( for 900 total replicates per treatment ) . The base and target problems are both hierarchical and share some , but not all , sub-problems ( Fig 4 ) . Evolution in the target environment continues until the new problem is solved or 25000 generations elapse . We quantify evolvability as the number of generations required to solve the target problem [17] . We performed three such experiments , each with different base and target problems . In all experiments , P&CC networks take significantly fewer generations to adapt to the new environment than PA networks . They also solve significantly more of the target problem’s sub-problems ( Fig 4 ) . One possible reason for the fast adaptation of P&CC networks is that their modular structure allows solutions to sub-problems to be re-used in different contexts [17] . A hierarchical structure may also be beneficial if both problems are hierarchical , even if the computation at points in the structure is different . For example , modules that solve XOR gates can quickly be rewired to solve EQU gates ( Fig 4A ) . Another reason for faster P&CC adaptation could be that these networks are sparser , meaning that fewer connections need to be optimized . To further understand why connection costs increase hierarchy , we generated 20000 random , valid networks for each number of connections a network could have ( Methods ) . A network is valid if it has a path from each of the input nodes to the output node . The networks were neither evolved nor evaluated for performance . Of these networks , those that are low-cost tend to have high hierarchy , and those with a high cost have low hierarchy ( Fig 5 , left ) . This inherent association between low connection costs and high hierarchy suggests why selecting for low-cost networks promotes the evolution of hierarchical networks . It also suggests why networks evolve to be non-hierarchical without a pressure to minimize connection costs . Indeed , most evolved PA networks reside in the high-cost , low-hierarchy region , whereas all P&CC networks occupy the low-cost , high-hierarchy region ( Fig 5 left ) . Similarly , there is also an inverse relationship between cost and modularity for these random networks ( Fig 5 , right ) , as was shown in [17] . All evolved P&CC networks are found in the low-cost , high-modularity region; PA networks are spread over the low-modularity , high-cost region ( Fig 5 , right ) . As Figs 3 and 5 suggest , network modularity and hierarchy are highly correlated ( S12 Fig , Pearson’s correlation coefficient = 0 . 92 , p < 0 . 00001 based on a t-test with a correlation of zero as the null hypothesis ) . It is thus unclear whether hierarchy evolves as a direct consequence of connection costs , or if its evolution is a by-product of evolved modularity . To address this issue we ran an additional experiment where evolutionary fitness was a function of networks being low-cost ( as in the P&CC treatment ) , high-performing ( as in all treatments ) , and non-modular ( achieved by selecting for low modularity scores ) . We call this treatment P&CC-NonMod . The results reveal that P&CC-NonMod networks have the same low level of modularity as PA networks ( Fig 6A and 6B , p = 0 . 23 ) , but have significantly higher hierarchy ( Fig 6A and 6C , p < 0 . 00001 ) and solve significantly more sub-problems than PA networks ( Fig 6D ) . These results reveal that , independent of modularity , a connection cost promotes the evolution of hierarchy . Additionally , P&CC-NonMod networks are significantly more evolvable than PA networks ( Fig 6E–6G ) , revealing that hierarchy promotes evolvability independently of the known evolvability improvements caused by modularity [17] . To gain better insight into the relationship between modularity , hierarchy , and performance , we searched for the highest-performing networks at all possible levels of modularity and hierarchy . We performed this search with the multi-dimensional archive of phenotypic elites ( MAP-Elites ) algorithm [56 , 57] . The results show that networks with the same modularity can have a wide range of different levels of hierarchy , and vice-versa , which indicates that these network properties can vary independently ( Fig 7 ) . Additionally , the high-hierarchy , high-modularity region , in which evolved P&CC networks reside , contains more high-performing solutions than the low-hierarchy , low-modularity region where PA networks reside ( Fig 7 ) , suggesting an explanation for why P&CC networks find high-performing solutions faster ( Fig 2D ) . To test the generality of our hypothesis , we ran three additional experiments . First , we repeated the main experiment with two different Boolean-logic problems that each have different logic gates , but have the same number of inputs and outputs . For all problem variants , P&CC networks are significantly more hierarchical , modular , and solve more sub-problems than PA networks ( Fig 8 ) . The P&CC treatment also evolved high-performing networks significantly faster on all of these additional problems ( S13 Fig ) . It is known that biological networks , such as neural systems , have long-range connections [58] . The equivalent in our model are connections that can skip any number of layers . Because such connections could affect the evolution of network modularity and hierarchy , we conducted a second experiment that investigates whether the results reported above hold when connections between any layers are allowed ( Methods ) . The results are qualitatively unchanged: a connection cost still promoted the evolution of modularity and hierarchy ( S14 Fig and S15 Fig ) . We also conducted a third experiment similar to the main experiment , but with a problem that has eight inputs and two outputs ( see S17A Fig ) . This experiment investigates whether hierarchy and modularity evolve only in the presence of problems with many inputs and a single output ( as in the main experimental problem: Fig 2A ) . Again the results are qualitatively unchanged ( S16 Fig and S17 Fig ) . Additionally , our results are qualitatively unchanged when measuring hierarchy via a different , independently created hierarchy metric , specifically that of Czégel and Palla [83] ( Methods ) . As with the default metric developed by Mones et al . [11] , under the alternate metric , final-generation P&CC networks are significantly more hierarchical than final-generation PA networks in the main treatment ( S18A Fig , p < 0 . 001 ) , the treatment that allows connections to skip layers ( S18B Fig , p < 0 . 001 ) , and the two-output treatment ( S18C Fig , p < 0 . 001 ) . This new metric also enabled us to verify that P&CC networks are significantly more hierarchical than PA networks on a problem with three outputs ( S18D Fig , p < 0 . 001 , problem shown in S19 Fig ) ; the Mones et al . hierarchy metric returned pathological results for the three-output case . Overall , the results from all of these additional experiments suggest that the hypothesis that connection costs promote the evolution of modularity and hierarchy is robust to varying key assumptions of the default model .
The evolution of hierarchy is likely caused by multiple factors . These results suggest that one of those factors is indirect selection to reduce the net cost of network connections . Adding a cost for connections has previously been shown to evolve modularity [17 , 38]; the results in this paper confirm that finding and further show that a cost for connections also leads to the evolution of hierarchy . Moreover , the hierarchy that evolves is functional , in that it involves solving a problem by recursively combining solutions to sub-problems . It is likely that there are other forces that encourage the evolution of hierarchy , and that this connection cost force operates in conjunction with them; identifying other drivers of the evolution of hierarchy and their relative contributions is an interesting area for future research . These results also reveal that , like modularity [17 , 51] , hierarchy improves evolvability . While modularity and hierarchy are correlated , the experiments we conducted where we explicitly select for networks that are hierarchical , but non-modular , reveal that hierarchy improves evolvability even when modularity is discouraged . An additional factor that is present in modular and hierarchical networks is sparsity , a term meaning that only a few connections exist of the total that could . It is possible that this property explains some or even all of the evolvability benefits of modular , hierarchical networks . Future work is needed to address the difficult challenge of experimentally teasing apart these related properties . Our study suggests a biologically plausible selection cost that leads to the evolution of sparse , hierarchical networks . Our finding is thus consistent with a previous study that found that sparse networks tend to be more hierarchical [29] . However , that previous work did not investigate whether sparsity tends to evolve on its own or which forces can encourage its evolution , nor did it investigate whether sparsity leads to hierarchy in an evolutionary context . Thus , our work both reaffirms the relationship between sparsity and hierarchy , and provides insight into their evolutionary origins . As has been pointed out for the evolution of modularity [17] , even if hierarchy , once present , is directly selected for because it increases evolvability , that does not explain its evolutionary origins , because enough hierarchy has to be present in the first place before those evolvability gains can be selected for . This paper offers one explanation for how sufficient hierarchy can emerge in the first place to then provide evolvability ( or other ) benefits that can be selected for . This paper shows the effect of a connection cost on the evolution of hierarchy via experiments on many variants of one class of problem ( hierarchical logic problems with many inputs and one output ) . In future work it will be interesting to test the generality of these results across different classes of problems , including non-hierarchical problems . The data in this paper suggest that a connection cost will always make it more likely for hierarchy to evolve , but it remains an open , interesting question how wide a range of problems hierarchy will evolve on even with a connection cost . In addition to shedding light on why biological networks evolve to be hierarchical , this work also lends additional support to the hypothesis that a connection cost may also drive the emergence of hierarchy in human-constructed networks , such as company organizations [45] , road systems [59] , and the Internet [9] . Furthermore , knowing how to evolve hierarchical networks can improve medical research [60] , which benefits from more biologically realistic , and thus hierarchical , network models [61 , 62] . The ability to evolve hierarchy will also aid fields that harness evolution to automatically generate solutions to challenging engineering problems [63 , 64] , as it is known that hierarchy is a beneficial property in engineered designs [19 , 22] . In fact , artificial intelligence researchers have long sought to evolve computational neural models that have the properties of modularity , regularity , and hierarchy [19 , 65–68] , which are key enablers of intelligence in animal brains [19 , 69–71] . It has recently been shown that combining techniques known to produce regularity [72] with a connection cost produces networks that are both modular and regular [38] . This work suggests that doing so can produce networks that have all three properties , a hypothesis we will confirm in future work . Being able to create networks with all three properties will both improve our efforts to study the evolution of natural intelligence and accelerate our ability to recreate it artificially .
There were 30 trials per treatment . Each trial is an independent evolutionary process that is initiated with a different random seed , meaning that the sequence of stochastic events that drive evolution ( e . g . mutation , selection ) are different . Each trial lasted 25000 generations and had a population size of 1000 . Unless otherwise stated , all analyses and visualizations are based on the highest-performing network per trial ( ties are broken randomly ) . The test of statistical significance is the Mann-Whitney-Wilcoxon rank-sum test , unless otherwise stated . We report and plot medians ± 95% bootstrapped confidence intervals of the medians , which are calculated by re-sampling the data 5000 times . For visual clarity we reduce the re-sampling noise inherent in bootstrapping by smoothing confidence intervals with a median filter ( window size of 101 ) . Networks evolve via a multi-objective evolutionary algorithm called the Non-dominated Sorting Genetic Algorithm version II ( NSGA-II ) , which was first introduced in [73] . While the original NSGA-II weights all objectives equally , to explore the consequence of having the performance objective be more important than the connection cost objective , Clune et al . [17] created a stochastic version of NSGA-II ( called probabilistic NSGA-II , or PNSGA , implemented via the Sferesv2 framework [84] ) where each objective is considered for selection with a certain probability ( a detailed explanation can be found in [17] and Supp . S21A Fig ) . Specifically , performance factors into selection 100% of the time and the connection cost objective factors in p percent of the time . Preliminary experiments for this paper demonstrated that values of p of 0 . 25 , 0 . 5 , and 1 . 0 led to qualitatively similar results . However , for simplicity and because the largest differences between P&CC and PA treatments resulted from p = 1 , we chose that value as the default for this paper . Note that when p = 1 , NSGA-II and PNSGA are identical . Evolutionary algorithms notoriously get stuck in local optima ( locally , but not globally , high-fitness areas ) , in part because limited computational resources require smaller population sizes than are found in nature [64] . To make computational evolution more representative of natural evolutionary populations , which exhibit more diversity , we adopt the common technique of promoting behavioral diversity [17 , 64 , 74–77 , 79] by adding another independent objective that rewards individuals for behaving differently than other individuals in the population . This diversity objective factored into selection 100% of the time . Preliminary experiments confirmed that this diversity-promoting technique is necessary . Without it , evolution does not reliably produce functional networks in either treatment , as has been previously shown when evolving networks with properties such as modularity [17 , 38] . To calculate the behavioral diversity of a network , for each input we store a network’s output ( response ) in a binary vector; output values > 0 are stored as 1 and 0 otherwise . How different a network’s behavior is from the population is calculated by computing the Hamming distance between that network’s output vector and all the output vectors of all other networks ( and normalizing the result to get a behavioral diversity measure between 0 and 1 ) . Following [17] , the connection cost of a network is computed after finding an optimal node placement for internal ( hidden ) nodes ( input and output nodes have fixed locations ) that minimizes the overall network connection cost . These locations can be computed exactly [78] . Optimizing the location of internal nodes is biologically motivated; there is evidence that the location of neurons in animal nervous systems are optimized to minimize the summed length of connections between them [39 , 40 , 78] . Network visualizations show these optimal neural placements . The overall network connection cost cc is then computed as the summed squared length of all connections: c c = ∑ i j ( A i j d i s t ( i , j ) 2 ) ( 1 ) Where Aij is 1 if node i and node j are connected and 0 otherwise , and dist ( i , j ) is the euclidean distance between node i and j after moving them to their optimal positions . Network nodes are located in a two-dimensional Cartesian space ( x , y ) . The locations of input and output nodes are fixed , and the locations of hidden nodes vary according to their optimal location as just described . For all problems , the inputs have x coordinates of {−3 . 5 , −2 . 5 , … , 2 . 5 , 3 . 5} , a y coordinate of ( 0 ) , and the output is located at ( 0 , 4 ) , except for the problem in Fig 8 , right , which has an output located at ( 0 , 5 ) because of the extra layer of hidden neurons . The default network model is multi-layered and feed-forward , meaning a node at layer n receives incoming connections only from nodes at layer n − 1 and has outgoing connections only to nodes at layer n + 1 . This network model is common for investigating questions in systems biology [52 , 53 , 80] , including studies into the evolution of modularity [17 , 51] . In the experiments that allow long-range connections that can skip layers , a node at layer n can receive incoming connections from nodes at any layer <n , and it can have outgoing connections to nodes at any layer >n . While the layered and feed-forward nature of networks may contribute to elevated hierarchy , this network architecture is the same across all treatments and we are interested in the differences between levels of hierarchy that occur with and without a connection cost . For the main problem , AND-XOR-AND , networks are of the form 8 \ 4 \ 4 \ 2 \ 1 , which means that there are 8 input nodes , 3 hidden layers each having 4 , 4 and 2 nodes respectively , and 1 output node . The integers {-2 , -1 , 1 , 2} are the possible values for connection weights , whereas the possible values for the biases are {−2 , −1 , 0 , 1 , 2} . Information flows through the network in discrete time steps one layer at a time . The output yj of node j is the result of the function: y j = tanh λ ∑ i ∈ I j ω i j y i + b j where Ij is the set of nodes connected to node j , wij is the connection strength between node i and node j , yi is the output value of node i , and bj is a bias . The bias determines at which input value the output changes from negative to positive . The function , tanh ( x ) , is an activation function that guarantees that the output of a node is in the range of [−1 , 1] . The slope of the transition between the two extreme output values is determined by λ , which is here set to 20 ( S21B and S21C Fig ) . To be consistent with [17] , in all treatments , the initial number of connections in networks is randomly chosen between 20 and 100 . If that chosen number of initial connections is more than the maximum number of connections that a network can have ( which is 58 ) , the network is fully connected . When random , valid networks are generated , the initial number of connections ranges from the minimum number needed , which is 11 , to the maximum number possible , which is 58 . Each connection is assigned a random weight selected from the possible values for connection weights and is placed between randomly chosen , unconnected neurons residing on two consecutive layers . Our results are qualitatively unchanged when the initial number of connections is smaller than the default , i . e . when evolution starts with sparse networks ( S20 Fig ) . To create offspring , a parent network is copied and then randomly mutated . Each network has a 20% chance of having a single connection added . In the default experiments , candidates for new connections are pairs of unconnected nodes that reside on two consecutive layers . For experiments that allow connections to skip layers , two different layers are randomly selected and then a node from within each layer is selected randomly and those nodes are joined by a new connection . Each network also has a 20% chance of a randomly chosen connection being removed . Each node in the network has 0 . 067% chance of its bias being incremented or decremented with both options equally probable; five values are available {−2 , −1 , 0 , 1 , 2} and mutations that produce values higher or lower than these values are ignored . Each connection in the network has 2 . 0 n chance of its weight being incremented or decremented , where n is the total number of connections in the network . Because weights must be in the set of possible values {−2 , −1 , 1 , 2} , mutations that produce values higher or lower than these four values are ignored . Network modularity is calculated by finding a partition of the network into modules that maximizes the commonly used Q modularity metric for directed networks [81] , with the resulting Q score being the best estimate of the true modularity of the network . Because this metric has been extensively described before [81] , here we only describe it briefly . The Q metric defines network modularity , for a particular division of the network , as the number of within-module edges minus the expected number of these edges in an equivalent network , where edges are placed randomly [82] . The formula for this metric is: Q = 1 m ∑ i j [ A i j - k i i n k j o u t m ] σ c i , c j Where k i i n and k j o u t are the in- and out-degree of node i and j , respectively , m is the total number of edges in the network , Aij is an entry in the connectivity matrix , which is 1 if there is an edge from node i to j , and 0 otherwise , and σci , cj is a function whose value is 1 if nodes i and j belong to the same module , and 0 otherwise . Because iterating over all possible network partitions is prohibitively expensive , even for the small networks presented here , we adopt the widely used , computationally efficient spectral method for approximating the true Q score , which is described in [81] . Our default hierarchy measure comes from [11] . It ranks nodes based on their influence . A node’s influence on a network equals the portion of a network that is reachable from that node ( respecting the fact that edges are directed ) . Based on this metric , the larger the proportion of a network a node can reach , via its outgoing edges , the more influential it is . For example , a root node has more influence because a path can be traced from it to every node in the network , whereas leaf nodes have no influence . The metric calculates network hierarchy by computing the heterogeneity of the influence values of all nodes in the network . Intuitively , node-influence heterogeneity is high in hierarchical networks ( where some nodes have a great deal of influence and others none ) , and low in non-hierarchical networks ( e . g . in a fully connected network the influence of nodes is perfectly homogeneous ) . Because of the non-linear function that maps a node’s inputs to its output , even a small change in the input to a node can change whether it fires . For that reason , it is difficult to determine the influence one node has on another based on the strength of the connection between them . We thus calculate hierarchy scores by looking only at the presence of connections between nodes , ignoring the strength of those connections . The score for a weighted directed network is calculated by: ∑ i ∈ V [ C R m a x - C R ( i ) ] N - 1 C R m a x is the highest influence value and V represents a set of all nodes in the network . N is the number of nodes in the network . Each node in the network is represented by i . CR ( i ) , the influence value of node i , is given by: C R ( i ) = 1 N - 1 ∑ j : 0 < d o u t ( i , j ) < ∞ [ 1 d o u t ( i , j ) ] Here , dout ( i , j ) is the length of the path that goes from node i to node j , meaning it is the number of outgoing connections along the path . This section gives a brief description on the second , alternate hierarchy metric that we employed to verify our main results: a complete description of the method can be found in Czégel and Palla [83] . The metric ranks network nodes based on their influence on other nodes in the network . Such influence can be measured by , for each step of the algorithm , starting random walkers at each node in the network and allowing them to randomly traverse a single network edge in reverse order based on the relative importance of each edge . The distribution of which nodes are visited is normalized after each step , and aggregated across many steps . In the limit , a unique , stable , distribution is reached that reflects the influence of each node in the network on the other nodes . Hierarchy is then quantified as the relative standard deviation of this distribution; more heterogenous distributions reveal more hierarchical structure , because they have a few nodes with high influence on other nodes whereas most have little influence . Conversely , a highly homogenous distribution indicates a non-hierarchical structure in which the influence of each node on the others is similar . As a proxy for quantifying functional hierarchy , we measure the percent of logic sub-problems of an overall problem that are solved by part of a network . Note that an overall logic problem can be solved without solving any specific sub-problem ( e . g . in the extreme , if the entire problem is computed in one step ) . We determine whether a logic sub-problem is solved as follows: for all possible inputs to a network , a neuron solves a logic gate ( a sub-problem ) if , for its outputs , there exists any threshold value that correctly separates all the True answers from the False answers for the logic gate in question . We also consider a sub-problem as solved if it is solved by a group of neurons on the same layer . To check for this case , we consider all possible groupings of neurons in a layer ( groups of all sizes are checked ) . We sum the outputs of the neurons in a group and see if there is a threshold that correctly separates True and False for the logic sub-problem on all possible network inputs . To prevent counting solutions multiple times , each sub-problem is considered only once: i . e . the algorithm stops searching when a sub-problem is found to be solved . | Hierarchy is a ubiquitous organizing principle in biology , and a key reason evolution produces complex , evolvable organisms , yet its origins are poorly understood . Here we demonstrate for the first time that hierarchy evolves as a result of the costs of network connections . We confirm a previous finding that connection costs drive the evolution of modularity , and show that they also cause the evolution of hierarchy . We further confirm that hierarchy promotes evolvability in addition to evolvability caused by modularity . Because many biological and human-made phenomena can be represented as networks , and because hierarchy is a critical network property , this finding is immediately relevant to a wide array of fields , from biology , sociology , and medical research to harnessing evolution for engineering . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"genetic",
"networks",
"neural",
"networks",
"engineering",
"and",
"technology",
"neuroscience",
"logic",
"circuits",
"network",
"analysis",
"genome",
"analysis",
"computer",
"and",
"information",
"sciences",
"animal",
"cells",
"evolutionary",
"genetics",
"cellular",
"neuroscience",
"cell",
"biology",
"gene",
"regulatory",
"networks",
"neurons",
"genetics",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"genomics",
"evolutionary",
"biology",
"computational",
"biology",
"electronics",
"engineering"
] | 2016 | The Evolutionary Origins of Hierarchy |
Plasmodium parasites invade and multiply inside red blood cells ( RBC ) . Through a cycle of maturation , asexual replication , rupture and release of multiple infective merozoites , parasitised RBC ( pRBC ) can reach very high numbers in vivo , a process that correlates with disease severity in humans and experimental animals . Thus , controlling pRBC numbers can prevent or ameliorate malaria . In endemic regions , circulating parasite-specific antibodies associate with immunity to high parasitemia . Although in vitro assays reveal that protective antibodies could control pRBC via multiple mechanisms , in vivo assessment of antibody function remains challenging . Here , we employed two mouse models of antibody-mediated immunity to malaria , P . yoelii 17XNL and P . chabaudi chabaudi AS infection , to study infection-induced , parasite-specific antibody function in vivo . By tracking a single generation of pRBC , we tested the hypothesis that parasite-specific antibodies accelerate pRBC clearance . Though strongly protective against homologous re-challenge , parasite-specific IgG did not alter the rate of pRBC clearance , even in the presence of ongoing , systemic inflammation . Instead , antibodies prevented parasites progressing from one generation of RBC to the next . In vivo depletion studies using clodronate liposomes or cobra venom factor , suggested that optimal antibody function required splenic macrophages and dendritic cells , but not complement C3/C5-mediated killing . Finally , parasite-specific IgG bound poorly to the surface of pRBC , yet strongly to structures likely exposed by the rupture of mature schizonts . Thus , in our models of humoral immunity to malaria , infection-induced antibodies did not accelerate pRBC clearance , and instead co-operated with splenic phagocytes to block subsequent generations of pRBC .
Clinical symptoms of malaria occur during the erythrocytic phase of infection , when Plasmodium parasites mature and replicate asexually in red blood cells ( RBC ) [1] . A key feature of the asexual life-cycle in RBC is the capacity for rapid population growth , because each parasite can produce many daughter parasites ( up to 32 daughter merozoites ) . The fold-increase in parasitised RBC ( pRBC ) from one cycle to the next is expressed by Parasite Multiplication Rate , PMR , a useful measure of parasite growth [2 , 3 , 4] . Theoretically , PMR can be influenced by host and parasite factors , such as how many merozoites are produced per replicating parasite , or how effectively the host clears parasites from the bloodstream . Since parasite biomass correlates strongly with disease severity in P . falciparum-infected humans [5] and in experimentally-infected mice [6] , there can be serious consequences for an infected host over just a few cycles of infection . For example , if PMR = 16 , four cycles of infection could theoretically result in a 65 , 536-fold ( 164 ) increase in pRBC density . Therefore , it is important to minimise parasite replication to reduce the risk and severity of malaria . Passive transfer of immune serum , firstly in rhesus monkeys [7] , and later in children with malaria [8] , was shown to reduce parasitemia and symptoms in the majority of recipients . This antibody-mediated control showed a slow decline in parasitemia with time , consistent with antibody-mediated “clearance” of pRBC . However , the rate of decay of pRBC appeared slow compared to that seen in drug treatment . In addition , because of the periodic life-cycle of the parasite , decays in parasite numbers might also occur by blocking parasite replication . Nevertheless , these passive transfer studies demonstrated that parasite-specific antibodies can control pRBC numbers effectively in humans and animal models , although precise mechanisms were not elucidated . Substantial effort has since been made to determine both the antigenic targets of protective antibodies [9 , 10] , and secondly , the mechanisms by which they control pRBC numbers in vivo ( reviewed in [9 , 11 , 12 , 13] ) . Broadly speaking , these antibodies exhibit two major specificities , either to Plasmodium proteins exported to the surface of pRBC , or to those expressed on the surface of merozoites [9] . Within either class , a number of possible in vivo mechanisms have been proposed and revealed in elegant in vitro assays . These include Growth Inhibition Assays ( GIAs ) [14 , 15] , which assess if antibodies can reduce parasite growth in RBC over a few replication cycles; opsonic phagocytosis assays [16 , 17 , 18] and antibody-dependent cellular inhibition ( ADCI ) assays [19] , which assess whether antibodies facilitate uptake and/or inhibition by phagocytes; and complement-fixing assays and killing assays [20 , 21] , which determine how well antibodies facilitate C1q deposition and complement-dependent direct killing . A clear correlation between functional efficacy in GIAs and immunity to malaria is lacking [22 , 23 , 24 , 25] . However , in recent years , the capacity to mediate opsonic phagocytosis or C1q-fixation in vitro , mostly of merozoites , has been shown to correlate with protection against P . falciparum malaria symptoms and high density parasitemia [18 , 20 , 21] . These reports suggested that antibody-mediated immunity to blood-stage infection can be mediated by binding to merozoites , but with necessary engagement of host factors including phagocytes and certain aspects of the complement system . More recently , it was also suggested that antibodies against the pRBC surface protein , PfEMP1 , which drove opsonic phagocytosis in vitro , afforded partial protection against malaria , and that these antibodies were acquired slightly earlier in life than merozoite-targeting antibodies [26] . Thus , the epidemiological data in humans supports the concept that multiple co-existing mechanisms acting against merozoites and pRBC can contribute to antibody-mediated pRBC control during blood-stage infection . However , few if any of these mechanisms have been convincingly observed in vivo . Here , we sought to explore in vivo mechanisms of action for protective antibodies , specifically those acquired via primary blood-stage infection . We employed our established RBC adoptive transfer system , which permits tracking of a single cohort of fluorescently-labelled pRBC , as well as direct estimation of PMR and its inhibition by antibodies . We recently reported this technique for P . berghei ANKA parasites , to study the effect of anti-malarial drugs and the innate immune system [27 , 28 , 29] . Here , we have employed our approach in two common , non-lethal blood-stage malaria mouse models , P . yoelii 17XNL and P . chabaudi chabaudi AS infections in C57BL/6J mice , in which antibody-mediated immunity to homologous re-challenge with either parasite is generated after a single , self-resolving or chronic primary infection[30 , 31 , 32] . Interestingly , like human-infective Plasmodium species , P . yoelii encodes hundreds of pRBC surface proteins , called yirs , and several merozoite proteins , all of which are targets of the host immune system in vivo [33 , 34 , 35] . Thus , using these two established in vivo models of antibody-mediated immunity , we sought to determine how infection-induced antibodies control pRBC numbers in vivo .
Passive transfer of immunoglobulin from immune adults has previously been shown to reduce parasitemias in children with malaria [8] , via mechanisms that remain to be elucidated in vivo . Here , we first examined the utility of a mouse model , Py17XNL infection of C57BL/6J mice , for studying antibody control of blood-stage parasites in vivo . We and others previously reported that in Py17XNL primary infection , parasitemia is largely resolved by 30 dpi and that Py17XNL-specific IgG levels subsequently peak around 60–70 dpi [36 , 37] . Firstly , we determined whether sera from mice that had resolved a primary Py17XNL infection , and which harboured parasite-specific IgG ( S1A Fig ) , could restrict parasitemia when administered therapeutically to mice infected with homologous parasites ( S1B Fig ) . At day 3 p . i . when parasitemia was patent ( ~0 . 05% ) , infected mice were transfused with a single dose of immune serum , or non-immune control serum . As expected , the majority ( 5/6 ) of those receiving immune serum controlled parasitemia over the next 36 hours more effectively than non-immune serum recipients ( S1B Fig ) . Next , we hypothesized that antibodies elicited by Py17XNL infection could accelerate pRBC clearance by the host . To assess this , we used our established RBC adoptive transfer protocol [27 , 28 , 29] . Briefly , we harvested RBC from Py17XNL-infected WT passage mice ( typically infected for 4-days , harbouring parasitemias of 2–10% , and exhibiting ~70% ring-stages ) , fluorescently-labelled them with CellTrace Far-Red ( CTFR ) ( Fig 1A ) , and transferred this first generation , termed Gen0 ( including un-infected , CTFR-labelled-RBC ) into mice given Py17XNL-immune serum or non-immune control serum 24 hours previously ( Fig 1A ) —we note in this protocol that mice were challenged with high doses of pRBC ( ~107 pRBC/mouse ) . We first assessed “total” parasitemia ( ie . disregarding CTFR-staining of RBCs ) , since this is a more conventional assessment of pRBC in experimental animals or humans ( S2A Fig ) . This confirmed that immune serum controlled total parasitemia over the first 3 days of high-dose homologous challenge compared to controls ( S2A Fig ) . Thus , as expected , primary Py17XNL infection had elicited immunity that could be passively transferred under both therapeutic and high-dose challenge conditions in vivo . Next , we made use of CTFR-labelling on pRBC to specifically study the initial generation ( Gen0 ) of pRBC over the first 25 hours post-challenge ( Fig 1B and 1C & S3 Fig ) . We used Hoechst33342 and Syto84 nucleic acid dyes to identify parasite life-cycle stages of Gen0 pRBC over the first 25 hours . Similar to our previous reports using P . berghei ANKA [27 , 28 , 29] , Gen0 Py17XNL parasites matured into schizonts ( S3 Fig ) and were lost from circulation , mostly likely due to a combination of maturation-induced rupture and pRBC clearance . We recently reported that host factors can slow parasite maturation in vivo [27] . However , Gen0 parasite maturation occurred equivalently in immune and non-immune serum recipients over the 25-hour period ( S3 Fig ) . Surprisingly , however , the loss of Gen0 pRBC was also equivalent between immune and non-immune serum recipients ( Fig 1C ) , suggesting that pRBC clearance had been unaffected . Thus , our data reveal that immune serum , though protective against increases in parasitemia , did not influence the rate of pRBC clearance by the host . We next assessed whether antibodies impeded Gen0 parasites from successfully infecting subsequent generations of un-infected RBCs ( termed Gen1+ ) , which we detected through the presence of parasites in CTFRneg RBC ( Fig 1B ) . While we observed rapid appearance of Gen1+ pRBC in non-immune serum recipients , immune serum substantially blocked the emergence of Gen1+ pRBC ( Fig 1D ) . The PMR ( the average number of Gen1 pRBC generated at 25h from each transferred Gen0 pRBC at 1h ) was reduced to 3 . 7±0 . 1 by immune serum , compared to 11 . 5±0 . 8 in non-immune serum recipients , indicating that passively transferred immune serum prevented transit from one generation of RBC to the next . However , given that donor Gen0 pRBC had been generated in WT passage mice , it is possible that natural antibodies or rapidly-generated IgM produced in the passage mice might coat donor pRBC prior to harvest and transfer , and thus interfere with the function of antibodies in immune serum . To test this , we repeated the above experiment using Gen0 donor pRBC generated either in rag1-/- passage mice that completely lack antibody responses , or WT passage controls ( S4A Fig ) . Both rag1-/- and WT-generated Gen0 pRBC were cleared similarly in immune and non-immune serum recipients ( S4 Fig ) , suggesting prior antibody coating of our donor pRBC was not a confounding factor in our studies . Next , to confirm that infection-induced IgG was responsible for the passive transfer of protection , we repeated above experiments with purified IgG from immune or control serum ( S5A Fig ) , alongside non-purified serum ( S5B and S5C Fig ) . We found that purified IgG from immune serum did not alter the rate of Gen0 pRBC clearance ( S5B Fig ) , but once again prevented emergence of Gen1+ pRBC in a similar manner to non-purified immune serum ( S5C Fig ) . Together , our data confirmed that parasite-specific IgG controlled infection , not via pRBC clearance or impaired parasite maturation , but by blocking the progeny of Gen0 pRBC from establishing subsequent generations of pRBC . The experiments described above involved passive transfer into naïve mice . We hypothesized that the lack of accelerated pRBC removal might be due to sub-optimal levels of antibodies transferred ( in 200μl volumes ) . However , neither doubling the volume of immune serum ( 400μl ) , nor passive serum transfer closer to the time of challenge had any effect on pRBC removal , despite again restricting the emergence of Gen1+ pRBC ( S6 Fig ) . It was also possible that additional immune mechanisms present in immune mice might be required to induce pRBC clearance by antibodies . To test this , we transferred Gen0 pRBC directly into immune mice ( that had resolved a primary infection ) . As before , we observed no changes in Gen0 pRBC loss in immune mice compared to controls ( Fig 2A ) , while total parasitemia ( S2B Fig ) , and emergence of Gen1 pRBC was blocked , with PMR substantially reduced from 17 . 3±7 . 7 in non-immune mice to 0 . 78±0 . 29 in immune mice . These data suggested that the combined cellular and humoral components of immunity present in Py17XNL immune mice did not affect the rate of clearance of pRBC . Finally , our experiments to date had examined antibody function in the absence of ongoing inflammation , where factors such as systemic pro-inflammatory cytokines might be expected to augment antibody function . To examine infection-induced antibodies in this context , we transferred immune and non-immune serum into mice stimulated with TLR4 or TLR9 agonists , since Plasmodium infection can stimulate the innate immune system and pro-inflammatory cytokine production via these receptors ( Fig 3 ) [38 , 39 , 40] . Despite clear upregulation of systemic cytokines IFNγ , TNF and IL-6 ( Fig 3A ) , neither TLR agonist induced antibodies to accelerate the clearance of Gen0 pRBC ( Fig 3B ) . Together our data indicated that the failure of infection-induced antibodies to accelerate pRBC clearance was unlikely due to the lack of accompanying cellular immunological memory or ongoing systemic inflammation . We next hypothesised that additional host factors might be required for infection-induced antibodies to function in vivo . In vitro experiments have implicated phagocytes in antibody-mediated immunity to malaria in humans , via opsonic phagocytosis and antibody-dependent cellular inhibition ( ADCI ) . To assess in vivo roles for phagocytes , we administered clodronate liposomes two days prior to transfer of immune or non-immune control serum . As expected , clodronate substantially depleted F4/80+ macrophages and CD11chi conventional dendritic cells in the spleen , with only minor effects on other cells such as monocytes and neutrophils ( S7A and S7B Fig ) . We then transferred and tracked CTFR+ ( Gen0 ) pRBC as above ( Fig 4A ) . Over the first 25 hours , clodronate-treatment modestly reduced the rate of clearance of Gen0 pRBC in mice receiving either immune or non-immune serum ( Fig 4B ) . More importantly , however , clodronate treatment reduced antibody effectiveness , since immune serum reduced PMR by 61 . 7% ( range: 50–70 ) in phagocyte-replete mice , yet by only 31 . 7% ( range: 25–40 ) in mice depleted of splenic macrophages/dendritic cells ( Fig 4C: P<0 . 0096 , two-way ANOVA , significant interaction between serum administration and clodronate treatment , 5 mice/group , three independent experiments ) . Thus Py17XNL-immune serum was ~50% less effective in the absence of phagocytes over the first 25 hours . Nevertheless , at days 2–3 post-transfer , immune serum limited the emergence of later generations ( Gen1+ ) of pRBC in both phagocyte-depleted and intact mice ( Fig 4C ) , suggesting that phagocyte-independent antibody-mediated mechanisms might exist in our model systems . In summary , parasite-specific antibodies co-operated with splenic macrophages and dendritic cells to optimally restrict parasites transitioning from Gen0 to Gen1 RBC . Finally , we also considered phagocyte-independent antibody mechanisms . Previous studies reported two mechanisms , complement-dependent merozoite invasion blockade requiring C1q-deposition , and complement-mediated direct lysis of merozoites via formation of C3/C5-dependent membrane attack complexes ( MAC ) . To test for the importance of MAC formation in antibody function , we depleted mice of complement proteins C3/C5 ( S8A and S8B Fig ) , using non-toxic , cobra venom factor ( CVF ) prior to transfer of immune or non-immune serum . Importantly , CVF-treatment does not target C1q . Having administered Gen0 pRBC we noted that as before , loss of Gen0 pRBC was equivalent amongst all groups ( S8C Fig ) . We then compared PMR reduction elicited by Py17XNL-immune serum ( compared to non-immune serum ) in mice either replete or CVF-treated . We found no evidence that immune serum failed to control PMR in CVF-treated mice compared to saline controls ( S8D Fig ) , with no significant effect observed in two experiments ( P>0 . 12 , two-way ANOVA interaction between immune serum administration and CVF treatment , 5 mice/group , 2 independent experiments ) . Therefore our data revealed no role for complement C3/C5-mediated killing in antibody function . Taken together , our data suggested that infection-induced antibodies co-operated with splenic macrophages and dendritic cells , but did not require complement-mediated direct killing for their function in vivo . Given that infection-induced IgG blocked parasite transit between RBC , but did not accelerate pRBC clearance , we hypothesized that antibodies in immune serum primarily target merozoites/internal structures that are exposed upon schizont rupture , rather than antigens on the surface of the infected RBC . To test this , we injected immune serum or non-immune control serum into mice followed by transfer of Gen0 pRBC ( Fig 5A ) . After 1 or 4 hours in ciruclation to afford time for in vivo binding , we then removed and examined Gen0 pRBC for evidence of cell-surface binding of mouse IgG ( Fig 5B ) . As a positive control for mouse IgG-deposition on RBC , control samples were incubated in vitro with a mouse monoclonal IgG ( clone 34-3C ) specific for murine RBC ( Fig 5B ) . After 1 or 4 hours of in vivo exposure to immune serum , only 2–7% of pRBC bound mouse IgG , the extent of which was weak compared to 34-3C-treated controls ( Fig 5B ) . Thus , consistent with our hypothesis , Py17XNL-specific IgG bound only weakly to the surface of pRBC in vivo . Next , to test whether infection-induced parasite-specific IgG could potentially bind internal parasite structures exposed during schizont rupture , we fixed and permeabilised pRBC ( from either WT or rag1-/- passage mice ) in vitro prior to exposure to immune serum ( Fig 6 ) . Interestingly , we note that while fixation and permeabilisation lysed all the un-infected RBC in our preparation and perhaps early-ring stages ( Fig 6A compared to 6B ) , later-stage pRBC remained intact , permitting immune-staining and flow cytometric assessment ( Fig 6B ) . In stark contrast to pRBC-surface assessment ( Fig 6A ) , ~90% of intact , permeabilised pRBC strongly bound IgG from immune serum , compared to <5% for control serum ( Fig 6B ) . Furthermore , pRBC with the highest DNA content , indicative of progression through schizogony , exhibited the strongest IgG staining ( Fig 6C ) . Finally , microscopy assessment revealed immune IgG binding to pRBC , particularly to multiple internal structures within schizonts , including but not limited to nascent , DAPI+ merozoites ( Fig 7 ) . Taken together , our data indicate that infection-acquired IgG antibodies bind poorly to the surface of pRBC , yet strongly to internal structures of late-stage pRBC , which may become exposed to antibodies in vivo upon schizont rupture . We next sought to determine whether similar antibody mechanisms act in a second model , P . chabaudi chabaudi AS infection . We harvested immune serum from PcAS-infected wild-type C57BL/6J mice at 40 dpi , the peak of parasite-specific IgG in this model [36] , and confirmed the presence of PcAS-specific IgG ( Fig 8A ) . Next , we employed the same approaches as above , transferring PcAS-infected CTFR+ pRBC into naïve mice given PcAS-immune serum or control serum ( Fig 8B & 8C ) . The loss of Gen0 pRBC was again not influenced by immune serum ( Fig 8B ) , although the emergence of Gen1+ pRBC was strongly blocked ( Fig 8C ) , with PMR reductions of ~50% for immune serum compared to controls . Consistent with no effect on pRBC clearance , yet strong effects on Gen0 to Gen1 transition , we noted that PcAS-specific IgG did not bind to the surface of pRBC in vitro ( Fig 8D ) , but rather to intracellular structures inside late-stage , permeabilised pRBC ( Fig 8E & 8F ) . Thus , in two independent models , infection-induced antibodies did not bind to molecules on the surface of pRBC , and protected by blocking parasite transit between RBCs .
Several mechanisms have been proposed to explain how Plasmodium-specific antibodies acquired through natural exposure may prevent or ameliorate malaria . Broadly speaking , these include three classes of mechanism: where antibodies act mostly alone , without other host processes required; where they serve as triggers for phagocyte-dependent processes ( opsonic phagocytosis and antibody-dependent cellular inhibition ( ADCI ) ) ; or where they activate complement-mediated direct killing . All three classes of mechanism could theoretically act on pRBC or free merozoites . However , it has been challenging to explore the importance of these in vivo . Here , we utilised two of the most common in vivo models of humoral immunity to malaria , P . yoelii 17XNL and PcAS infection of C57BL/6J mice , to dissect how infection-induced antibodies control parasite numbers during secondary challenge . This was combined with an RBC adoptive transfer technique to study parasite clearance and replication in vivo [27 , 28 , 29] . We provided evidence that effective control of pRBC numbers was mediated not by targeting pRBC for clearance by the host , either in naïve mice , or when cellular components of adaptive immune memory or ongoing systemic inflammation were added . Moreover , we provide evidence that splenic macrophages and dendritic cells are necessary for optimal in vivo antibody function , while complement C3/C5-mediated killing played no discernible role . Moreover , while infection-induced IgG antibodies bound very poorly to the surface of intact pRBC in either model , they bound well to internal structures of schizonts . Our flow cytometric and microscopy data suggest merozoites and other internal structures could be targets for antibodies once schizont rupture has occurred . Thus , we propose here a model in which infection-induced antibodies target merozoites and internal parasite structures exposed upon schizont rupture , and co-operate with splenic phagocytes to impair transit of parasites from one RBC to another . Although our study indicated a role for splenic phagocytes in preventing parasites transitioning from one generation of RBC to the next , which of the two main phagocyte-dependent mechanisms , opsonic phagocytosis or ADCI , were involved remains to be determined . Given that ADCI is mediated in vitro by a monocyte-derived cytokine , Tumor Necrosis Factor ( TNF ) [19] , future studies should explore an in vivo requirement for TNF in our system . Another important question arising from the reduction in antibody-efficacy after phagocyte depletion , is whether the partial effect was due to incomplete phagocyte depletion by clodronate liposomes , or indicates the existence of multiple mechanisms of antibody action . Although a single-dose of clodronate liposomes does not completely deplete phagocytes for more than a few days , it nonetheless was a very efficient way to remove splenic macrophages in particular . For that reason , we believe that other mechanisms likely contribute to parasite control in our models , such as blockade of merozoite invasion or pro-inflammatory responses to antibody-coated fragments of ruptured schizonts [20] . A further possibility is that non-phagocytic cells expressing Fc receptors may also mediate antibody-dependent control in our in vivo models . For example , a recent study demonstrated human NK cells mediated antibody-dependent cellular cytotoxicity against Plasmodium falciparum-infected RBC in vitro [41] . Future experiments and technological advances may be required to assess these mechanisms in vivo since it is not trivial to reliably visualise pRBC , merozoites , or schizont fragments interacting with phagocytes or other immune cells in vivo . A recent study revealed in human samples that the capacity of merozoite-specific antibodies to facilitate deposition of the polymeric complement C1q complex strongly correlated with protection from high density parasitemia [20] . Moreover , in vitro assays indicated that although C1q deposition could increase direct complement-mediated lysis of merozoites , C1q deposition alone , ie . with no other components of the complement system present , was sufficient to enable antibodies to block merozoite invasion of RBC . Thus a conclusion from this paper was that merozoite-specific antibodies from children protected against high density parasitemia probably acted by blocking invasion , not by direct complement-mediated lysis of merozoites . Our study in mice is consistent with infection-acquired IgG blocking merozoite transit to the next generation of RBC , with no major role for complement-mediated direct lysis . However , in our study , we did not directly assess a role for antibody-mediated C1q deposition . Thus , it may be important in the future to examine whether parasite-specific antibodies are impaired in their capacity to block Gen0 to Gen1 transit in C1q-/- mice . Another important question is whether the complement system is adequately functional in the C57BL/6J genetic background employed in this study , for any similarities to be drawn to human immunity to malaria . A previous assessment of complement in numerous inbred mouse strains focussed on the capacity to directly lyse cellular targets [42] . This report indicated that strains with lower complement lysis capabilities tended to have reduced amounts of C3 , C5 , C6 and C7 in their serum compared to strains with better complement lysis capacity , but similar levels of C1 proteins [42] . Given these findings , it is perhaps unsurprising that direct complement mediated lysis was not an important mechanism in our system for controlling infection . However , given that levels of C1 proteins are likely equivalent amongst most mouse strains [42] , it remains to be determined whether C1q deposition contributes to antibody function in the Py17XNL and PcAS systems . We hypothesise this to be the case since a previous report showed that genetic C1q deficiency in mice impaired immunity to re-challenge with blood-stage Plasmodium parasites [43] . Many , but not all , of the suggested mechanisms by which antibodies could serve to control pRBC numbers in vivo can be explored during Py17XNL infection in mice . For example , it has been suggested that antibodies could prevent cytoadherence , and thus sequestration , of pRBC to endothelial surfaces in various tissues [11 , 44] . Similarly , antibodies could prevent adherence of RBC to each other , also known as rosetting [11 , 44] . Neither of these two phenomena are well-established in the Py17XNL model . Therefore , our data cannot be taken as complete assessment of all the antibody mechanisms that prevent malaria in humans . Instead , our study emphasises that antibody binding to the pRBC surface is not necessary for control , while targeting free merozoites and/or ruptured schizonts appears to predominate in two of the most common mouse models of antibody immunity . It has previously been shown that Py17XNL parasite export variant surface proteins , yirs , to the surface of RBC during in vivo infection of mice . Moreover , yirs are clearly detected by the mouse immune system , since their variation is dependent upon the presence of adaptive immunity . In our study , we found evidence that intact schizont stage pRBC did bind parasite-specific IgG . It is possible that yir proteins could be the target for such antibodies , in a similar way that PfEMP1 appears to be a major target for pRBC binding antibodies in human samples [26] . However , an important question from our study is why the IgG response to infection is dominated by antibodies specific to the contents of permeabilised schizonts , most likely merozoites and proteins released into circulation during rupture , and not to pRBC surface proteins . The answer to this question remains unclear at present but could be related to the ability of antigen-presenting cells to access and process pRBC surface proteins compared to merozoite proteins and the cellular debris released by pRBC rupture . We previously described in our models , the kinetics with which parasite-specific IgM and IgG isotypes accumulate in mice infected with P . chabaudi and P . yoelii 17XNL [36 , 37] . However , this study was designed primarily to assess peak levels of parasite-specific IgG ( at 40–60 days post-infection ) . Therefore , our experiments do not shed light on how the sequential generation of IgM and IgG classes might mediate immunity , nor how different immune cell-types might interface with these Ig classes . Future experiments could explore the functional roles of IgM and IgG sub-classes in vivo . In this study , the serum transfer achieved only a fraction of the antibody concentration present in the serum of an immune mouse , yet elicited clear protection even against high inocula . Mice were therefore administered sufficient antibodies to protect against large infectious doses . However , this model appears different to natural P . falciparum infection scenarios , where protective antibody levels might be relatively low , and where individuals are infected with much lower numbers of blood-stage parasites . Nevertheless , we speculate that antibody mechanisms outlined in our mouse studies may be relevant for understanding immunity in humans . Our study also raises questions regarding how antibody-mediated immunity might be boosted in a rational manner . For example , could increasing the plasma concentration of antibodies that recognise pRBC further improve protection ? This may indeed be the case . However , given that antigenic variation is a major feature of pRBC surface proteins , it will likely be challenging to induce pRBC-targeting antibodies that elicit long-lasting immunity to high-density parasitemia in humans . Nevertheless , in summary , our data using in vivo models suggest that pRBC clearance is not a crucial mechanism for naturally-acquired antibodies to control parasitemia . Instead , we propose that binding of affinity-matured IgG to the surface of merozoites and/or to fragments of ruptured schizonts protects against blood-stage Plasmodium infection via multiple mechanisms including those mediated by splenic macrophages and dendritic cells . Our pre-clinical in vivo data provide further insight into the types of antibody functions that might protect naturally-exposed and immunised humans against malaria .
Female C57BL/6J mice aged 6–12 weeks were purchased from the Australian Resource Centre ( Canning Vale , Perth , WA , Australia ) and maintained under conventional conditions . This study was carried out in strict accordance with guidelines from The National Health and Medical Research Council of Australia , as detailed in the document Australian Code of Practice for the Care and Use of Animals for Scientific Purposes , 7th edition , 2004 . All animal procedures and protocols were approved ( A1503-601M ) and monitored by the QIMR Berghofer Medical Research Institute Animal Ethics Committee . P . yoelii 17XNL ( Py17XNL ) or P . chabaudi chabaudi AS ( PcAS ) parasites were used after defrosting frozen , infected stabilate blood and performing a single in vivo passage in wild-type ( WT ) or rag1-/- C57BL/6J mice ( Animal Resource Centre , Perth , Australia ) until a parasitemia of 1–5% was reached . Immune mice ( and thus , immune sera ) were generated by infecting WT mice with 104 Py17XNL-containing pRBC or 105 PcAS-containing pRBC , via intravenous ( i . v . ) injection using a lateral tail vein ( 200μl using 26G needles ) , and confirming firstly that infection had been initiated around day 7–14 days post-infection ( dpi ) , and secondly that peripheral blood parasitemia had dropped below 0 . 1% by flow cytometry ( described below ) by 40–50 dpi for PcAS , and 60–80 dpi for Py17XNL . Age-matched naïve control mice were housed in the animal facility to generate non-immune mice and control , non-immune serum . Unless stated otherwise , mice were administered 200μl of immune or control serum , or 500μg of purified IgG , via intravenous injection using a lateral tail vein . Adoptive transfer of a single cohort of fluorescently-labelled pRBC was carried out as previously described [27 , 28 , 29] . pRBC , including un-infected RBC , were collected from Py17XNL or PcAS-infected passage mice by cardiac puncture . Heparinised donor blood was washed twice in Ca2+/Mg2+-free phosphate buffered saline ( PBS-A ) , and stained in CellTrace Far Red ( CTFR ) according to manufacturer’s instructions . Briefly , 50μg CTFR was dissolved for ten minutes in 25μl tissue culture grade dimethyl sulphoxide . This was then added to 5ml of resuspended RBC in PBS-A . RBC were stained in the dark , at room temperature with constant rolling for 15 minutes , and then washed twice in 10x volumes of PBS-A . Successful CTFR-labelling of RBC was confirmed by flow cytometry using an LSRII Fortessa analyser ( BD Biosciences ) and FlowJo software ( Treestar , CA , USA ) . CTFR-labelled RBC were resuspended in 2ml volumes per donor mouse , and injected in 200μl volumes via i . v . injection using 26G needles . Our established flow cytometric method for P . berghei ANKA parasites , was employed here to track the first generation of injected Py17XNL or PcAS-containing pRBC , termed Gen0 pRBC , and to distinguish these from subsequent generations of pRBC ( termed Gen1+ ) , which emerged as un-labelled , endogenous RBC became infected by the merozoite progeny of Gen0 pRBC . Finally , our staining approach also permitted assessment of parasite life-stages ( ring , trophozoite and schizont ) , as previously reported [27 , 28 , 29] . Briefly , a single drop of blood from a tail bleed was diluted and mixed in 200 μl of RPMI medium containing 5 U/ml heparin sulphate . Diluted blood was simultaneously stained for 30 min in the dark at room temperature with cell-permeant RNA/DNA stain , Syto84 ( 5 μM; Life Technologies ) and with DNA stain , Hoechst 33342 ( 10 μg/ml; Sigma ) . Staining was quenched with 10 volumes of ice-cold RPMI medium , and samples immediately analysed by flow cytometry using an LSRII Fortessa analyser ( BD Biosciences ) and FlowJo software ( Treestar , CA , USA ) . FSC-Area ( FSC-A ) and FSC-Height ( FSC-H ) was used to exclude doublets ( the result of two cells being detected simultaneously by the flow-cytometer ) . FSC-A and side scatter ( SSC-A ) was used to distinguish RBCs from debris and white blood cells on the basis of size and granularity . Infected RBCs were detected as co-expressing Hoechst 33342 and Syto84 , with expression of either one alone being insufficient to identify the presence of parasites . Adoptively-transferred Gen0 pRBCs were readily distinguished from endogenous RBC , and importantly from Gen1+ pRBC , by CTFR-labelling . Purification of IgG from immune serum or non-immune serum was performed using Protein G Sepharose4 Fast Flow ( GE Healthcare ) column , according to the manufacturers’ protocol . Briefly , the column was equilibrated with Running Buffer ( 0 . 02M sodium phosphate , 0 . 01% sodium azide , pH 7 ) prior to running serum ( diluted 1:3 with Running Buffer ) through at ~1 . 5 ml/minute . Flow-through ( FT ) was collected and prior to elution , the column was washed with Running Buffer to remove any unbound protein . IgG was eluted from the column with Elution Buffer ( 0 . 1M glycine , 0 . 01% sodium azide , pH 2 . 7 ) and the pH of the eluate was neutralised by addition of 1M Tris-HCl pH 9 . Buffer exchange of the eluate to 0 . 9% saline was performed using Amicon Ultra-15 10K Centrifugal Filter devices ( Merck Millipore ) and sterilised using Acrodisc 0 . 2μm Mustang E Membrane Units ( PALL Life Sciences ) . The concentration of purified IgG was determined using a NanoDrop ND-1000 and the IgG was then stored at -80°C until use . For SDS-PAGE analysis , samples were resuspended in 4x NativePAGE Sample Buffer ( ThermoFisher Scientific ) and β-mercaptoethanol ( 10% final concentration ) was added prior to denaturing at 95°C for 5 minutes . Samples were cooled on ice for 5 minutes and then separated by 10% SDS-PAGE . Protein sizes were estimated by loading of a BenchMark Pre-stained Protein Ladder ( Life Technologies ) and an anti-mouse CD3 rat IgG ( Biolegend ) was used as a control . The SDS-PAGE gel was stained with Coomassie Brilliant Blue R250 and imaged on an Odyssey CLx imaging system ( LI-COR Biosciences ) . Crude antigen extract from Py17XNL-infected pRBC or PcAS-infected pRBC was prepared using an adapted version of a previously described protocol [45] . Briefly , mice were infected as described above . When parasitemias reached 20–30% , blood was collected by cardiac puncture into heparinized tubes . RBC were washed once in RPMI at 1200rpm for seven minutes at room temperature , and then lysed using ultrapure water followed by four washes in ice-cold PBS at 16 , 000xg for 25 minutes at 4°C , as well as three cycles of freezing ( two hours at -80°C ) and thawing ( 30 minutes at room temperature ) . Extracts were also processed from RBCs of uninfected C57BL/6J mice , for use as negative controls in ELISA . The concentration of proteins in the purified extracts was determined by Bradford assay ( Thermo Scientific ) . All extracts were stored at -80°C until use . Costar EIA/RIA 96-well flat bottom plates were coated overnight at 4°C with 2 . 5μg of soluble antigen/ml in bicarbonate coating buffer ( pH9 . 6 ) . Wells were washed three times ( all washes in 0 . 005% Tween in PBS ) and then blocked for 1hr at 37°C with 1% BSA in PBS . Wells were washed three times , 100ul of sera diluted 1/400 , 1/800 , 1/1600 or 1/3200 was added and incubated for 1hr at 37°C . Following six washes , wells were incubated in the dark with biotinylated total IgG ( Jackson ImmunoResearch ) for 1hr at room temperature . Unbound antibodies were washed off ( six times ) prior to incubating wells in the dark with streptavidin HRP ( BD pharmagen ) for 30 minutes at room temperature . Wells were washed six times prior to development ( 100μl , OPD; Sigma-Aldrich ) for five minutes in the dark before termination with an equal volume of 1M HCl . Absorbance was determined at 492nm using a Biotek synergy H4 ELISA plate reader ( Biotek , USA ) . Data were analysed using Gen5 software ( version 2 ) and GraphPad Prism ( version 6 ) . TLR4 agonist , Lipopolysaccharide ( LPS ) ( Sigma-Aldrich , derived from E . coli O127:B8 ) , and TLR9 agonist , CpG1826 ( Sigma-Aldrich , TCCATGACGTTCCTGACGTT with phosphorothioate linkages ) were diluted in 0 . 9% saline ( Baxter ) . Mice received 75μg of either , 2 hours prior to infection in 200μl volumes via intraperitoneal ( i . p . ) injection using 26G needles . To confirm efficacy of TLR agonists , serum cytokine concentrations at 9 hours post-treatment were assessed using Cytometric Bead Arrays ( BD Bioscience ) according to manufacturers’ instructions . Samples were acquired on a LSRII Fortessa analyser ( BD Biosciences ) and analysed using FCAP Array Version 3 ( BD Biosciences ) software . To assess binding of IgG within immune serum to pRBC in vivo , blood from immune serum ( or control non-immune serum ) recipient mice was diluted ( RPMI , 5 U/ml heparin sulphate ) and for positive controls only , incubated with an anti-mouse RBC , mouse IgG2a antibody , clone 34-3C ( 1 . 5μg/ml; Abcam , Cambridge , MA , USA ) . RBC were washed three times in cold FACS buffer ( 1% w/v BSA , 5mM EDTA in PBS ) prior to incubation on ice with detection reagent , FITC-conjugated goat anti-mouse IgG ( 3μg/ml; Biolegend , San Diego , CA ) . RBC were again washed three times , prior to staining with cell-permeant RNA/DNA stain , Syto84 ( 5 μM; Life Technologies ) and DNA stain , Hoechst 33342 ( 10 μg/ml; Sigma ) , followed by immediate flow cytometric analysis on an LSRII Fortessa analyser ( BD Biosciences ) and FlowJo software ( Treestar , CA , USA ) . To assess in vitro immune serum IgG binding to pRBC , the same protocol was employed , except that RBC were first incubated in vitro with immune serum diluted in RPMI , and then stained as above . To further investigate IgG specificity for parasites , pRBC from infected mice were first fixed and permeabilised to expose parasite structures to antibodies ( BD Cytofix/Cytoperm; BD Life Sciences ) , and were then stained and analysed as above with the exception that reagents were prepared in Cytofix/Cytoperm Wash buffer . Immunofluorescence assays were performed as previously described [46] . Briefly , thin blood smears were made on glass slides , air dried and fixed in ice cold Acetone/Methanol ( 50:50 ) for 10 minutes . The slides were removed from the fixative and air dried prior to performing the assay . The slides were washed three times in 1X PBS and blocked in 3% BSA/1XPBS for 1 hour . The blocking solution was removed and the pooled non immune and immune sera added to the slides and incubated at room temperature for 2hrs ( 1:5dilution in 3%BSA/1XPBS ) . Rabbit anti-GAPDH antibodies were used as a counterstain ( 1:2000 ) . The slides were washed 3 times in 1X PBS before addition of the secondary antibodies . Goat anti-mouse Alexa Fluor 488 and anti-rabbit 568 ( 1:400 ) secondary antibodies were prepared in 3%BSA/1XPBS and incubated for 1 h at room temperature . Slides were washed as described above and the parasite nuclei were stained with DAPI ( 2 μg/mL ) for 10 min at room temperature . The slides were again washed prior to the mounting in p-phenylenediamine antifade and addition of a coverslip . Imaging was performed on a DeltaVision Elite Restorative Widefield Deconvolution Imaging System ( GE Healthcare ) using the oil immersion 100x UPLS Apo ( 1 . 4NA ) objective . Samples were excited with solid state illumination ( Insight SSI , Lumencor ) . The following filter sets with excitation and emission wavelengths were used: DAPI Ex390/18 , Em435/48; FITC , Ex475/28 , Em523/26; TRITC , Ex542/27 , Em594/45; Cy5 Ex 632/22 , 676/34 nm . Identical exposure settings were used for the non-immune and immune sera groups . Images were processed using the Fiji ImageJ software . Host phagocytes were depleted in vivo as previously described [29] , with a single i . v . injection ( via a lateral tail vein using 26G needles ) of 200μl of clodronate-containing liposomes ( www . clodronateliposomes . com ) 2 days prior to passive transfer of serum . To deplete mice of the complement component , C3 , 10μg of Cobra Venom Factor ( CVF ) ( Quidel ) was injected intraperitoneally 16 hr prior to infection ( day -1 ) , and then daily on days +1 , +2 & +3; control mice were injected with saline . To evaluate the efficacy of CVF treatment , C3-depletion efficacy was measured in serum by sandwich ELISA . Costar EIA/RIA 96-well flat bottom plates were coated overnight at 4°C with 2 . 0μg/ml primary antibody ( Goat IgG Fraction to Mouse Complement C3 , MP Biomedicals ) in PBS . Wells were washed three times ( all washes in 0 . 005% Tween in PBS ) and then blocked for 1 hr at room temperature ( RT ) with 1% BSA in PBS . Wells were washed three times , and 100μl of sera in duplicate at 1/9000 and 1/27000 was added and incubated for 1 hr at RT . Following six washes , wells were incubated in the dark with a secondary antibody ( Peroxidase-Conjugated Goat IgG Fraction to Mouse Complement C3 , MP Biomedicals ) for 1hr at RT . Unbound antibodies were washed off six times prior to incubating wells in the dark with streptavidin HRP ( BD Biosciences ) for 30 mins at RT . Wells were washed six times prior to development with substrate ( OPD; Sigma-Aldrich ) for five minutes in the dark before termination with 1M HCl . Absorbance was determined at 492nm using Biotek synergy H4 ELISA plate reader ( Biotek , USA ) . To assess host cell depletion , 200μl of clodronate containing liposomes ( www . clodronateliposomes . com ) was administered to mice by a single i . v . injection ( via a lateral tail vein using 26G needles ) , or control untreated , 3 days prior to assessment by flow cytometry . For flow cytometry analysis , spleens were collected , chopped , collagenase/DNase treated ( 0 . 1% ( 1mg/ml ) Collagenase D and 0 . 05% ( 0 . 5mg/ml ) DNase in RPMI 1640 at room temperature for 30 minutes with constant mixing , then homogenized through a 100μm cell strainer to create a single cell suspension . RBC were lysed by incubating samples with RBC lysing buffer ( Sigma-Aldrich ) for 5 minutes at RT and splenocytes were washed in FACS buffer . Cells were assessed for viability by staining with LIVE/DEAD Fixable Aqua Dead Cell stain kit ( Life Technologies ) , according to the manufacturers’ protocol . Cells were FcγR-blocked ( αCD16/32 ( clone 93 ) ) , and stained for surface markers by incubating with the following antibodies on ice for 20 minutes: B220-PB ( RA3-6B2 ) , Ly6G-APCCy7 ( 1A8 ) , MHC-II-PE ( M5/114 . 15 . 2 ) , CD11b-PercpCy5 . 5 ( M1/70 ) , CD11c-APC ( N418 ) , Ly6c-FITC ( HK1 . 4 ) , F4/80-PeCy7 ( BM8 ) . Samples were fixed in 2% paraformaldehyde and acquired on a BD LSR Fortessa Cell Analyser ( BD Biosciences ) . Data were analysed using FlowJo software ( Tree Star ) . To determine Parasite Multiplication Rate ( PMR ) for those parasites transitioning from Gen0 to Gen1 RBC , which also indicates the average number of new pRBC produced from each injected CTFR+ Gen0 pRBC over the first replication cycle in vivo , we calculated: [Gen1 pRBC at 25h]/[Gen0 pRBC at 1h] . Group means and standard error of the mean are presented in the text . All comparisons of group means were performed using a one-way ANOVA followed by a post-hoc contrast analysis , unless otherwise stated ( using the anova1 . m and multcompare . m functions in MATLAB R2015b ( 8 . 6 . 0 . 267246 ) ) . Analysis of interactions between groups were assessed using a two-way ANOVA analysis ( using the anovan . m function in MATLAB R2015b ( 8 . 6 . 0 . 267246 ) ) . Results from flow cytometric analysis were plotted in figures using GraphPad Prism ( v7 . 02 ) , with statistical comparisons between two groups at a single timepoint performed using non-parametric , Mann-Whitney tests unless stated otherwise . | Malaria occurs when Plasmodium parasites replicate inside red blood cells , with the number of parasitised cells ( pRBC ) correlating with disease severity . Antibodies are highly effective at controlling pRBC numbers in the bloodstream , and yet we know very little about how they function in vivo . Human in vitro studies predict that antibodies may function in a number of ways , including via phagocytes or different complement mechanisms . However , to date it has been challenging to explore how antibodies might control parasite numbers in vivo . Here , we have used a unique method in mice , where clearance and replication of a single cohort of pRBC was closely tracked in the presence of protective antibodies . Surprisingly , antibodies played no role whatsoever in accelerating the removal of pRBC . Instead , antibodies were highly effective at preventing parasites from progressing from one generation of pRBC to the next . This process partly depended on host phagocytes . However , we found no role for complement-mediated direct killing . Together , our in vivo data suggest in mouse models that naturally-acquired antibodies do not clear pRBC , and instead prevent transition from one red blood cell to the next . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"blood",
"serum",
"complement",
"system",
"medicine",
"and",
"health",
"sciences",
"parasite",
"groups",
"immune",
"physiology",
"body",
"fluids",
"immunology",
"parasitic",
"diseases",
"parasitic",
"protozoans",
"parasitology",
"parasitemia",
"animal",
"models",
"apicomplexa",
"model",
"organisms",
"protozoans",
"experimental",
"organism",
"systems",
"antibodies",
"quantitative",
"parasitology",
"research",
"and",
"analysis",
"methods",
"immune",
"system",
"proteins",
"malarial",
"parasites",
"animal",
"studies",
"proteins",
"merozoites",
"mouse",
"models",
"immune",
"system",
"biochemistry",
"eukaryota",
"blood",
"anatomy",
"physiology",
"biology",
"and",
"life",
"sciences",
"immune",
"serum",
"organisms"
] | 2019 | Plasmodium-specific antibodies block in vivo parasite growth without clearing infected red blood cells |
Aiming at the design of an allosteric modulator of the neonatal Fc receptor ( FcRn ) –Immunoglobulin G ( IgG ) interaction , we developed a new methodology including NMR fragment screening , X-ray crystallography , and magic-angle-spinning ( MAS ) NMR at 100 kHz after sedimentation , exploiting very fast spinning of the nondeuterated soluble 42 kDa receptor construct to obtain resolved proton-detected 2D and 3D NMR spectra . FcRn plays a crucial role in regulation of IgG and serum albumin catabolism . It is a clinically validated drug target for the treatment of autoimmune diseases caused by pathogenic antibodies via the inhibition of its interaction with IgG . We herein present the discovery of a small molecule that binds into a conserved cavity of the heterodimeric , extracellular domain composed of an α-chain and β2-microglobulin ( β2m ) ( FcRnECD , 373 residues ) . X-ray crystallography was used alongside NMR at 100 kHz MAS with sedimented soluble protein to explore possibilities for refining the compound as an allosteric modulator . Proton-detected MAS NMR experiments on fully protonated [13C , 15N]-labeled FcRnECD yielded ligand-induced chemical-shift perturbations ( CSPs ) for residues in the binding pocket and allosteric changes close to the interface of the two receptor heterodimers present in the asymmetric unit as well as potentially in the albumin interaction site . X-ray structures with and without ligand suggest the need for an optimized ligand to displace the α-chain with respect to β2m , both of which participate in the FcRnECD–IgG interaction site . Our investigation establishes a method to characterize structurally small molecule binding to nondeuterated large proteins by NMR , even in their glycosylated form , which may prove highly valuable for structure-based drug discovery campaigns .
In order to discover new chemical drugs , fragment screening followed by structure-based design is an efficient way to sample chemical space and find hits for challenging target classes such as protein-protein interactions [1–3] . In addition to discovering orthosteric ligands , fragment screening has the potential to locate secondary binding sites on a protein that may be exploited for allosteric regulation [4] . In the development process , a methodology that includes detection of allosteric effects is highly welcome . Magic-angle-spinning ( MAS ) NMR has the potential to contribute via the detection of long-range chemical-shift changes when the investigated protein is too large for solution-state NMR and can even not be deuterated . It is applied here to a soluble 42 kDa construct of the neonatal Fc receptor ( FcRn ) within a search for allosteric regulators , employing very fast MAS ( 100 kHz ) . FcRn facilitates new-born humoral immunity by regulating Immunoglobulin ( IgG ) transport across the epithelium [5] . In addition , it has been shown to bind to IgG and Human Serum Albumin ( HSA ) at nonoverlapping sites in a pH-dependent manner ( Fig 1 ) [6 , 7] . This allows maintenance of IgG and HSA homeostasis , accounting for the long serum half-life of both proteins [8–11] . At low pH , the interaction of FcRn with IgG occurs through protonation of ionizable residues , located at the CH2–CH3 hinge of the IgG Fc , which produces transient , intermolecular salt bridges with negatively charged residues on FcRn [12] . The interaction of FcRn with IgG and HSA occurs in acidified early endosomes , diverting the proteins from catabolism and carrying them back to the neutral pH environment of the extracellular compartment . At near-neutral pH , the affinity of the interaction decreases , and the complex dissociates [10 , 13] . Existing as a heterodimer composed of β2-microglobulin ( β2m ) and a membrane-anchored α-chain ( Fig 1 ) , FcRn is homologous to the class I major histocompatibility complex ( MHC1 ) [14] . MHC1 presents antigenic peptide fragments to the T cell receptor ( TCR ) in complex with cluster of differentiation 8 ( CD8 ) [14] . In contrast , FcRn is not involved in endogenous peptide presentation to the TCR; its peptide-binding groove is closed and nonfunctional [14] . The effect of FcRn loss has been studied using β2m-deficient mice , which develop normally but are defective in T cell–mediated cytotoxicity [15] . Deletion of β2m precluded FcRn expression , resulting in a reduction in IgG half-life [16–18] . Additionally , random mutation of residues proximal to the IgG–FcRn binding site allowed selection of Fc variants with increased affinity for FcRn at pH 6 . These Fc constructs maintained pH-dependent binding and showed an extended half-life relative to wild-type IgG [19] . FcRn has been proposed as a drug target in the treatment of autoimmune diseases , in which pathogenic autoantibodies are detrimental to health [20] . Examples of such diseases include , but are not restricted to , myasthenia gravis , Guillain–Barré syndrome , and dermatomyositis [21–24] . Antibodies produced against the FcRn heavy chain ameliorated myasthenia gravis symptoms in rats , and it has been shown that mice without FcRn are resistant to autoimmune disease [25–27] . The current treatment for these conditions is intravenous application of immunoglobulin , which increases the turnover of pathogenic IgG by saturating FcRn [26] . Recently , rozanolixizumab , an antihuman FcRn monoclonal antibody , reduced the serum IgG concentration in a randomized phase 1 study , providing clinical evidence for the potential of an anti-FcRn therapeutic [28] . A chemical inhibitor to control FcRn trafficking is therapeutically desirable for potential treatment of autoimmune disorders . Orthosteric peptide inhibitors with in vivo efficacy have previously been reported [29–33] . Additionally , structure–activity relationships of small molecule antagonists using ELISA assays have also been described [34] . However , to date , no allosteric modulators of FcRn–IgG or FcRn–HSA interactions have been reported . Should such molecules be found , they could be used therapeutically or as tools in biomedical investigations . In this study , we present the discovery of a compound that binds to the soluble extracellular domain of FcRn , abbreviated here as FcRnECD . In the initial screening process , complementary in silico methods were used to predict binding sites on the basis of charge and topography or sequence conservation , alone and in combination [35–37] . The interaction of the ligand with FcRnECD is investigated by a combination of X-ray crystallography and proton-detected , ultrafast MAS NMR , revealing its localization in an evolutionarily conserved binding pocket . At high MAS frequencies of 100 kHz , we could acquire well-resolved proton-detected NMR spectra on sedimented , fully protonated FcRnECD , allowing de-novo chemical-shift assignments of residues in the binding pocket and in β2m through establishing sequential connections . MAS NMR pinpoints chemical-shift changes upon ligand binding close to the binding site and in regions distant to it . In this context , MAS NMR helps to exclude the influence of crystal packing effects by making use of protein solutions . Both crystal structures with and without ligand were available , but structural changes were not obvious when using global fitting procedures , with the molecules possibly “locked” into a conformation by crystal contacts . To identify sections of the structure not affected by ligand binding for a fit that is also sensitive to small allosteric effects , residues that did not show chemical-shift changes were used to produce “chemical-shift–informed” overlays of FcRnECD X-ray structures with and without ligand . Our findings suggest that therapeutic intervention in autoimmune diseases may be achieved through allosteric small molecules that bind to FcRnECD . In addition , such compounds could potentially be used as chemical probes to study FcRn trafficking . The presented approach highlights further the use of MAS NMR for detecting structural changes in nondeuterated proteins expressed in mammalian cells upon ligand binding .
Small molecule binding sites on proteins can be identified based on surface shape , charge , and functionality . We used SiteMap software to identify binding sites on the FcRnECD [35] . To obtain structural data for our analysis , diffraction data for FcRnECD crystals , at pH 3 and pH 8 . 5 , were collected at cryogenic temperatures and structures solved to 2 . 0 Å and 2 . 45 Å , respectively , by molecular replacement . Two copies of FcRnECD were found in the asymmetric unit . In addition to the structures at acidic and basic pH , a pH 7 . 2 structure was generated using molecular dynamics simulation ( S1 Fig and S1 Text ) . The SiteMap software detected a number of regions with a druggability score >1 . 0 , a nominal quantifier indicating that nM binding might be achieved with a conventional small molecule of <500 Da molecular weight . In particular , SiteMap identified a ligandable site at the interface between the α-chain and β2m ( S1 Fig and S1 Text ) . The respective boundaries were predicted to vary with pH , with the area described as either one large or three distinct cavities . No regions near the IgG binding site were found by SiteMap . Based on our analysis , should an orthosteric pocket for an IgG blocking small molecule exist , it is likely to be transient in nature and not stabilized in the crystal lattice . On the premise that evolutionarily conserved cavities may have an associated function , we evaluated such conservation of the cavities found by SiteMap ( S2 Fig ) . We identified ortholog sequences using OrthoDB and performed sequence alignments using Clustal Omega [38 , 39] . A homology model was then created for each sequence using MEDELLER that enabled us to visualize mutations according to the evolutionary conservation in PyMOL [40 , 41] . Our sequence analysis broadly supports the notion that regions of the protein important for structural integrity or function are conserved , in particular the interface between the α-chain and β2m . It contains key contacts , such as those between D53β2m , Q34α-chain , and S37α-chain , and conservation of these contacts preserves the structural integrity of the noncovalently linked heterodimer . Additionally , motifs within the HSA binding site were also conserved . One of them has previously been identified as being critical for albumin binding and was the site of a protein contact in our pH 3 and pH 8 . 5 crystal structures , between the two copies of FcRnECD in the asymmetric unit [8] . We were initially surprised to see the IgG binding site is poorly conserved with the exception of key residues , such as D130α-chain . This may be attributed to the heterogeneity in Fc moieties between mammalian orthologues , which presumably reduces species’ cross-reactivity . Of the regions identified by SiteMap , the central cavity was highly conserved , potentially due to the proximity to the dimer interface . Our in silico analysis suggests the presence of an evolutionarily conserved binding site at the interface between the α-chain and β2m . To find ligands for this cavity and , potentially , other sites , a solution-state NMR fragment screen was performed . For this purpose , we used an in-house library of approximately 1 , 100 molecules containing fluorine atoms to enable ligand interrogation by 19F Carr-Purcell-Meiboom-Gill NMR [42 , 43] . This screen detected 143 potential binders . In order to select compounds for crystallographic studies , active molecules were further tested in Saturated Transfer Difference NMR experiments and by Surface Plasmon Resonance ( SPR ) , yielding an estimated KD of 80 μM for the highest affinity fragment , the racemate UCB-FcRn-84 ( S3 Fig and S1 Data ) [44 , 45] . Chiral separation showed preferential binding of the R enantiomer . Competition of UCB-FcRn-84 with IgG was tested in a FcRnECD–IgG FRET assay; however , no measurable inhibition was observed . Confirmed hits from the fragment-screen were soaked into FcRnECD crystals that had been grown under acidic conditions . Based on the derived crystal structures , we aimed to further improve both affinity and solubility of the ligand . UCB-FcRn-84 was found in the conserved cavity at the interface of β2m and the α-chain ( S4 Fig and S2 Text ) , as predicted by SiteMap . Key properties of binding interactions include burial of the 3-fluorophenyl group in a hydrophobic cavity composed of Y26β2m , S52β2m , Y63β2m , L65β2m , W29α-chain and P228α-chain . The most visually striking feature of the binding pocket is a tunnel-like cavity that extends through the middle of the protein ( S5 Fig and S2 Text ) . The bicyclic ring of UCB-FcRn-84 occupies this cavity , being involved in one direct and one water-mediated hydrogen bond to the main-chain of Q34α-chain . As the fragment was centrally located in this region , it afforded us a tractable chemical platform to explore the binding site further ( S4 Fig and S2 Text ) . We attempted to improve the affinity of UCB-FcRn-84 by better occupying space around the 3-fluorophenyl group , which was nested in a well-defined hydrophobic pocket . Notably , we identified a small area available for growth adjacent to the unsubstituted meta position of the 3-fluorophenyl group . After scanning a range of disubstituted phenyl derivatives , we identified 3 , 5 difluorophenyl as the best option , yielding a ligand with 2 . 4 μM affinity for the racemate . Moreover , adding a 3-pyridile group in position 5 led to an equipotent compound ( UCB-FcRn-303 ) but with improved solubility ( S5 and S6 Figs , S1 Data and S2 Text ) . The crystal structure of UCB-FcRn-303 bound to FcRnECD displays the R enantiomer ( Fig 2 and S5 Fig ) . The binding mode is consistent with the position observed for UCB-FcRn-84 ( Fig 2A and 2B , S4 and S5 Figs ) . The di-substituted phenyl ring better fills the hydrophobic pocket in the α-chain/β2m interface , while hydrogen bond interactions are maintained with the α-chain and the local water network . In biochemical experiments molecules from our series did not inhibit IgG binding when tested in a FcRnECD–IgG FRET competition assay . This was consistent with our crystallographic studies which showed no discernible change in the IgG binding site , although accurate interpretation of small changes may be hampered by crystal contacts . The FcRnECD crystallized with two copies of the heterodimer present in the asymmetric unit , packing in an anti-parallel fashion ( Fig 2A ) . Overlay of the IgG heavy chain–FcRnECD complex ( Protein Data Bank [PDB] code 1FRT ) with the ligand-free crystal structure shows significant overlap between CH2 and CH3 domains of IgG and symmetry related copies of FcRnECD ( S7 Fig ) [46] . In order to determine whether the ligand might induce small conformational and/or dynamic changes in regions beyond the direct binding site , we have used orthogonal solution-based techniques , which will avoid conformational restrictions imposed by crystal packing . Until recently , solution-state and proton-detected MAS NMR typically required deuteration of proteins in the size-range of the 42 kDa FcRnECD , as the signal-to-noise ratios obtained in triple-resonance experiments involving 13Cβ and 13Cα chemical-shifts critically depend on the T2 of 13C coherences [47 , 48] . Since we were unable to produce deuterated FcRnECD in sufficient amounts it was not possible to acquire 3D solution-state NMR spectra that would allow sequential assignments . The spectral quality of a 2D 15N-1H correlation spectrum recorded in solution , however , is remarkably high ( S8 Fig ) [49] . It indicates that FcRnECD most likely does not form a stable dimer of heterodimers in solution at the applied concentrations , which agrees with biochemical data and earlier studies [50] . At least 227 of the expected 349 backbone amide signals could be observed , with the remainder of the signals most probably obscured by overlap . To circumvent these experimental limitations , we applied MAS NMR . Progress in recent years , in particular ever faster sample spinning , enables the acquisition of proton-detected MAS NMR spectra on fully protonated samples which increases sensitivity and facilitates resonance assignments through triple-resonance experiments [51–54] . At the highest MAS frequency routinely available ( 110 kHz ) , well resolved proton spectra have been reported for fully protonated microcrystalline and membrane proteins , sedimented assemblies and fibrillar proteins [55–57] . In the present case , we have investigated soluble FcRnECD by MAS NMR experiments via sedimentation . For this purpose , the availability of dedicated filling tools is a prerequisite [58 , 59] . With the help of such appliances , we sedimented the 42 kDa soluble , fully protonated [13C , 15N]-labeled FcRnECD by ultracentrifugation directly into a 0 . 7 mm MAS rotor ( Fig 3A ) . The 2D 15N-1H correlation spectrum measured at 100 kHz MAS is shown in Fig 3B . The observed linewidths demonstrate a remarkable spectral quality by MAS NMR standards ( Fig 3C ) . Recorded at a magnetic field of about 20 T , the 1H and 15N linewidths of a selected typical cross peak are 184 Hz and 41 Hz , respectively . The amide 1H linewidth of FcRnECD is still higher than is observed in the model protein GB1 , by a factor of approximately 1 . 8 , but similar to the precipitated viral capsid coat protein AP205 [60] . The spectrum displays a large number of unresolved signals and a considerable fraction of signals that are well dispersed , both of which may be assigned on the basis of suitable triple-resonance spectra . In order to monitor the folding state of FcRnECD in the sedimented sample , we compared the MAS 15N-1H spectrum to the 15N-1H correlation recorded in solution ( S8 Fig and S3 Text ) . The agreement of the spectra indicates that the structure of FcRnECD in solution is very similar to the one in the sedimented sample . FcRnECD did not form the stable dimers of heterodimers in solution as had been observed in the crystal structure . Analytical ultracentrifugation , however , reveals a concentration-dependent protomer–diprotomer equilibrium of the FcRnECD heterodimer in solution , with a very low populated diprotomer fraction ( S9 Fig and S4 Text ) . It is possible that this equilibrium is changed in the sedimentation process which may explain small chemical-shift differences . However , due to the signal overlap in the 15N-1H MAS NMR spectrum , a more detailed chemical-shift comparison is not possible . To allow interpretation of chemical-shift perturbations ( CSPs ) upon binding of UCB-FcRn-303 to FcRnECD , resonance assignments are critical . The high MAS frequencies now available facilitate an assignment procedure based on triple-resonance MAS NMR experiments as in solution-state NMR . These include ( H ) CANH , ( H ) CBCANH , ( H ) CA ( CO ) NH , and ( H ) CBCA ( CO ) NH spectra , yielding assignments of 15N , 1HN , 13Cα0 , and 13Cβ chemical-shifts [61 , 62] . If it is not possible to obtain a ( H ) CBCA ( CO ) NH spectrum with sufficient signal-to-noise due to too short 13Cα and 13CO relaxation times , the first two experiments allow for identification of amino acids that , in a following step , can be sequentially connected according to the protein sequence by using a ( H ) CA ( CO ) NH spectrum . In this study , such spectra were acquired on the fully protonated [13C , 15N]-labeled FcRnECD enabling assignments of 25 α-chain residues close to the binding pocket and of 73% of all β2m residues . To ease the assignment of β2m resonances , we made use of data in Beerbaum and colleagues , in which [2H , 13C , 15N]-labeled β2m was investigated in complex with unlabeled MHC1 [63] . However , the final assignments were achieved by establishing sequential connections along the protein backbone , as shown for the sequence from K41β2m to R45β2m ( Fig 4 ) . All chemical-shift assignments are listed in S1 Table and S2 Data and are deposited in the Biological Magnetic Resonance Data Bank ( BMRB ) ( accession number 27437 ) . In total , signals of 98 residues were assigned this way ( S1 Table , S2 Data , and S5 Text ) . The more sensitive ( H ) CANH spectrum contained 84 additional resolved cross peaks that could not be assigned further due to overlap or lacking correlations in the less sensitive ( H ) CA ( CO ) NH and ( H ) CBCANH spectra . Interestingly , a large number of the measured 13Cα , 13Cβ , 15N , and 1H chemical-shifts of β2m in the FcRnECD heterodimer match the observed solution-state NMR resonances for deuterated β2m in MHC1 complexes of the previous study ( S1 Table and S5 Text ) [63] . Differences in chemical-shifts can be explained by 1H/2H isotope effects , the limited number of amino acid substitutions in MHC1 , potential dimers of FcRnECD heterodimers in the sedimented sample , and differences in buffer and temperature [64] . The similarity of the chemical-shifts supports the notion that β2m adopts its globular fold in the sedimented sample of FcRnECD . Based on these findings and on the similarity between the solution-state 15N-1H and MAS 15N-1H spectra , we assume the overall FcRnECD structure to be highly similar in solution and in sedimented samples ( S8 Fig , S3 Text , S1 Table and S5 Text ) . In summary , ( H ) CANH , ( H ) CA ( CO ) NH , and ( H ) CBCANH spectra represent an acceptable basis for obtaining backbone resonance assignments and allowed us to exploit CSPs as monitors for structural alterations in FcRnECD upon UCB-FcRn-303 binding . CSPs ( Δδ ) are probes for both direct effects of ligand binding and concomitant long-range structural changes in receptor proteins that may hint at allosteric effects . Since differences between X-ray structures with and without ligand may be masked by crystal contacts , we analyzed this possibility by comparing 3D ( H ) CANH spectra recorded on samples of FcRnECD with and without UCB-FcRn-303 ( Fig 5A and 5B ) . The introduction of a third dimension compared to a 2D 15N-1H correlation leads to better spectral resolution . The observed minimal chemical-shift differences are displayed in Fig 5A . Overall , many of the assigned signals show very similar chemical-shifts in both spectra ( Δδ < 0 . 02 ppm , white in Fig 5C and 5D ) , whereas signals that shift significantly ( 0 . 02 ppm < Δδ < 0 . 03 ppm , cyan; 0 . 03 ppm < Δδ < 0 . 04 ppm , marine; 0 . 04 ppm < Δδ , dark blue in Fig 5C and 5D ) appear well clustered , suggesting a conformational/dynamic change extending from the ligand binding site . In the following discussions , we consider all residues with Δδ < 0 . 02 ppm as not perturbed . Selected cross peaks of strongly affected residues ( 0 . 04 ppm < Δδ ) are shown in 2D planes of the ( H ) CANH spectra with and without ligand ( Fig 5B ) . The calculated CSPs shown in Fig 5A are based on the chemical-shift changes in all three spectral dimensions . Large CSPs displayed in Fig 5A may therefore not be obvious from the 2D planes shown in Fig 5B , see D96β2m . As expected , strong effects are seen in the vicinity of the binding pocket of UCB-FcRn-303 observed by X-ray crystallography ( G232α-chain , C48α-chain , G49α-chain and S52β2m , Fig 5A and 5C ) since the chemical environment of these residues is altered upon binding . More interestingly , several residues that experienced changes in their chemical-shifts are distant from the binding pocket . These include , for example , E54α-chain at the C-terminal end of the short α-helix close to the binding region ( Fig 5C and 5D ) . Such an effect could potentially be explained by a small movement of the α-helix . Furthermore , strong CSPs can be observed in a rather remote region composed of loops in β2m , involving D96β2m , G18β2m , K19β2m , S20β2m , and H13β2m ( Fig 5C and 5D ) . These residues cluster at the surface close to or at the interface of β2m and the α-chain of two different heterodimers seen in the asymmetric unit ( Fig 6A and 6B ) . A second ligand binding site in this region of FcRnECD can be excluded since stoichiometric ratios >1 were not observed in SPR experiments or by X-ray crystallography ( Fig 2A , S3 , S4 , S5 , S6 Figs , and S1 Data ) . Interestingly , strong shift changes were observed at comparably long distances from the small molecule binding site . Those strong CSPs may be direct effects of ligand binding , such as introducing a slight displacement of the β2m subunit with respect to the α-chain . Alternatively , it is possible that changes in the protomer–diprotomer equilibrium occur , and thus the chemical-shifts of residues at a potential diprotomer interface are affected . Although we do not observe a stable dimer of heterodimers in solution , a small fraction is present as seen in analytical ultracentrifugation ( AUC ) experiments ( S9 Fig and S4 Text ) . A diprotomer may also be more highly populated at very high protein concentrations after the applied sedimentation through ultracentrifugation in preparation of 100 kHz MAS experiments . However , the samples with and without ligand have been prepared under identical conditions , making structural effects unrelated to ligand binding unlikely . For a ligand-induced alteration of the equilibrium , long-range structural changes towards the potential diprotomer interface are required , highlighting again the possibility for the occurrence of allosteric effects upon ligand binding . In the HSA interaction site , a number of strongly affected residues could be found . These included R12β2m , H13β2m , C48α-chain , G49α-chain , A50α-chain , and E54α-chain with CSPs around 0 . 04 ppm or larger ( Figs 5A , 6C and S2 Data ) . Some of these amino acids are close to the binding site of UCB-FcRn-303 , but especially R12β2m , H13β2m , and E54α-chain are distant to it and may be therefore structurally altered through allosteric effects ( Figs 5C and 6C ) . It is possible that the FcRn–HSA interaction can be modulated with UCB-FcRn-303 . The binding site to HSA , however , is located at the possible diprotomer interface and CSPs in this region may be caused by potential ligand-induced changes in the protomer–diprotomer equilibrium of FcRnECD as discussed above . Of the NMR assigned residues , only a few from β2m are found in the IgG binding area , such as R3β2m and T86β2m ( Fig 6C ) , for which negligible chemical-shift changes occur ( Fig 5A ) [46] . In general , residues in the β-sheet region of β2m distant from the binding site or the interface between heterodimers do not exhibit notable CSPs ( Fig 5D ) . In order to further investigate the potential for allosteric interference with IgG binding , we generated chemical-shift–informed overlays of the two FcRnECD crystal structures , with and without UCB-FcRn-303 . All residues with Δδ < 0 . 02 ppm ( Fig 5A ) were used for fitting , including the area around R3β2m , T86β2m , and L87β2m in β2m that is involved in IgG interaction ( Fig 6C ) , yielding a root mean square deviation ( r . m . s . d . ) of 0 . 095 Å . If a movement of the α-helical region of the α-chain could be induced upon ligand binding , it may disseminate towards the loop from D130α-chain to E133α-chain , which is part of the binding site to IgG ( Fig 6C ) [46] . Unfortunately , the overlay does not reveal any significant structural changes of α-chain residues close to this loop , except for subtle differences , which we are unable to isolate as a compound-induced effect due to its proximity to crystal packing interactions . Still , the CSPs observed distant to the ligand cavity reveal the potential for allosteric effects in FcRnECD induced by UCB-FcRn-303 by affecting the HSA interaction site . This provides evidence that binding of an optimized ligand could be aimed at displacing the β2m and α-chain subunits and distally disrupting the IgG binding interface of FcRnECD . In the context of a druggability study , we identified a compound of <500 Da molecular weight , UCB-FcRn-303 , that binds to the extracellular domain of FcRn with low μM affinity . The conserved binding site is located at the interface of β2m and the α-chain , featuring a tunnel-like cavity with solvent access from two different sides . The pocket was identified by computational chemistry methods in a theoretical druggability examination and was corroborated by fragment screening , X-ray crystallography , and NMR spectroscopy . The respective crystal structures show a dimer of heterodimers , with a 1:1 stoichiometry for the ligand/protein complex . Due to the distance of about 35 Å between the small molecule binding pocket and the site of IgG interaction ( Fig 6C ) , substantial allosteric effects would be required for pharmacologically relevant interference with IgG binding [46] . Such allosteric effects can be monitored by comparing X-ray structures with and without ligand and by NMR spectroscopy via the analysis of ligand-induced CSPs . We therefore established a new approach for testing ligand binding effects on soluble proteins , employing proton-detected 100 kHz MAS NMR spectroscopy on fully protonated FcRnECD . After successful sedimentation of the protein by ultracentrifugation using dedicated filling tools , remarkably well-resolved proton-detected MAS NMR spectra could be obtained and partially assigned . CSPs upon UCB-FcRn-303 binding are observed around the small molecule binding pocket but also distant residues are affected , however , overlapping with the HSA interaction site . Since a large fraction of residues in β2m towards the IgG binding site and some in the α-chain do not show substantial CSPs ( Figs 5A , 6A and 6C ) , we generated alignments of the two X-ray structures with and without ligand , superposing the nonshifting residues . This approach revealed no substantial changes of residues close to the IgG interaction site upon FcRn-UCB-303 binding , which could modulate the FcRnECD–IgG interaction . It may be envisaged that an optimized ligand could produce a shift of the α-helical region in the α-chain , with potential effects on IgG interaction . Such an optimization could include derivatives with additional functional groups that enter the α-chain/β2m interface and thus produce a slight displacement of the two with respect to each other . Strong CSPs could be observed in the HSA interaction site of FcRnECD , highlighting the potential to achieve a functional modulation of FcRn with UCB-FcRn-303 . However , it cannot be fully excluded that changes in oligomeric state of FcRnECD caused by the ligand under MAS conditions lead to the observed CSPs in this region . The presented MAS NMR methodology provides an appealing approach for structural investigations of large , soluble proteins expressed in mammalian or other types of eukaryotic cells in which deuteration is challenging , especially in cases when glycosylation is crucial . Moreover , it extends the range of NMR applications to pharmacologically relevant targets , which are often inaccessible by solution-state NMR methods due to size . The MAS NMR approach works with high molecular weight targets and is independent of the physical state of the sample , providing spectral complexity can be handled , e . g . , by applying appropriate labelling concepts . This facilitates protein-ligand interaction studies by NMR for any type of protein or biomolecular complex , making previously intractable pharmacological targets accessible , such as the ribosome , G protein-coupled receptors , and the like . Our investigation shows that MAS NMR complements X-ray crystallography in structure-based drug discovery campaigns . It is particularly useful to explore allosteric changes beyond small molecule binding sites , which may be difficult to observe by X-ray crystallography .
The coding sequences of extracellular domain ( ECD ) ( amino acids 1–297 ) of human FcRn α-chain and human β2m were synthesized by Entelechon ( Entelechon , Regensburg , Germany ) . The α-chain fragment was cloned into the expression vector pMH ( UCB ) and the β2m sequence was cloned into pMK ( UCB ) as HindIII and EcoRI fragments . Both vectors were digested with SalI and NotI and the relevant fragments excised and ligated to generate a vector containing both the α-chain and β2m genes ( pM-ECDFcRn-B2M ) . The vector was further digested with SalI and a neomycin cassette was ligated in to generate a double gene vector with antibiotic selection ( pMFcRnECD-B2M-Neo ) . HEK293 cells were transfected with pMFcRnECD-B2M-Neo using 293fectin ( Thermofisher ) according to the manufacturer’s instructions . Transfected cells were incubated in a static incubator at 37 °C and 8% CO2 for 24 hours . The cells were then diluted to the required concentration with medium supplemented with 0 . 5 mg/L G418 ( Invitrogen ) and subsequently divided into 1-mL pools in 24-well plates and incubated in a static incubator at 37 °C with 8% CO2 . Every 7 days , the medium was removed from each well and replaced with fresh medium . After a further 14 days , the pools that exhibited cell growth were transferred to 6-well plates . These were expanded up to 50-mL cultures in E250 flasks in shaking incubators . To determine total expression of the FcRnECD heterodimer composed of β2m and α-chain , batch overgrows were set up and incubated for 10 days . Samples of the 50-mL batch overgrow supernatants were analyzed for FcRnECD expression using western blotting . 15 μL of each supernatant was run on a ( denatured Tris/Gly gel western blot ) alongside known concentrations of purified human FcRn as a control . The highest expressing pool was expanded ( without Neomycin selection ) up to 10-L scale in a Wave Bioreactor 20/50 EHT at 37 °C , 8% CO2 for 4 days , at which point the temperature was reduced to 32 °C and incubated for a further 6 days . The cell culture was harvested by centrifugation ( 1 , 000x g for 1 hour ) and supernatant put through a 0 . 22-μm filter . Mammalian cell culture supernatant containing the heterodimer FcRnECD was concentrated with a 10 , 000 MWCO membrane using tangential flow filtration ( Centramate , Pall ) or Amicon stirred cell ( Millipore ) , depending on scale . The sample buffer was exchanged into 50 mM sodium phosphate , pH 5 . 8 , 30 mM NaCl by diluting and concentrating . FcRnECD was loaded onto a KappaSelect ( GE Healthcare ) column , which had been preloaded with an IgG4 monoclonal antibody to bind FcRnECD , before washing with 50 mM sodium phosphate , pH 5 . 8 , 30 mM NaCl , and eluting with 50 mM sodium phosphate , pH 8 . 0 , 30 mM NaCl . Elution fractions were analyzed by SDS-PAGE and relevant fractions pooled . The sample was concentrated using an Amicon spin concentrator ( Millipore ) , 10 , 000 MWCO membrane . The protein was purified by Gel Filtration with a Superdex 200 ( GE Healthcare ) column using 25 mM sodium phosphate , pH 7 . 4 , 100 mM NaCl . Peak fractions were analyzed by SDS-PAGE before pooling and concentrating if required . SPR was carried out using BIAcore 4000 instruments ( GE Healthcare ) . Reagents including CM5 sensor chips , N-hydroxysuccinimide ( NHS ) , N-ethyl-N- ( 3-dimethylaminopropyl ) carbodiimide ( EDC ) , and ethanolamine HCl , 10 mM sodium acetate buffers ( pH 5 . 0 , pH 4 . 5 ) and HBS-P ( 10x buffer ) were obtained from GE Healthcare . FcRnECD was diluted into 10 mM sodium acetate buffer , pH 5 . 0 and immobilized on a CM5 Sensor Chip via amine coupling chemistry to a capture level of approxmimately 4 , 000 response units . Compounds were screened in a 10-point titration from 200 μM , at 2% DMSO , pH 6 . 0 , and the surface was re-immobilized after each 384-well plate . Injections were performed at a flow rate of 10 μL/min . All data were double-referenced for blank injections and reference surface , following standard procedures . Both data processing and fitting were performed using Activity Base template protocols developed in house . A library of approximately 1 , 100 fluorine-containing fragments were cocktailed into groups of 12 ensuring no overlap of 19F signals . Cocktails were initially prepared at a concentration of 4 . 2 mM in d6-DMSO and diluted to 800 μM in PBS pH 7 . 4 before a final dilution to 40 μM ligand concentration ( 1% d6-DMSO ) in either PBS containing 10% D2O ( for control samples ) or 20 μM FcRnECD containing 10% D2O ( for protein samples ) . NMR spectra were acquired at 25 °C on a Bruker 600 MHz AVIII-HD spectrometer equipped with a QCI-F cryoprobe and a SampleJet autosampler . Data were collected using a CPMG pulse sequence with a total echo time of 160 ms across a sweep width of 126 ppm with an acquisition time of 1 s . All spectra were processed using TopSpin 3 . 2 . Fragments were considered binders to FcRnECD when the 19F signal intensity was significantly reduced in the spectra with FcRnECD present compared to the spectra recorded in the absence of protein . STD NMR samples were prepared with a ligand to protein ratio of 50:1 ( 500 μM ligand , 10 μM FcRnECD ) in 500 μL phosphate buffered saline , pH 7 . 4 ( 90% H2O , 10% D2O ) with 5% d6-DMSO to help solubilize the ligand . STD NMR spectra were recorded using a Bruker Avance III HD 600 MHz spectrometer equipped with a 5 mm QCI-F Cryoprobe . Data were acquired and processed using the standard Bruker software and were collected at 298 K . The protein was saturated in the methyl region of the spectrum at 0 ppm , and off-resonance saturation was performed at 33 ppm . A series of 120 EBurp2 pulses ( 50 ms each ) were applied with a 4-μs delay between each pulse , resulting in total saturation time of 6 s . Protein signals were removed by applying a spinlock of 100 ms . Interleaved on- and off-resonance data were recorded , processed separately , and then the difference spectra obtained by subtracting the on- from the off-resonance spectra . Data were zero filled once and an exponential multiplication window function applied ( LB 2 Hz ) . The parent construct for α-chain/β2m expresses two Open Reading Frames ( ORFs ) within one baculovirus multiple-target expression plasmid , pBacugs4X-1 . Protein targets are based on the following amino acid sequences ( both α-chain and β2m each contain their native leaders ) : α-chain 1–297 based on NCBI reference sequence NP_004098 , β2m 1–119 based on NCBI reference sequence NP_004039 . Codons for both ORFs were engineered by GeneComposer for highly expressed baculovirus genes , such that BamHI , HindIII , BglII , and EcoRI restriction sites were eliminated from the inserts to facilitate cloning [65] . Both genes , including flanking restriction sites , were synthesized at GeneArt . The ORF for α-chain was cloned behind the polyhedrin promoter via unique BamHI and HindIII sites; in this case , the BamHI site preceded the signal sequence while HindIII followed the sequence “TGAT” such that two stops are introduced after the C-terminal residue . The ORF for β2m was cloned behind the p10 promoter via unique BglII and EcoRI sites; the BglII site preceded the signal sequence , while EcoRI followed the sequence “TGATAA” such that two stops are introduced after the C-terminal residue . Using this cloning scheme , polyhedrin and p10 promoters are arranged in divergent/opposing orientations within the baculovirus transfer vector . ORFs were sequence verified prior to the commencement of expression studies . The α-chain/β2m construct was transfected into Sf9 insect cells ( Expression Systems ) using BestBac 2 . 0 , v-cath/chiA Deleted Linearized Baculovirus DNA ( Expression Systems , Cat#-91-002 ) . Virus from each transfection was amplified through 3 rounds to produce virus stock for large-scale production . The large-scale preparations were grown in ESF921 medium ( Expression Systems , Cat#96–001 ) . Large-scale preparations were infected using the titerless infected-cells preservation and scale-up ( TIPS ) method [66] . Approximately 106 Tni cells ( Trichopulsia ni , Expression Systems ) per mL were infected using 1 mL of TIPS cells . Secreted proteins were harvested after 2–3 days by Tangential Flow Filtration ( Spectrum KrosFlo , 0 . 2 μm filter Cat# P-NO2-E20U-05-N ) . Harvested baculovirus medium ( Trichoplusia ni ) containing secreted FcRnECD was concentrated 10-fold and buffer exchanged ( 50 mM sodium phosphate pH 5 . 8 and 30 mM NaCl via Tangential Flow Filtration ( TFF ) ( Spectrum Labs ) . The concentrated medium was centrifuged using a JA-10 rotor at 9 , 000 RPM for 15 minutes at 4 °C and then filtered through a 0 . 2 μm bottle top filter . Three complete EDTA free protease inhibitor tablets were added to the concentrated media prior to chromatography . The filtered concentrated medium was applied to 35 mL of IgG Sepharose FF resin ( GE Healthcare ) equilibrated with 50 mM sodium phosphate buffer , pH 5 . 8 , and 30 mM NaCl ( Sigma ) and rotated end over end for one hour at 4 °C . The resin was then poured into a gravity flow column and washed with 10 column volumes of the same equilibration buffer . FcRnECD was eluted from the resin with eight column volumes of 50 mM sodium phosphate buffer pH 8 . 0 and 30 mM NaCl . The elution fractions containing FcRnECD were pooled and loaded onto two 5-mL HiTrap Q FF columns ( GE Healthcare ) and eluted over a 1 M NaCl gradient . The fractions of interest were pooled and glycerol was added to a final concentration of 10% . The pool was concentrated to 15 mg/mL via centrifugal concentration ( Amicon Regenerated Cellulose , 10 kDa MWCO , Millipore ) and further purified via size exclusion chromatography over a HiLoad 16/600 Superdex 200 pg ( GE Healthcare ) column in 50 mM HEPES ( 4- ( 2-hydroxyethyl ) -1-piperazineethanesulfonic acid ) pH 7 . 0 and 75 mM NaCl . The fractions containing FcRnECD were pooled and concentrated for crystallography via centrifugal concentration ( Amicon Regenerated Cellulose , 10 kDa MWCO , Millipore ) to 9 . 9 mg/mL prior to being aliquoted and flash frozen in liquid nitrogen for later use in crystallization experiments . In order to search for crystallization conditions for FcRnECD expressed in insect cells , sitting-drop vapor diffusion crystallization trials were set up at 291 K using a variety of commercial spare-matrix ( Rigaku Reagents: JCSG+ , Wizard 1/2 , Wizard 3/4; Hampton Research: Crystal Screen HTIndex; Molecular Dimensions: PACT , Morpheus , Proplex; Microlytics: MCSG1 ) using 0 . 4 μL of protein solution at 5 mg/mL that were mixed with 0 . 4 μL of reservoir solution and equilibrated against 80 μL of reservoir solution . The initial crystallization trials produced small kite crystals in several conditions that contain PEG 3350 , PEG 6000 , or PEG 8000 at low pH . Low pH crystals of FcRnECD were produced in an optimized crystallization condition screen ( Rigaku ) containing 12%–16% PEG 6000 , 100 mM Citric Acid/Ammonium citrate tribasic pH 3 . 00–3 . 09 . Crystals of FcRnECD appeared within 24 hours and grew larger overtime , typically to 50–150 microns in size . The crystals were harvested using 20% glycerol as a cryoprotectant and flash frozen in liquid nitrogen prior to data collection . Additionally , small rod-like crystals appeared in a single condition at a significantly higher pH ( condition PACT H3 , 100 mM Bis-Tris Propane/HCl pH 8 . 5 , 200 mM NaI , and 20% PEG 3350 ) [67] . High/neutral pH crystals of FcRnECD were produced in optimized crystallization condition containing 100 mM Bis-Tris Propane/HCl pH 8 . 5 , 200 mM NaI , and 20% PEG 3350 . Crystals of FcRnECD appeared within 48 hours and grew larger overtime , typically greater than 150 microns in size . The crystals were harvested using 20% ethylene glycol and flash frozen in liquid nitrogen prior to data collection . To obtain compound bound crystals , apo FcRnECD crystals grown at pH 3 were soaked for three days in buffer containing 0 . 1 M Citric Acid/NaOH at pH 3 . 0 , 20% w/v PEG 6000 , 20% glycerol , and 12 . 5–20 mM compound dissolved in 100% DMSO . For crystallization of the UCB-FcRn-303 bound structure of FcRnECD , protein expressed in Sf9 insect cells was used . Datasets were collected at Canadian Light Source ( CLS ) on beamline 08ID-1 ( CMCF ) equipped with a Rayonix MX300 CCD X-ray detector and Advanced Photon Source ( APS ) on beamline 21-ID-F ( LS-CAT ) equipped with a Marmosaic 225 CCD X-ray detector . Diffraction data were reduced and scaled with XDS/XSCALE [68] . The structures of FcRnECD at low and high/neutral pH were solved by molecular replacement using a pre-existing structure of the complex . A significant portion of the structure required remodeling; therefore , Phenix . autobuild was run to generate a starting structure for further refinement . All structures were refined using iterative cycles of TLS and restrained refinement with Phenix . refine and model-building using the Crystallographic Object-Oriented Toolkit ( COOT; Version 0 . 8 . 1-pre ) and were validated using Molprobity prior to deposition in the PDB ( IDs 6C97 , 6C98 , 6C99 ) [69–73] . Diffraction data and refinement statistics for apo FcRnECD at pH 3 and both ligand bound structures are listed in S2 Table . The FcRnECD structures at pH 3 and pH 8 . 5 were used as the starting points for molecular dynamics simulation , after adding hydrogen . Protonation states and missing loops were predicted with Maestro ( Schrödinger LLC ) . The structure was solvated in a dodecahedron such that no protein atom was within 10 Å of the edge of the solvent . Monatomic ions were added to a salt concentration of 0 . 15 M . All simulations were 1 , 000 ns in length and were carried out with GROMACS 4 . 6 . 2 [74] . Particle mesh Ewald was used for long-range electrostatics along with 10 Å cutoffs for Coulomb and Lennard–Jones potential functions . FcRnECD was expressed using a stable HEK293 cell line , as described above , but the growth media was replaced with Bioexpress 6000 media ( Cambridge Isotope Laboratories ) . The solution-state 2D 15N-1H TROSY spectrum was acquired from a 0 . 35-mL sample of 300 μM [13C , 15N]-labeled FcRnECD in a 25 mM Na2HPO4 , 100 mM NaCl , 50 μM EDTA , 0 . 02% ( w/v ) NaN3 buffer at pH 7 . 4 containing 5% D2O/95% H2O [49] . NMR data were acquired at 35 °C on a 600 MHz Bruker AVIII HD spectrometer fitted with a cryogenically cooled probe . The spectrum was acquired for 40 minutes with acquisition times of 40 ms in F1 ( 15N ) and 60 ms in F2 ( 1H ) . It was processed using Topspin 3 . 5 ( Bruker Biospin Ltd ) with linear prediction used to extend the effective acquisition time to 60 ms in F1 . A 0 . 5-mL sample of 100 μM [13C , 15N]-labeled FcRnECD in a 25 mM Na2HPO4 , 100 mM NaCl , 50 μM EDTA , 0 . 02% ( w/v ) NaN3 buffer at pH 7 . 4 containing 3% DMSO was directly ultracentrifuged into a 0 . 7 mm MAS NMR rotor at 100 , 000 x g and 4 °C for 40 hours using a swinging bucket ultracentrifuge rotor . Dedicated home-made filling tools were used for this step . Both bottom and top caps of the 0 . 7 mm MAS rotor were glued to avoid the loss of liquid or removal of caps during MAS . Excess protein and liquid after ultracentrifugation was removed before closing the 0 . 7 mm NMR MAS rotor . For experiments to detect CSPs , 3 mM UCB-FcRn-303 was added to a 0 . 5-mL sample of 100 μM [13C , 15N]-labeled FcRnECD in a 25 mM Na2HPO4 , 100 mM NaCl , 50 μM EDTA , 0 . 02% ( w/v ) NaN3 buffer at pH 7 . 4 containing 3% DMSO ( ligand:protein ratio of 30:1 ) . The ligand-bound sample was sedimented into a second 0 . 7 mm MAS rotor under the same conditions as the ligand-free sample . All MAS NMR experiments were performed on a Bruker AVANCE III 850 MHz spectrometer . The spectra were recorded at 80 , 90 , or 100 kHz MAS with a triple-resonance 0 . 7 mm MAS probe ( Bruker ) and referenced to 4 , 4-dimethyl-4-silapentane-1-sulfonic acid ( DSS ) ( see S3 Table for experimental parameters ) . The sample temperature of 283 K was monitored by the frequency of the water resonance line . To achieve this sample temperature , the sample was cooled during MAS using nitrogen gas through a cooling unit ( BCU II , Bruker ) in strong mode with the gas flow set to 400 L/h . ( H ) NH , ( H ) CANH , ( H ) CA ( CO ) NH , and ( H ) CBCANH spectra were recorded using Cross Polarization ( CP ) ( heteronuclear transfers ) and DREAM ( homonuclear transfers ) magnetization transfer steps according to the procedures described in Penzel and colleagues [61 , 75 , 76] . All spectra were processed with Topspin 3 . 5 ( Bruker ) and analyzed with CCPNmr Analysis v . 2 . 4 . 2 . [77] . All assigned chemical-shifts and observed CSPs are listed in S1 Table and S2 Data , and are deposited in the Biological Magnetic Resonance Data Bank ( BMRB ) ( accession number 27437 ) . Sedimentation velocity experiments were performed at 8 °C and 35 , 000 rpm with an An-60Ti rotor using 12-mm Epon 2-sector centerpieces . For each of the three measurements at different FcRnECD concentrations ( 4 μM , 14 μM , and 52 μM ) , a 400-μL sample in 10 mM Na2HPO4 , 137 mM NaCl , 2 . 7 mM KCl , 1 . 8 mM KH2PO4 buffer at pH 7 . 2 was used . The data were analyzed and plotted with the GUSSI implementation and SEDFIT [78] . All solvents and reagents were used as received from commercial suppliers , unless noted otherwise . The compounds were named using the Biovia Draw 2016 package ( IUPAC ) . NMR spectra were recorded on a Bruker Avance III HD 500 MHz or 250 MHz spectrometer . The chemical-shifts ( δ ) reported are given in parts per million ( ppm ) and the coupling constants ( J ) are in Hertz ( Hz ) . The spin multiplicities are reported as s = singlet , d = doublet , t = triplet , q = quartet , dd = doublet of doublet , ddd = doublet of doublet of doublet , dt = doublet of triplet , td = triplet of doublet , and m = multiplet . uPLC-MS was performed on a Waters Acquity UPLC system coupled to a Waters Acquity PDA detector , an ELS detector and an MSD ( Scan Positive: 150–850 ) . Method ( pH 3 ) : Phenomenex Kinetix-XB C18 ( 2 . 1 x 100 mm , 1 . 7 μm ) column . Elution with a linear gradient of Water + 0 . 1% Formic acid and Acetonitrile + 0 . 1% Formic acid at a flow rate of 0 . 6 mL/min . Chiral SFC analysis: Waters Thar 3100 SFC system connected to Waters 2998 PDA detector , Chiralcel OD-H 25 cm . Chiral SFC separation: Water Thar SFC system with a Waters Thar FDM pump , Waters Thar Alias autoinjector , Waters Thar fraction collector and a Waters 2998 PDA detector . | In drug design , a detailed characterization of structural changes induced by drug binding is useful for further optimizing lead compounds . In many cases , structural alterations are distant from the compound binding site , potentially acting through allosteric effects . These allosteric effects are often difficult to observe by static methods , i . e . , X-ray crystallography , but can be monitored by NMR spectroscopy . The latter method , however , has size-limitations when investigating the protein backbone structure in solution-state . To overcome this , we present an innovative approach employing ultrafast magic-angle-spinning ( MAS ) NMR on the extracellular domain of the neonatal Fc receptor ( FcRnECD ) . This is a validated drug target in autoimmune diseases , and we aim to identify and characterize novel compounds to serve as starting points to develop allosteric inhibitors of this receptor . After sedimentation , we could record well-resolved proton-detected MAS NMR spectra of the fully protonated [13C , 15N]-labeled protein , enabling the observation of structural changes . In combination with computational methods , X-ray crystallography , and other biophysical tools , we present new compounds that may be used as allosteric modulators of FcRn after further optimization . The introduced MAS NMR approach can be applied to a large variety of proteins to support structure-based drug design , facilitating the detection of allosteric effects . | [
"Abstract",
"Introduction",
"Results",
"and",
"discussion",
"Materials",
"and",
"methods"
] | [
"biotechnology",
"medicine",
"and",
"health",
"sciences",
"methods",
"and",
"resources",
"crystal",
"structure",
"chemical",
"compounds",
"small",
"molecules",
"nmr",
"spectroscopy",
"immunology",
"condensed",
"matter",
"physics",
"organic",
"compounds",
"crystals",
"materials",
"science",
"centrifugation",
"materials",
"physics",
"sedimentation",
"crystallography",
"materials",
"by",
"structure",
"crystallographic",
"techniques",
"research",
"and",
"analysis",
"methods",
"separation",
"processes",
"immune",
"system",
"proteins",
"structural",
"characterization",
"solid",
"state",
"physics",
"proteins",
"x-ray",
"crystallography",
"chemistry",
"ultracentrifugation",
"physics",
"biochemistry",
"signal",
"transduction",
"cell",
"biology",
"organic",
"chemistry",
"fc",
"receptors",
"biology",
"and",
"life",
"sciences",
"immune",
"receptors",
"physical",
"sciences",
"spectrum",
"analysis",
"techniques"
] | 2018 | Insight into small molecule binding to the neonatal Fc receptor by X-ray crystallography and 100 kHz magic-angle-spinning NMR |
Successful behavior requires selection and preferred processing of relevant sensory information . The cortical representation of relevant sensory information has been related to neuronal oscillations in the gamma frequency band . Pain is of invariably high behavioral relevance and , thus , nociceptive stimuli receive preferred processing . Here , by using magnetoencephalography , we show that selective nociceptive stimuli induce gamma oscillations between 60 and 95 Hz in primary somatosensory cortex . Amplitudes of pain-induced gamma oscillations vary with objective stimulus intensity and subjective pain intensity . However , around pain threshold , perceived stimuli yielded stronger gamma oscillations than unperceived stimuli of equal stimulus intensity . These results show that pain induces gamma oscillations in primary somatosensory cortex that are particularly related to the subjective perception of pain . Our findings support the hypothesis that gamma oscillations are related to the internal representation of behaviorally relevant stimuli that should receive preferred processing .
Within the continuous flow of sensory information , a huge number of events compete for neural representation and perception . This sensory overflow requires the selection and preferential processing of relevant information in order to optimize the utilization of cerebral processing resources . Recently , induced neuronal oscillations in the gamma frequency range ( about 40–100 Hz ) have been suggested to represent one mechanism of the selection and preferred processing of sensory information [1–7] . These induced gamma oscillations represent event-related modulations of neuronal oscillations , are often observed in early sensory cortices and differ from evoked neuronal responses in a lack of phase locking to the sensory stimulus . Functionally , the association between induced gamma oscillations , and selection and preferred processing of sensory stimuli suggests that these responses may not only be related to the physical stimulus attributes , but also related particularly to the subjectively weighted percept of a sensory event . Painful stimuli signal threats and are therefore of utmost behavioral relevance [8 , 9] . Thus , we hypothesized that painful stimuli induce gamma oscillations in somatosensory cortices . Moreover , we speculated that these pain-induced gamma oscillations may not only relate to the objective attributes of painful stimuli , but may also particularly reflect the subjective experience of pain . To address this issue , we used magnetoencephalography to record neural responses to noxious stimuli in healthy human subjects . We investigated the effects of noxious stimuli on neuronal activity in the gamma band and related these effects to objective stimulus intensity and subjectively perceived pain intensity . Our results show that pain induces gamma oscillations in the contralateral primary somatosensory cortex . Amplitudes of pain-induced gamma oscillations increase with objective stimulus intensity and subjective pain intensity . However , around pain threshold , perceived stimuli induce significantly stronger gamma oscillations than unperceived stimuli of equal stimulus intensity . These observations provide direct evidence for a close association between induced gamma oscillations and the conscious and subjective perception of behaviorally relevant sensory events .
First , we aimed to identify and to characterize spatially and temporally pain-induced gamma oscillations in human somatosensory cortices . In 12 healthy male participants , 40 moderately painful cutaneous laser stimuli ( intensity 600 mJ ) were applied to the dorsum of the right hand . Participants were instructed to passively perceive the stimuli without any further task . The contralateral primary ( S1 ) and bilateral secondary ( S2 ) somatosensory cortices were localized by analyzing the well-known [10] pain-evoked ( phase-locked ) responses from these areas ( Figure 1A ) . Next , we investigated possible pain-induced gamma oscillations in these areas . To this end , time-frequency representations ( TFRs ) were calculated for each trial and area . The analysis revealed that pain induces strong and significant increases in gamma power at frequencies between 60 and 95 Hz in the contralateral S1 cortex ( Figure 1B ) . These pain-induced gamma oscillations were observed between 100 ms and 300 ms after stimulus application coinciding with the pain-evoked response from S1 ( Figure 1C ) . Please note that these pain-induced gamma oscillations were observed without any particular task and , thus , do not depend on the task relevance of painful stimuli , but rather on their sensory quality and their inherent behavioral relevance . No pain-induced changes in gamma power were observed in the bilateral S2 cortices . Analysis of amplitude and phase dynamics confirmed that gamma oscillations were not phase locked to stimuli , and therefore represent induced , but not evoked oscillations ( Figure 2 ) . Second , we investigated the relationship between amplitudes of induced gamma oscillations in S1 , stimulus intensity , and perceived pain intensity . To this end , randomly varied intensities of noxious laser stimuli were applied to the right hand of 13 healthy human participants . Possible stimulus intensities were 150 , 300 , 450 , and 600 mJ , which yield sensations ranging from barely detectable to moderately painful . Forty stimuli were presented for each stimulus intensity , and subjects were asked to rate the perception of each stimulus on a scale from 0 to 100 anchored at “no pain” and “worst imaginable pain . ” Pain ratings were cued by an auditory signal presented 3 s after each stimulus . The contralateral S1 cortex was localized , and evoked responses and induced gamma oscillations to stimuli of different intensities were analyzed . Mean amplitudes of evoked responses and induced gamma oscillations from S1 were calculated during the time window from 100 ms to 300 ms as compared to baseline amplitudes . Figure 3 shows amplitudes of evoked ( phase-locked ) responses and induced ( non–phase-locked ) gamma oscillations as a function of stimulus intensity . The results reveal that amplitudes of induced gamma oscillations and amplitudes of evoked responses from S1 increase with stimulus intensity . This increase in response amplitudes was paralleled by an increase in perceived pain intensity . These observations show that amplitudes of pain-induced gamma oscillations in S1 vary with objective stimulus intensity and subjective pain intensity . Third , we aimed at further defining the relationship between pain-induced gamma oscillations in S1 and the subjective experience of pain . Low intensity ( 150 and 300 mJ ) trials around pain threshold were chosen for the analysis . Per definition , some of these trials are perceived as painful ( “percept” trials , rating > 0 ) and some are not ( “no percept” trials , rating = 0 ) . We matched both trials for stimulus intensity and number of stimuli , i . e . , for each individual , the “percept” and “no percept” sets of trials did not differ with respect to objective stimulus intensity or number of stimuli , but only with respect to subjective perception . Mean amplitudes of gamma oscillations at 100 ms to 300 ms as compared to baseline were computed for “percept” and “no percept” trials . The analysis reveals that gamma oscillations were significantly stronger for “percept” than for “no percept” trials ( p = 0 . 013 , Wilcoxon signed-rank test , Figure 4 ) . Gamma power during baseline did not differ between conditions ( p = 0 . 15 ) and only “percept” trials ( p = 0 . 003 ) , not “no percept” trials ( p = 0 . 735 ) , yielded significant increases of gamma power as compared to baseline . In contrast , amplitudes of phase-locked , evoked S1 responses did not differ between “percept” and “no percept” trials ( p = 0 . 38 ) . These observations show that pain-induced gamma oscillations are particularly related to the subjective perception of pain .
Our study demonstrates that painful stimuli induce gamma oscillations in the contralateral S1 cortex . These non–phase-locked gamma oscillations differ from evoked responses in a trial-by-trial jitter in latency , and occur at latencies around 200 ms and at frequencies between 60 and 95 Hz . Amplitudes of pain-induced gamma oscillations increase with both objective stimulus intensity and subjective pain intensity . However , around pain threshold , differences in the subjective perception of objectively similar stimuli were related to differences in amplitudes of induced gamma oscillations . These results show that pain-induced gamma oscillations in S1 are particularly related to the subjective perception of pain . Here , latencies of pain-induced gamma oscillations between 100 ms and 300 ms indicate that these responses are mediated by A-delta-fibers relating to first pain sensation . Later sensations of warmth or second pain mediated by slowly conducting C-fibers occur at latencies of about 1 , 000 ms [11] , and are unlikely to relate to cortical responses at latencies around 200 ms . Correspondingly , in the present study , “percept” and “no percept” refer to the presence and absence of A-delta-fiber–mediated first pain sensation . Moreover , our findings provide evidence for first pain–related gamma oscillations in S1 , but do not preclude gamma oscillations from outside the somatosensory cortices , which were beyond the scope of our analysis . Induced gamma oscillations have been demonstrated in tasks that require activation and further processing of object representations [12 , 13] . As common features across sensory modalities , these induced gamma oscillations share a similar timing ( around 200 ms to 400 ms ) and a focal localization [12 , 14–16] as compared to low-frequency components . A few previous studies investigated pain-related changes in gamma oscillations [17–19] . However , these studies did not apply selective nociceptive stimuli and could not provide consistent results on location , timing , and functional characteristics of possible pain-induced gamma responses . Other studies investigating the relationship between S1 and S2 responses and pain intensity [20–25] revealed a positive correlation between response amplitudes and pain intensity . However , these studies did not investigate induced but evoked responses or blood oxygen level-dependent ( BOLD ) effects and , thus , cannot be directly compared to the present results . Painful stimuli possess utmost behavioral relevance that invariably interrupts ongoing processes , demand full access to system resources , and thereby lead to preferred processing [8 , 9] . Our study demonstrates that induced gamma oscillations are particularly related to the subjectively weighted percept of noxious stimuli . Thus , our results provide an important link between gamma oscillations and a neural filtering mechanism selecting behaviorally relevant information for action . Indeed , neural oscillations as regular excitability changes of neuronal populations have the capability of gating information flow and adding relevance to spike trains [4 , 26–28] . Thus , the gamma oscillations observed in the present study may modulate the cerebral processing of painful stimuli and the perceptual quality of the stimulus rather than representing the neural substrate of perception per se . Pain-induced gamma oscillations may thereby participate in activating the sensory representation of a painful stimulus and its selection for further processing . Preferred processing of these stimuli may facilitate behavioral responses aimed at preserving the integrity of the individual . Beyond , our findings suggest that pain-induced gamma oscillations in S1 are related to a complex cerebral network subserving conscious perception of sensory events . This network includes sensory areas as well as higher order frontal and parietal association cortices [1 , 29–34] . Within this network , perception appears to depend on the complex relationship between ongoing neuronal activity [30 , 35] , phase-locked evoked responses [30 , 34] , and non-phase-locked–induced [1 , 32] responses . Taken together , the present findings show that noxious stimuli induce gamma oscillations in S1 that are particularly related to the subjective experience of pain . These observations are compatible with the hypothesis that induced gamma oscillations are related to the internal representation of behaviorally relevant stimuli which should receive preferred processing . Our findings may , thus , contribute to our understanding of the neural mechanisms of perception and , in particular , to the understanding of the highly subjective experience of pain in health and disease .
Twelve healthy male participants ( mean age: 33 y , range 22–41 y ) participated in the experiment . All participants gave informed consent , and the study was performed according to the Declaration of Helsinki with the local ethics committee's approval . Subjects were comfortably seated in a reclining chair . Forty noxious cutaneous laser stimuli were delivered to the dorsum of the right hand , and subjects were instructed to passively perceive the stimuli with closed eyes [36] . Stimuli were cutaneous laser stimuli that selectively activate nociceptive afferents without concomitant activation of tactile afferents . The laser device was a Tm:YAG-laser ( Carl Baasel Lasertechnik , Starnberg , Germany ) with a wavelength of 2 , 000 nm , a pulse duration of 1 ms , and a spot diameter of 6 mm . An optical fiber transmitted the laser beam into the magnetically shielded recording room . Stimulation site was slightly changed after each stimulus . Interstimulus intervals were randomly varied between 10 and 14 s . Applied stimulus intensity was 600 mJ , which evoked moderately painful sensations . Neural activity was recorded with a Neuromag-122 whole-head neuromagnetometer [37] with passbands of 0 . 03–170 Hz and digitized with 514 Hz . The exact position of the head with respect to the sensor array was determined by measuring magnetic signals from four coils placed on the scalp . High-resolution T1-weighted magnetic resonance images ( MRI ) were obtained for each subject . Anatomical landmarks ( nasion and preauricular points ) were localized in each individual and used for the alignment of the MRI- and magnetoencephalography ( MEG ) coordinate systems . Thirteen healthy , right-handed men participated ( mean age: 28 y , range 25–33 y ) in experiment 2 . In this experiment , stimulus strength varied randomly with possible values of 150 , 300 , 450 , or 600 mJ . Forty stimuli for each intensity were applied . Three seconds after each laser stimulus , an auditory signal prompted the participants to rate the intensity of the initial “pinprick”-like first pain on a rating scale from 0–100 . Zero was defined as “no pain , ” and 100 was defined as the “worst imaginable pain . ” Because the rating was explicitly focused on first pain , a rating of zero did not preclude later sensations of warmth or second pain . The sample rate was 483 Hz , and signals were band-pass filtered between 0 . 03 and 160 Hz [24] . The other parameters were the same as in experiment 1 . Recorded signals were high-pass filtered ( 1 Hz ) and visually inspected for artifacts . Contaminated epochs were excluded , leaving a minimum of 36 trials per participants and stimulus intensity . First , somatosensory cortices were localized by analyzing the well-known pain-evoked responses from these areas [10] . Somatosensory cortices were selected for the analysis because previous studies showed that induced gamma oscillations mainly occur in early sensory cortices [2 , 7 , 12 , 14 , 16] . To this end , covariance matrices across all sensors were calculated for a prestimulus baseline interval ( −400 to 0 ms ) and a poststimulus interval ( 0 to 400 ms ) including strongest pain-evoked activity as evident in global field power . From these covariance matrices , neural activity during both intervals was localized by using a spatial filtering algorithm [38–40] . The spatial filter was used with a realistic head model to estimate power in the whole brain , resulting in individual tomographic power maps with voxel sizes of 6 × 6 × 6 mm . For each voxel , the ratio of poststimulus power to prestimulus power was computed resulting in individual functional tomographic power maps that show cortical areas with a strong increase of neural activity following noxious stimuli . These functional maps were individually normalized to one , spatially normalized using SPM2 ( Wellcome Department of Cognitive Neurology , Institute of Neurology , London , United Kingdom: http://www . fil . ion . ucl . ac . uk/spm ) , thresholded at 0 . 8 of the individual maximum , and averaged across participants . This procedure yields group-mean tomographic maps of pain-evoked power increases with a dimensionless maximum value of approximately 0 . 4 . An arbitrary threshold was used for visualization because only the local maxima , but not the extent of activations , were used for subsequent analysis . The Analysis of Functional Neuroimages/AFNI Surface Mapper ( AFNI/SUMA ) programs were used for surface rendering ( National Institute of Mental Health , Bethesda , Maryland , United States: http://afni . nimh . nih . gov/afni ) . As in previous MEG studies [10] , strongest pain-evoked activity was seen in contralateral S1 and bilateral S2 . Note that this analysis aimed at localizing somatosensory cortices and does not preclude activations of additional areas , e . g . , insular or cingulate cortex , which may yield lower signal-to-noise ratios due to their deep location and/or radial orientation barely detected by MEG . Second , individual locations of evoked S1 and S2 responses were optimized beyond the 6-mm grid of the first analysis step . To this end , a multi-dimensional , constrained nonlinear minimization ( Nelder-Mead , modified fminsearch function in Matlab , Mathworks: http://www . mathworks . com ) was employed . The S1 and S2 maxima from the individual tomographic power maps were used as starting points for the optimization . The position was allowed to change by 1 cm in each direction , whereas orientation was constrained to be tangential to the center of the head . For each position , the ratio of poststimulus activity ( 0 to 400 ms ) to prestimulus activity ( −400 to 0 ms ) was calculated , and the position with the maximum stimulus-evoked power increase was chosen for further analysis . The same optimization procedure was applied to induced gamma power in S1 , revealing that optimized locations of evoked S1 responses and induced gamma oscillations in S1 did not differ ( x-coordinates , p = 0 . 15; y-coordinates , p = 0 . 10; z-coordinates , p = 0 . 84; two-tailed Wilcoxon signed-rank tests ) . Third , for optimized locations of S1 and S2 , time courses of activity were computed for all single trials , using the adaptive spatial filter [38 , 39] . Note that time courses of all activations were analyzed in source space . These time courses were subjected to a time-frequency analysis based on multi-tapers [41] using the fieldtrip toolbox ( F . C . Donders Centre for Cognitive Neuroimaging: http://www . ru . nl/fcdonders/fieldtrip ) . A multi-taper–based analysis was chosen since this approach provides a robust and optimal way to smooth spectra in the frequency domain , and thereby enhances higher frequency oscillatory components with large-frequency jitter-like induced gamma oscillations . The analysis yields TFRs showing power as a function of time and frequency . TFRs were computed from 30 to 100 Hz in 400-ms–long windows with a spacing of 20 ms between windows . A 400-ms time window was chosen to allow a multi-taper frequency smoothing of ±5 Hz . For each frequency , relative change to a 1 , 000-ms baseline was computed . Power was coded as z-scores calculated from the 1 , 000-ms baseline . Significance of differences between poststimulus and prestimulus activity in TFRs was determined by applying permutation statistics . To this end , the 12 prestimulus baseline ( −1 , 000 to 0 ms ) and the 12 poststimulus ( 0 to 1 , 000 ms ) parts of the TFRs of the 12 different subjects were randomly permuted 5 , 000 times . Each time , the maximum difference was computed across time and frequency . The 95th percentile of all 5 , 000 maximum differences was taken as threshold for the TFRs . This maximum statistics takes multiple comparisons into account [42] . Fourth , in order to distinguish between phase-locked ( evoked ) and non–phase-locked ( induced ) neural responses , phase locking of stimulus-related neural activity was determined . For each cortical area , single trial time courses were bandpass filtered ( forward and reverse with a fourth-order Butterworth filter ) in the gamma frequency band ( 60–95 Hz ) defined from TFRs . The Hilbert transformation yielded instantaneous phase and amplitude estimates for each single trial with an optimum temporal resolution . Phase-locking value ( plv , bounded between zero and one ) was computed as the absolute value of the mean of complex phase Φ across N trials . Evoked components show a consistent phase relationship to stimulus that is evident in a high plv . Amplitude and phase dynamics were calculated with reference to a 1 , 000-ms prestimulus baseline . To establish a confidence level for stimulus-induced phase locking , the time course for each region of interest was randomly permuted 5 , 000 times and then subjected to the same phase-locking analysis ( i . e . , filtering , Hilbert transformation , averaging , and baseline correction ) . The maximum value was used as the confidence level . | Pain is a highly subjective sensation of inherent behavioral importance and is therefore expected to receive enhanced processing in relevant brain regions . We show that painful stimuli induce high-frequency oscillations in the electrical activity of the human primary somatosensory cortex . Amplitudes of these pain-induced gamma oscillations were more closely related to the subjective perception of pain than to the objective stimulus attributes . They correlated with participants' ratings of pain and were stronger for laser stimuli that caused pain , compared with the same stimuli when no pain was perceived . These findings indicate that gamma oscillations may represent an important mechanism for processing behaviorally relevant sensory information . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"neuroscience",
"homo",
"(human)"
] | 2007 | Gamma Oscillations in Human Primary Somatosensory Cortex Reflect Pain Perception |
The phenotype of any organism on earth is , in large part , the consequence of interplay between numerous gene products encoded in the genome , and such interplay between gene products affects the evolutionary fate of the genome itself through the resulting phenotype . In this regard , contemporary genomes can be used as molecular records that reveal associations of various genes working in their natural lifestyles . By analyzing thousands of orthologs across ∼600 bacterial species , we constructed a map of gene-gene co-occurrence across much of the sequenced biome . If genes preferentially co-occur in the same organisms , they were called herein correlogs; in the opposite case , called anti-correlogs . To quantify correlogy and anti-correlogy , we alleviated the contribution of indirect correlations between genes by adapting ideas developed for reverse engineering of transcriptional regulatory networks . Resultant correlogous associations are highly enriched for physically interacting proteins and for co-expressed transcripts , clearly differentiating a subgroup of functionally-obligatory protein interactions from conditional or transient interactions . Other biochemical and phylogenetic properties were also found to be reflected in correlogous and anti-correlogous relationships . Additionally , our study elucidates the global organization of the gene association map , in which various modules of correlogous genes are strikingly interconnected by anti-correlogous crosstalk between the modules . We then demonstrate the effectiveness of such associations along different domains of life and environmental microbial communities . These phylogenetic profiling approaches infer functional coupling of genes regardless of mechanistic details , and may be useful to guide exogenous gene import in synthetic biology .
An important challenge in biology is to understand the relationships between an organism's genotype and its phenotype , involving dissection of myriad interdependencies among various cellular components , such as genes , proteins , and small molecules [1] , [2] . Recent efforts to map protein-protein interactions [3] , [4] , together with efforts to reconstruct metabolic and regulatory networks at the genome scale [5] , [6] , [7] , offer promising opportunities to investigate emergent biological phenomena of interacting biomolecules inside cells . Complex interdependencies among the products of individual genes do not necessarily imply molecular interactions by direct physical contacts but also include more macroscopic associations exhibited at e . g . biochemical pathway levels , as has been claimed in epistasis research [8] . These genetic interdependencies , regardless of their underlying mechanisms , may leave a trace on the composition of genomes through evolutionary processes . One important set of analyses that have been used successfully over the past decade is based on phylogenetic profiles as introduced in [9] . Here , we analyze a type of phylogenetic profile – the patterns of the presence or absence of orthologs across many organisms – to find genes with favored co-occurrence in the same genomes ( called herein correlogs ) or disfavored co-occurrence ( called anti-correlogs ) suggesting their putative functional coupling . Such analysis of correlogy and anti-correlogy can help uncover global gene associations conserved at the biome level , beyond those specific to any particular organism . This information is distinct from genetic interactions inferred from , for example , double-mutant data [10] , [11] , which relate to gene relationships within the specific organisms employed in the experiments . In particular , knowledge of the association of genes not coexisting inside considered specific organisms can be applied for heterologous gene expression in synthetic biology [12] , and interest in anti-correlogy itself inevitably tends to target heterologous genes . It should also be noted that fitness of an organism in its natural habitat does not necessarily coincide with fitness in a laboratory [13] , and ortholog profiles of organisms may thus encompass genetic relationships in environmental and ecological contexts not readily captured by laboratory experiments . The importance of co-occurring orthologs as a means to gain insight into gene relationships is well appreciated in many previous studies [9] , [14] , [15] , [16] , [17] , [18] , [19]; however , these studies have largely focused on improving the identification of molecular-level interactions rather than on systematic analysis of the global organization of gene associations , as is the focus herein . In addition , we have included the removals of indirect gene-gene correlations in analyzing co-occurrence patterns to reduce false positive associations . We here take a comprehensive and systematic approach to study from a global perspective the biome-wide associations of genes as manifest through patterns of correlogy and anti-correlogy across sequenced bacteria .
We surveyed the presence or absence of gene orthologs across 588 different bacterial species ( Table S1 ) on the basis of orthology data available from the Kyoto Encyclopedia of Genes and Genomes ( KEGG ) [20] . Beyond simply measuring co-occurrence of these genes , we evaluated degrees of correlogy or anti-correlogy between the genes within the context of direct associations in biological activities , using methods to help avoid vertical co-inheritance effects , transitivity effects , and other spurious correlations . In particular , we attempted to reduce the contribution of indirect ( anti- ) correlogous relationships that result from transitivity effects: if genes i and j are each correlated with third common genes in terms of the presence or absence across species , genes i and j can also appear to be correlated with each other even in the absence of direct association between them . Therefore , simple correlations calculated from the co-occurrence pattern can suffer from these indirect correlations . Such removal of indirect correlations in this manner has not generally been taken into account in previous studies on ortholog profiles [9] , [14] , [15] , [16] , [17] , [18] , [19] , or also with similar types of correlation calculations in other fields to infer disease comorbidity networks [21] , [22] , [23] or social networks [24] . Nevertheless , filtering out these transitivity effects has been of critical importance in e . g . reverse engineering of transcriptional regulatory networks [25] , [26] , and we employed such ideas to reduce transitivity effects when quantifying correlogy and anti-correlogy . As a result , for every pair of a total of 2085 genes , we assigned wij of which magnitude increase away from zero measures the magnitude of correlogy ( wij>0 ) or anti-correlogy ( wij<0 ) between genes i and j in a pair ( see Methods ) . It is worthwhile to address how much correlogous relationships overlap with molecular-level interactions . To test this idea , we compared the distribution of wij for physically-binding proteins in Escherichia coli with that for arbitrary pairs of the proteins [27] , and found that highly correlogous proteins are much likely to physically interact ( Figure 1A ) . Indeed , the average of wij for directly-binding proteins was 6 . 8 times larger than that for the second nearest proteins in the protein interaction network ( P = 7 . 1×10−55 ) , and 10 . 5 times larger than that for all the protein pairs ( P = 8 . 8×10−57; Figure 1B and Methods ) . Instead of wij , if we use the simple co-occurrence measure rij that precedes wij before the alleviation of transitivity effects ( Methods ) , the overall distribution of rij is heavily biased toward positive rij 's ( Figure 1C ) , and the enrichment of high rij's in physically-binding protein pairs is relatively weak ( P = 4 . 0×10−8 against the second nearest proteins ) although still significant . Moreover , the network-topological signature appearing for wij in Figure 1B becomes very distorted for rij with a hump at the tenth nearest proteins in Figure 1D . Note that a hump at the tenth nearest proteins represents average rij larger than averages at nearer proteins , possibly contributed to by indirect correlations from transitivity but unlikely to be biologically meaningful . Thus , the elimination of effects caused by transitivity is critical to identifying biologically meaningful correlogous and anti-correlogous relationships between genes . Another simple co-occurrence measure we tried , the mutual information Iij of genes i and j ( Methods ) , reflects better the physical interactions than rij but still less than wij ( P = 3 . 4×10−28 against the second nearest proteins; Figures 1E and 1F ) . Again , Iij leaves a small hump at the tenth nearest proteins in Figure 1F , and by its definition does not directly provide positive or negative signs of gene relationships themselves which are important in our study . The data in Figure 1A suggest a natural boundary for separating a regime enriched with functionally-obligatory protein interactions ( wij>0 . 045 ) from that with conditional or transient interactions [28] , [29] . Specifically , the probability density of wij for physically-binding proteins deviates strongly from that for arbitrary pairs of proteins at wij>0 . 045 , but almost overlaps with that for the arbitrary pairs at the rest wij ( except for the absence of strong anti-correlogy ) . In the former regime of wij , there thus might be present functionally-obligatory relationships between physically interacting proteins to constrain them with large correlogous associations , as indicated by the rich presence of operonic genes whose transcriptions are precisely co-regulated in time ( Text S1 and Figure S1 ) . Even if we exclude these operonic gene pairs , the transcripts of the gene pairs with wij>0 . 045 still tend to be more co-expressed than the others ( P = 0 . 03 ) , implying more obligatory interactions between them ( Text S1 and Figure S1 ) . Taken together , integration of protein-protein interaction data with the heterogeneous data of correlogous relationships provides a powerful means for identifying potentially functionally-obligatory interactions conserved across evolution from mere binding events . It is of course expected that other forms of molecular interactions than physically-binding interactions would also contribute to correlogy and anti-correlogy . Table 1 presents some cases of the highest |wij|'s; rfbF encodes the enzyme to catalyze the reaction , CTP+α-d-glucose 1-phosphate→diphosphate+CDP-glucose , and the produced CDP-glucose is subsequently converted by the enzyme from the correlogous gene , rfbG , into CDP-4-dehydro-6-deoxy-d-glucose . The correlogous relationship between rfbF and rfbG thus appears to be a consequence of the need for the second reaction to proceed when the first occurs in order to achieve the relevant biological functions . On the other hand , two NAD+ synthases , one utilizing ammonia as an amide donor to produce NAD+ and the other utilizing glutamine as an amide donor instead , are highly anti-correlogous to each other , and the anti-correlogy reflects that both processes operating simultaneously in the same organism have been consistently selected against . This fact might be related to the distinct modes of regulating cellular nitrogen inside bacteria , as pleiotropic nitrogen utilization mutations can be found at NAD+ synthases [30] . In general , genes associated with similar functions were enriched with correlogous and anti-correlogous relationships ( Figure S2 and Text S1 ) . Indeed , the average of wij>0 for isozymes based on the same Enzyme Commission numbers was 5 . 8 times larger than that for arbitrary pairs of enzymes ( Z = 83 . 69 ) , and the average of wij<0 for the isozymes was 2 . 9 times larger than that for arbitrary enzyme pairs ( Z = 24 . 11 ) . Thus , isozymes exhibit high levels of correlogy and anti-correlogy compared to arbitrary enzyme pairs , although the effect is more pronounced for the correlogy side . To evaluate the overall correlogous couplings around individual genes , we defined for a given gene i , , where the summation was taken over all other gene j's satisfying wij>0 . Likewise , we can define . Sip and Sin quantify how tightly gene i is associated to other genes through correlogy and anti-correlogy , respectively . Having both the largest Sip and Sin was gene rpmJ , which encodes ribosomal protein L36 . The next largest were Sip and Sin of DNA-modifying genes , while the smallest ones were owned by flagella-related genes ( Table S2 ) . In general , Sip and Sin of each gene i are very positively correlated ( Text S1 and Figure S3 ) . How broadly genes are phylogenetically distributed would affect or be affected by the strength of interactions with other genes . Therefore , the degree of phylogenetic spread of any given gene is expected to be correlated with its Sip and Sin . Rather surprisingly , our analysis suggests that species-level spread of genes does not exhibit such correlations ( Figure 2A ) , but spread at a higher taxonomic level – at the phylum level – reveals clear correlations with Sip and Sin; as genes inhabit diverse phyla , their Sip and Sin tend to increase continuously until saturated at a plateau ( Z = 79 . 94; Figure 2B and Methods ) . Since different phyla ( phylogenetically distant ) may represent disparate cell types more clearly than different species ( phylogenetically close ) , our result suggests that how many such disparate cellular conditions genes inhabit at the phylum level could substantially evolve or be influenced by the strength of the genetic interdependencies around the genes . To systematically view the essential structure of correlogous and anti-correlogous relationships from a global architectural perspective , we constructed a maximum relatedness subnetwork ( MRS ) [24] , in which each gene i points to two other genes j and j′ by different categories of edges that represent the most correlogous ( maxj wij>0 ) and anti-correlogous ( minj′ wij′<0 ) genes to gene i , respectively ( Figures 3A–3C and Table S3 ) . By definition , genes in MRS are not necessarily reciprocally linked . As such , one can easily recognize that following a series of genes in either of correlogy or anti-correlogy direction of links , the magnitude of wij of each link increases until encountering reciprocal links . MRS provides the ‘backbone’ structure of a considered network , and has been demonstrated as particularly useful for detecting modular structures of complex networks [24] . One example is for two isocitrate dehydrogenases , IDH1 and IDH3 , which encode similar enzymes depending on NADP+ and NAD+ , respectively , and are anti-correlogously associated in the MRS . Consistent with our observations , their distinct phylogenetic profiles have been discussed in previous studies [31] . Also in the MRS , IDH1 is correlogously associated with gltA that encodes citrate synthase , and the enzyme activities from IDH1 and gltA are indeed known as elaborately coordinated in a cell for efficient growth on acetate [32] . For another example , β-glucosidase ( bglB ) , which breaks down cellobiose into β-d-glucose and can be industrially useful for lignocellulose conversion for biofuel production [33] , is correlogously associated in the MRS with l-rhamnose mutarotase ( rhaM ) . We expect that this mutarotase may convert the product of β-glucosidase into α-d-glucose if the host cell prefers the α form to the β form , and thus expressing both the two genes , bglB and rhaM , may introduce a synergetic effect in cellobiose metabolic engineering . Furthermore , Figure 3C illustrates that , along correlogy direction , the MRS arranges sequentially xylose transport protein ( xylF ) , xylose isomerase ( xylA ) , and xylulokinase ( xylB ) , the same order as their arrangement in the xylose metabolism pathway ( Figure 3D ) . It is interesting to note that the MRS brings relatively less characterized genes such as xylF to attention by informing of their strong correlogous relationships with other genes . As such , Figure 3C does specify the identities of genes important for xylose metabolic engineering [34] that may benefit from simultaneous heterologous expression of xylF , xylA , and xylB . Also , by their link directionality the MRS provides the information that xylA and xylB are more associated than xylA and xylF . We found that all genes in the MRS are decomposed into 483 different small subgroups ( Table S3 ) , of which each includes correlogously associated genes yet not linked correlogously to any genes in the other subgroups . Nevertheless , the vast majority ( 99 . 5% ) of anti-correlogy links bridge the gaps between different modules in the MRS , binding all of them as a single giant component ( Figures 3B and 3C ) . Thus , the MRS results in clear modular structure among correlogs , with these modules interrelated through anti-correlogous relationships . Different genes in each correlog group are highly likely to perform biological tasks in a common functional category ( Z = 43 . 16 for KEGG categories; see Methods ) . For every pair of possible functional categories , we also analyzed how likely genes of the functional categories in a pair would belong together to the same correlog groups ( Methods ) . As expected , genes of the same KEGG functional categories tend to belong to the same correlog groups , but deviations from this behavior are also informative ( Figure 4A and Table S4 ) . For example , genes of functional category Folding , sorting and degradation and those of another category Metabolism of other amino acids significantly tend to be affiliated to the same correlog groups ( P<10−4 ) . The former category involves a number of molecular chaperones and RNA helicases , while the latter involves enzymes to synthesize glutathione and d-glutamate . Since glutathione serves as the major endogenous antioxidant and d-glutamate decorates bacterial cell walls , our results support the previous observation that the corresponding enzymes are likely to be actively in concert with chaperones and RNA helicases when subject to oxidative or cell-wall stress [35] , [36] . Likewise , genes of functional category Xenobiotics biodegradation and metabolism and those of category Folding , sorting and degradation are highly likely to be together in the same correlog groups ( Figure 4A and Table S4 ) , which can be understood in the similar context of cellular stress response induced by xenobiotic compounds such as benzoate and bisphenol [37] . We found that the number of genes in each correlog group approximately follows an exponential distribution ( Figure 4B ) . We expect that each correlog group may serve as a toolbox for importing exogenous genes and functions into a cell , as xylF , xylA , and xylB above form a single correlog group themselves ( Figures 3C and 3D ) . Furthermore , the prevalence of anti-correlogy links between correlog groups as shown in Figures 3B and 3C extends the concept of anti-correlogy from single genes to correlog groups and allows for evaluating the anti-correlogous associations between , rather than within , the correlog groups ( Table S5 and Text S1 ) . For example , a correlog group containing aldehyde-related dehydrogenases was anti-correlogously associated with another group containing perR , peroxide stress response regulator , and this anti-correlogy between the two groups might be involved in a problem that arises in controlling cellular redox state [38] . If a correlog group represents a repertoire of genes that tend to coexist in the same organisms , genes in individual organisms when mapped to MRS should be densely distributed around the correlog groups rather than distributed uniformly . This hypothesis can be straightforwardly checked by enumerating the number of different correlog groups that the genes are mapped to , and comparing this value with the number of different correlog groups harboring the same number of genes but mapped randomly . If the former is smaller than the latter , the genes can be regarded as more clustered around correlog groups than by chance . As would be expected , genes from each bacterial species show much more clustered behaviors than by chance ( −13 . 20≤Z≤−3 . 68 for all bacteria; Figures 4C–4E and Methods ) . To the MRS , we can also map orthologs in each species from the other superkingdoms , archaea and eukaryotes . Interestingly , although the MRS in this study is based on bacterial data , orthologs from archaea and eukaryotes show somewhat weakened but still significantly clustered behaviors around the correlog groups , reflecting a significant conservation of gene associations across different domains of life [−5 . 97≤Z≤−2 . 64 for archaea except for one species with Z = −1 . 32 ( Figure 4C ) and −6 . 47≤Z≤−2 . 23 for eukaryotes except for five species with −1 . 93≤Z≤−1 . 39 ( Figure 4D ) ; see Methods] . Among these observations , less significantly clustered behaviors come from the species having small numbers of genes mapped to the MRS , and this result might be due to insufficient mappings or minimally-required diversity of genes for any biological systems . On the other hand , considering that correlogs tend to coexist in individual organisms , it might be challenging to examine the above clustering issues for environmental microbial communities [39] , as each of the environmental samples typically contains a number of different species . Again , we mapped to the MRS the genes from twelve diverse environmental sources such as human and mouse guts , deep-sea whale fall carcasses , and uranium contaminated ground water [40] , and found still significant clustering of those genes around the correlog groups ( −9 . 26≤Z≤−2 . 10 except for one sample with less significant Z = −0 . 95; Figure 4E and Methods ) . Accordingly , this result encourages us to even conjecture the identities of undetected but existing genes in environmental samples based on those of detected genes by applying the knowledge of correlogous gene associations . It would also be interesting to identify correlog groups harboring cosmopolitan genes [41] in a given environment , as these groups can represent together environment-specific genetic contents rather than species-specific ones .
Recent advances in understanding genome-scale intracellular networks have been enabled by the availability of high-throughput experimental techniques that permit network data to be collected on a scale far larger than previously possible [1] , [2] . These networks help capture and quantify the functional interplay amongst genes that drives essentially all cellular phenotypes in the environment . Correlogy and anti-correlogy can be regarded as the natural outcome of such ubiquitous interactions between genes , molded through a long evolutionary process that tests beneficial effects of numerous gene combinations for a cell . The resulting ‘sociology’ of genes provides rich implications in rapidly growing fields such as function prediction of uncharacterized genes [27] , gap-filling in genome-scale biochemical model reconstruction [42] , and synthetic biology [12] . In synthetic biology , information on both correlogs and anti-correlogs may be useful to screen additional gene candidates to be expressed or silenced contingent on primary genes of interest , and such considerations should inform attempts to construct optimal recombinant strains . Intriguingly , the information on correlogy and anti-correlogy can be uncovered even when the underlying mechanisms of these gene associations are unknown . Finally , future extension of our work towards archaea and eukaryotes will allow comparative analysis of correlogy and anti-correlogy for different domains of life , offering a fascinating opportunity of understanding evolution of genetic associations .
We downloaded orthology data from the KEGG database [20] , and surveyed the presence or absence of each ortholog across different bacterial species ( see Table S1 for the full list of the bacterial species considered here ) . In order to capture functional interactions of genes reflected in their co-occurrence patterns across species , we take into account the genes ( i . e . , orthologs ) not too lowly nor too highly prevalent across species; in the case of too lowly ( highly ) prevalent genes , there do not exist so many species with ( without ) the genes , making it hard to judge whether these few co-presences ( co-absences ) of the genes actually come from their functional interactions . In other words , without filtering , spurious correlations from non-functional origins may emerge , simply by vertical co-inheritance of genes or by chance . Specifically , if Ei denotes the number of species containing gene i and N denotes the total number of species , one can define Xi = min ( Ei , N−Ei ) for each gene i . The probability density of Xi approximately follows the power-law decay as long as Xi≥Xth = 80 , and we chose the genes with Xi≥Xth to prevent spurious correlations that could occur at low Xi deviating from the power-law trend observed at large Xi . To evaluate direct gene associations while alleviating transitivity effects in charge of indirect correlations between genes , we applied the partial correlation method employed in graphical Gaussian models [26] , of which superiority over many other methods was demonstrated in reverse engineering of transcriptional regulatory networks [25] . To implement this method , we start with calculating the Pearson correlation rij for binary variables of presence and absence of genes i and j:where Cij is the number of species containing both genes i and j . Next , to reduce indirect correlations between genes i and j , we calculate the partial correlation wij using rij:where pij is the ( i , j ) th component of an inverse matrix of rij . However , in this study , the number of genes ( = 2085 ) is much larger than the number of species ( = 588 ) , yielding an ill-conditioned problem for matrix rij . To overcome this problem , we applied the shrinkage estimation derived by Schäfer and Strimmer [26] . Specifically , Schäfer and Strimmer obtained a regularized estimator of rij combining analytic determination of shrinkage intensity from the Ledoit-Wolf theorem [43] . The following is the resultant estimator r*ij that simply substitutes for rij in the above calculation of wij:where δij is the Kronecker delta symbol and λ is given by Here xki = 1 if gene i is present in species k , otherwise , xki = 0 , <···>k denotes the average over species k's , and σi is the standard deviation of xki over species k's . As a result , the obtained wij was used to quantify correlogy ( wij>0 ) or anti-correlogy ( wij<0 ) between genes i and j . For comparative analysis , the mutual information Iij of genes i and j between {xki} and {xkj} across species k's [14] was also calculated . We calculated the average of wij ( wppi ) from the pairs of physically-binding proteins in E . coli [27] , and obtained its P value by generating the distribution of average wij ( wnull ) from the same number of , but arbitrarily-mated pairs of the proteins as distant as the given shortest path length in the protein interaction network . The central limit theorem ensured that this null distribution converged well to the Gaussian distribution , providing the P value for how frequently wnull exceeds wppi . Similar analyses were also performed for rij and Iij . In order to characterize how tightly each gene i is correlogously associated to other genes , we calculated , where the summation was taken over all other gene j's satisfying wij>0 . In a similar way , we calculated for anti-correlogous couplings around gene i . Let nphyla be the number of different phyla where genes are present . For genes with nphyla<7 ( Figure 2B ) , we obtained the slope of Sip against nphyla by linear regression , and normalized it by multiplying . From surrogate data with randomly-permuted gene presences across species , we also generated an ensemble of such normalized slopes for nphyla<7 , and calculated the Z score of the actual value . For any given weighted network , one can simplify its structure by constructing the MRS composed only of highly weighted edges [24] . Specifically , in the MRS of this study , each gene i points to only two genes j and j′ by different categories of edges that represent the most correlogous ( maxj wij>0 ) and anti-correlogous ( minj′ wij′<0 ) genes to gene i , respectively . We found that all genes in the MRS here are decomposed into 483 different small subgroups , of which each includes correlogously associated genes yet not linked correlogously to any genes in the other subgroups . These subgroups in the MRS were termed correlog groups . For given correlog group g in the MRS and given functional category c of genes , we can calculate = Ñcg/Ñg , where Ñcg is the number of genes affiliated to both correlog group g and functional category c , and Ñg is the total number of genes affiliated at least to one functional category in correlog group g . Therefore , fcg represents the uniformity of gene functions in a correlog group . Majority ( 57 . 1% ) of correlog groups with Ñg>1 were shown to have at least one functional category c satisfying fcg = 1 in each g . To calculate the corresponding Z score , we generated an ensemble of correlog groups with Ñg>1 by randomly exchanging genes of the same number of the affiliated functional categories . For a given pair of functional categories c1 and c2 , we can also define their overlapping ratio ( Figure 4A and Table S4 ) as:where i and j are indices of genes , ai , c is 1 if gene i belongs to functional category c , otherwise 0 , di , g is 1 if gene i belongs to correlog group g , otherwise 0 , and H ( x ) is 1 if x>0 , otherwise 0 . Yc1 , c2 quantifies how likely genes in c1 and c2 belong to the same correlog groups ( 0≤Yc1 , c2≤1 ) . The corresponding P value was obtained by generating the null distribution in the same way as in the case of fcg above . For each species or environmental sample , we counted the number ( n ) of correlog groups harboring the genes mapped to the MRS . We also obtained the mean ( η ) and the standard deviation ( σ ) of such numbers of correlog groups when the same number of genes are randomly mapped to the MRS ( Figure 4C–4E ) :where N is the number of genes mapped to the MRS , g is the index of each correlog group , Ng is the total number of genes in correlog group g , and is the total number of correlog groups in the MRS . Accordingly , we calculate the Z score of n [Z = ( n−η ) /σ] . | Genes in organisms have a number of interactions with one another in their biological contexts . For example , proteins produced from one gene may interact with other proteins produced from another gene to perform together a particular biological task , and such pairs of cooperative genes may often reside together in the same organisms . We analyzed thousands of genes across ∼600 bacterial species , and found genes with favored co-occurrence in the same organisms ( termed correlogs ) or disfavored co-occurrence ( termed anti-correlogs ) . These co-occurrence patterns are significantly reflective of actual biochemical interplays between genes , and distinct cliques of correlogous genes are seamlessly interrelated through anti-correlogous links between the cliques . The ‘sociology’ of genes inferred by this approach provides useful information on how to engineer a cell , such as for production of a desired byproduct . For example , an important gene in cellobiose digestion for biofuel production , bglB , is suggested to function better in a cell factory when co-activated with another gene rhaM , the correlogous partner we found in our analysis . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"physics",
"biochemistry",
"protein",
"interactions",
"theoretical",
"biology",
"interdisciplinary",
"physics",
"molecular",
"biology",
"biophysics",
"theory",
"synthetic",
"biology",
"biology",
"computational",
"biology",
"metabolic",
"networks",
"biophysics",
"proteomics"
] | 2011 | Genetic Co-Occurrence Network across Sequenced Microbes |
The world is rapidly becoming urban with the global population living in cities projected to double by 2050 . This increase in urbanization poses new challenges for the spread and control of communicable diseases such as malaria . In particular , urban environments create highly heterogeneous socio-economic and environmental conditions that can affect the transmission of vector-borne diseases dependent on human water storage and waste water management . Interestingly India , as opposed to Africa , harbors a mosquito vector , Anopheles stephensi , which thrives in the man-made environments of cities and acts as the vector for both Plasmodium vivax and Plasmodium falciparum , making the malaria problem a truly urban phenomenon . Here we address the role and determinants of within-city spatial heterogeneity in the incidence patterns of vivax malaria , and then draw comparisons with results for falciparum malaria . Statistical analyses and a phenomenological transmission model are applied to an extensive spatio-temporal dataset on cases of Plasmodium vivax in the city of Ahmedabad ( Gujarat , India ) that spans 12 years monthly at the level of wards . A spatial pattern in malaria incidence is described that is largely stationary in time for this parasite . Malaria risk is then shown to be associated with socioeconomic indicators and environmental parameters , temperature and humidity . In a more dynamical perspective , an Inhomogeneous Markov Chain Model is used to predict vivax malaria risk . Models that account for climate factors , socioeconomic level and population size show the highest predictive skill . A comparison to the transmission dynamics of falciparum malaria reinforces the conclusion that the spatio-temporal patterns of risk are strongly driven by extrinsic factors . Climate forcing and socio-economic heterogeneity act synergistically at local scales on the population dynamics of urban malaria in this city . The stationarity of malaria risk patterns provides a basis for more targeted intervention , such as vector control , based on transmission ‘hotspots’ . This is especially relevant for P . vivax , a more resilient parasite than P . falciparum , due to its ability to relapse and the operational shortcomings of delivering a “radical cure” .
Addressing health problems associated with urban growth will be one of the major challenges of the 21st century , especially for the developing world [1] . City life is associated with significant variation in socioeconomic and environmental conditions of potential relevance to vector-borne diseases [2–4] . In particular , the pronounced and on-going increase in urban population [5 , 6] , combined with climate change and economic disparities could act synergistically on the transmission dynamics of malaria [7–9] . Although , Indian cities harbor both malaria parasites , Plasmodium falciparum and Plasmodium vivax , there is an increasing appreciation of the latter as a threat to global health [10] , in particular in urban environments where Plasmodium vivax has become the most prevalent parasite . Plasmodium vivax has re-emerged in areas previously cleared of malaria , and has done so with higher mortality and morbidity than previously documented [11–13] . In Indian cities with seasonal transmission , the incidence of vivax malaria starts to rise earlier than that of falciparum malaria , and also earlier than transmission via the vector would allow . This earlier part of the vivax season is dominated and enabled by relapses . The parasite has the ability to delay the development of a fraction of the infectious load of sporozoites in the liver [14] , which results in the relapse of the disease after the primary infection is cleared from the bloodstream [15] . The later part of the season is shared with P . falciparum , reflecting largely vector-borne transmission following the monsoon rains [10] . A better understanding of the spatial heterogeneity of malaria vivax risk ( defined based on the wards’ incidence ) within cities remains an open area of research in the population dynamics of the disease , of relevance to both prediction and intervention . Historically , urbanization has led to economic and social transformations associated with profound improvements in sanitation and hygiene [16 , 17] . For malaria , the process of urbanization is generally thought to reduce transmission , primarily because urban environments are largely unsuitable for malaria vectors due to a lack of breeding sites and the pollution of potential larval habitats [18] . Other explanations for reduced malaria risk include better access to health care services and an increased ratio of humans to mosquitoes [2] . There is concern however that areas with rapid unplanned urbanization and poor sanitation may not experience this marked decrease in malaria transmission [5] . The most common hypotheses for the persistence of malaria in cities include spatial variation in: 1 ) environmental conditions ( relative humidity , temperature , precipitation ) , land use , and stored water , which create a favorable environment for Anopheles breeding in cities [7 , 19 , 20]; 2 ) socioeconomic factors ( income , human movement , population density and the failure of local malaria intervention among others ) , which hamper the effectiveness of case management and the promotion of intermittent antimalarials [8 , 21–24]; and 3 ) at more local scales , variation in mosquito behavior and ecology which can influence transmission intensity [22 , 23] . Despite increased interest in the role of spatial heterogeneities in the population dynamics of vector-borne diseases [25–28] , malaria models typically assume spatially homogeneous transmission and tend to aggregate temporal dynamics over space . In particular , they do not take into account how spatial variation in environmental , climatic and socio-economic conditions affect vector habitat , contact rates , host susceptibility and the effectiveness of control [27–29] . Importantly , these considerations are focused on Africa where endemic malaria remains a predominantly rural problem , because the main mosquito vectors are themselves rural , and in cities , largely peri-urban [2 , 3 , 5 , 7 , 8] . By contrast , the Indian subcontinent harbors a truly urban vector , Anopheles stephensi , particularly thrives in man-made environments , and breeds in various artificial containers within homes and in water collected in construction sites ( catch basins , seepage canals , wells ) [30] , whereas its sister ( sub ) species ( myorensis ) is associated with rural areas . The existence of this particular vector within cities poses a unique challenge to the elimination of malaria in India . Cities can act as a reservoir for the persistence of the disease beyond their administrative limits , and prevent elimination despite considerable gains in the fight against rural transmission . But even beyond the Indian subcontinent , malaria can no longer be considered only a rural problem , given the increasing proportion of the world population living in cities and transmission of the disease in urban environments sometimes surpassing that of rural ones [2 , 3 , 9] . Here , we describe the spatial pattern of urban Plasmodium vivax risk within a large city of India , and investigate its association with socio-economic and environmental heterogeneity . By combining statistical analyses and adapting a probabilistic modeling framework previously proposed for cholera [31] , we show that spatial heterogeneity in population size , environmental and economic factors , modulates malaria risk , and that the temporal effect of climate variables on malaria risk interacts with this spatial heterogeneity . These findings emphasize the importance of considering the interplay of climate forcing and socio-economic heterogeneity in the population dynamics of urban malaria in India . They also provide a basis for more targeted intervention , such as vector control , based on identifying transmission ‘hotspots’ . Comparison to findings for Plasmodium falciparum reinforce the evidence for a role of spatial heterogeneity in the transmission of urban malaria in this region .
We take advantage of a highly disaggregated dataset of monthly malaria cases collected by the Municipal Corporation of the city of Ahmedabad , the capital of state of Gujarat in Northwest India ( Fig 1A and 1B ) . In this largely semi-arid state where malaria is seasonally epidemic , Ahmedabad reports more than 1000 cases every year . The city presents ideal conditions to investigate malaria transmission dynamics in an urban environment , since it has experienced rapid development , unplanned urbanization and large population growth . It is also located on the banks of the Sabarmati River which creates environmental variation in this otherwise arid setting . The malaria data consist of monthly cases for the dominant parasite , Plasmodium vivax , over the last decade , confirmed through microscopy examination of blood slides from clinical ( febrile ) individuals self-presenting at the health facilities ( Fig 1B and 1C ) . The resulting time series span a total of 12 years ( from 2002 to 2014 ) . Because some administrative units ( known as wards ) were subdivided into smaller wards in 2007 , we aggregated the 64 units of the city into 59 wards , in order to maintain the same geographical subdivision through time . For comparison purposes , a similar data set for Plasmodium falciparum is also considered . Apart from decadal census data , annual population data were provided by the Ahmedabad Municipal Corporation to approximate the population of each ward . Socio-economic data were obtained from the District Census Handbook of the concerned district for the year 2001 from the Directorate of Census Operations , Gujarat . Monthly time series ( from 2002 to 2014 ) for mean temperature , mean rainfall and relative humidity at 8 am are those from the meteorological station of the city of Ahmedabad ( and were provided by the Indian Institute of Technology in Gandhinagar ) . In order to investigate the existence of a spatial pattern in malaria incidence within the city of Ahmedabad , we performed a series of statistical analyses to address whether malaria risk varied within the city and what factors explained this variation . First , we analyzed the spatial and temporal variation of malaria incidence and identified regions of high and low risk based on incidence . Second , we performed a series of statistical analyses on the role of socioeconomic and environmental factors in the spatial , and spatio-temporal patterns of malaria incidence . These analyses ranged from a simple t-test comparing socio-economic factors between the two regions of differential malaria risk , to time series models incorporating the autocorrelation in the data and the external drivers ( including climatic ones ) , to a full spatio-temporal general linear mixed model with random effects . Third , based on these results a probabilistic dynamical model was formulated for malaria transmission at the ward level , and predictions of this model were evaluated at the city level . To consider a measure of vivax malaria risk independent from interannual variation , we normalized malaria incidence for each ward in a given year by the total number of cases throughout the city . We then ranked these normalized values across wards to determine if high risk locations were consistently so over time . To examine the robustness of the patterns , we complemented the estimation of the intensity of infection with the Slide Positivity Rate ( SPR ) [32] S1 Fig 1 . SPR , defined as the number of laboratory-confirmed malaria cases per 100 suspected cases examined , provides a rapid and inexpensive means of assessing the burden of malaria in the population that relies on health care facilities . SPR provides a better estimate of the true intensity of malaria than malaria incidence ( reported cases divided by population ) , as it is more robust to spatial sampling errors/biases resulting from variation in the population size of the wards [32] . SPR is independent from incidence , although it can be prone to different biases ( as the result for example of other febrile diseases which would increase the number of blood tests taken ) . Confirming that the patterns are consistent across these two measures of malaria intensity is important because the population denominator that is relevant for the reporting may vary in ways that are difficult to determine . Also , the local population size might be under- or overestimated by a census carried out only once every 10 years in India . To further characterize spatial variation in risk we applied a 2k-means cluster algorithm to the incidence data at the wards level , and examined the existence of at least two regions differing in malaria risk . ( We pre-determined two groups of wards to consider the hypothesis of different transmission intensity in the core and periphery of the city ) . Also , two groups allow us to consider the subdivision of wards into high and low risk regions . We hypothesized that differences in malaria risk in these two main areas are largely explained by demographic and socio-economic factors . To test if the two regions differed significantly in those variables , we first extracted socioeconomic indicators , including slum density ( number of slums/ward area ) , unemployment ( number of unemployed people ) , number of marginal workers , literacy , population below 6 years , total population , number of households , vulnerable and economically deprived communities , from the 2001 Ahmedabad census and calculated the density of slums per ward based on cartographic information on the slums’ distribution within the city S2 Fig . A t-test was applied to evaluate if these two regions differed significantly in these covariates . We then addressed whether the temporal variation in malaria incidence responded differentially to climate variables ( rainfall , temperature and humidity ) across the two regions . To determine which predictors explain the temporal variation in vivax cases , we considered models with autoregressive terms to account for serial correlation in the data . The correlation structure in the malaria time series was assessed by inspecting the autocorrelation function ACF . Then we applied a generalized linear model ( GLM ) framework . Because observed count data , such as reported cases in infectious diseases , often exhibit significant over-dispersion [33] , a negative binomial distribution of cases was used . Backwards model selection was based on the minimum Akaike Information Criterion ( AIC ) ( S1 Table ) . Variables with coefficient significantly different from zero were selected . Spatial correlation was incorporated by assuming a conditional autoregressive ( CAR ) process in the random effect vi: vivj=N ( ∑jaijvj∑jaij , σv2∑jaij ) where σ2 controls the strength of the local dependence , and aij are neighborhood weights for each ward based on distance to the river . We additionally compared the best model to a model with a different distribution ( Zero Inflated Poisson ) but this model was not significantly better ( S2 Table ) . Finally , we considered the full spatio-temporal variation at the level of wards , by fitting a generalized linear mixed model including as covariates the effects of temperature and humidity , and a spatially-structured random effect weighted by the distance to the river ( SSRE ) . Parameters and their distributions were estimated with Bayesian Markov Chain Monte Carlo ( MCMC ) parameter sampling implementation in WINBUGS . To model malaria risk within the city of Ahmedabad an inhomogeneous Markov chain model was used , following the theoretical framework developed for cholera by Reiner et al . 2012 . In this approach the monthly malaria cases are categorized into discrete states of malaria incidence , which we chose as “low malaria” , “mild malaria , ” and “high malaria” . The three discrete states partitioned the distribution of monthly incidence based on the 25th , lower than 75th and above 75th quantiles . Then , the model assigns baseline probabilities Pi , j to the transitions between these states in a defined time step as described by the following transition probability matrix P: Pi , j=[P1 , 1 ( 1−P1 , 1−P1 , 3 ) P1 , 3P2 , 1 ( 1−P2 , 1−P2 , 3 ) P2 , 3P3 , 1 ( 1−P3 , 1−P3 , 3 ) P3 , 3] Where the baseline probabilities depend on the state of the system only in the previous time step . Modification of this basic matrix can introduce the effect of covariates , including the consideration of different regions such as the two groups of wards , identified in previous analyses . Thus , the resulting Markov chain model can be made inhomogeneous by allowing transition probabilities to depend on temporal and spatial environmental drivers; here seasonality , the state of the neighboring wards , temperature , Relative humidity and socioeconomic heterogeneity ( summarized by the two groups of wards ) : Pi , j , k , t=Pi , j , d⋅Seasi , j , t , d⋅Neighi , j , v , d⋅tempt , d⋅RHt , d Where Pi , j , k , t is the probability that ward k goes from state i to j from time t to time t+1 . This probability is dependent on: ( 1 ) Pi , j , d , the baseline transition probabilities of moving from state i to state j for a ward in risk region d; ( 2 ) a seasonal factor Seasi , j , t , d= ( 1+βi , j , d ) Se ( t , d ) , where the seasonality exponent Se ( t , d ) is periodic over the 12 months of the year and each group d has its own seasonality; ( 3 ) a neighborhood effect Neighi , j , v , d= ( 1+∝i , j , d ) t , dv , where v corresponds to the malaria state ( 0 , 1 , 2 ) of the neighboring ward with the highest value . This function reduces to 1 when none of the neighboring wards have malaria reported . The parameters of these functions , βi , j , d and ∝i , j , d are estimated . Finally , the effects of temperature and humidity are included as sigmoidal functions , similar to the formulation for ENSO and its effect on cholera in Reiner et al 2012: temp/Rhi , j , t , d= ( Adtan ( hdTemp/RHt−γ ( 2Md ) ) tan ( hd2 ) ) where Temp/Rhti , j , dis the temperature or humidity with temporal lag γ , in this case 2 months for the humidity and 1 month for the temperature , A is the maximal amplitude , M controls the scale ( normalization by the maximum values allowed for the climate factors ) , and h is shape factor varying between 0 and ~ π to go from a linear to a nonlinear effect ( S3 Fig ) . We considered different models obtained by including or neglecting the effect of a subset of the following factors: temperature , relative humidity , the state of the neighboring wards and the two different risk regions . We compared each of the models to a null model ) employing a likelihood ratio test . The most complex model has 78 parameters ( S3 Table ) . Under Markovian assumptions , the transitions for the different time steps ( months ) are independent from each other [31 , 34] , which allows us to explicitly write a likelihood . The constraint that each transition probability must be between 0 and 1 was imposed by a barrier method ( i . e . by setting the probability to 0 whenever its estimated value falls outside these limits ) [31] . Each model was fitted by maximizing the likelihood with a Nelder-Meade simplex algorithm , which allows for the incorporation of such constraints . Finally , to assess prediction performance , a cross-validation approach was implemented by sequentially removing the epidemic months ( August-November ) that follow the monsoon in a given year , refitting the model to the remaining data , and simulating it four months ahead starting from August to predict the course of the seasonal outbreak for the omitted period . Forecasting accuracy was estimated by computing the likelihood of the observed state . To that end , we inferred the probability distribution of the predicted state by performing 5000 independent simulations . This procedure is then sequentially repeated removing , one at a time , all the epidemic seasons available . To quantify the accuracy of our predictions , we calculate the percentage mean absolute error in our predictions , as well as a second quantity more practical and possibly relevant to public health , based on the definition of a ‘large’ outbreak . We defined such as event as one where the peak of the epidemic at the whole city level exceeds the 75% quantile of the distribution of this quantity . We quantify the fraction of times the model correctly predicts the observed malaria incidence state ( above the 75% quantile ) . Then , to examine and illustrate the importance of the climate covariates to the predictions , we simulated the model using different combinations of humidity and temperature ‘data’ . In particular , predictions for the epidemic months in the anomalous , low incidence years , were obtained using: 1 ) monthly observed temperature and average humidity 2 ) monthly average temperature and observed humidity 3 ) monthly average humidity and average temperature . Monthly averages were computed based on the mean of all previous years for a given month . We also examined the performance of the model by obtaining a one-step ahead prediction , where we removed 1 month of data at a time for all the wards ( S4 Fig ) .
We initially addressed whether the spatial distribution of normalized Plasmodium vivax risk is heterogeneous throughout the city , and how the ranking of risk varies in time . The top left panel of Fig 2 shows that several locations systematically rank high or low based on their P . vivax incidence through time . The top right panel in this figure reveals a stable spatial regularity of the locations with the highest malaria burden through time . This remarkable temporal stability of the spatial pattern suggests the existence of strong underlying determinants that are largely stationary at the temporal scales of malaria transmission within the city . The spatial pattern observed in Fig 2 also indicates the existence of two distinct malaria risk regions within Ahmedabad: one comprised of the wards close to the Sabarmati River and the core of the city ( region 1 ) , and the other , of those in the newer urban periphery ( region 2 ) . Fig 3 shows the results of the cluster analysis using the rankings data , demonstrating that the group of wards that are close to the river and in the inner part of the city ( referred to as high risk region hereafter ) have a different malaria risk than those in the in the periphery ( low risk region ) . Plasmodium vivax cases in the region defined as high risk are significantly ( p<0 . 001 ) higher than those in the region defined as low risk , whereas the seasonal pattern remains the same for both areas . This results hold for both parasites S5 Fig . We then asked whether the differences in malaria risk between the low and high malaria risk regions are associated with differences in population density ( number of slums , population size , number of households ) or economic level ( income , unemployment , literacy ) . Table 1 summarizes the results from statistical comparisons ( t-test ) between the high and low risk regions ( for P . vivax ) based on socioeconomic indicators from the 2001 census . We found that the high risk region has significantly higher unemployment , slum density , total population and number of households . In addition , we did not find significant differences in the mean area of the wards . The differences in slum density , unemployment , literacy , economically disadvantaged people and vulnerable communities ( both Scheduled Castes , SC , and Scheduled Tribes , ST , the official designations given to two different groups of historically disadvantaged people ) are pronounced , with the high risk region encompassing for example a density of slums that is at least 1 . 2 higher than that of the low risk region ( Table 1 ) . Moreover , the temporal variation in malaria incidence between the two regions could be influenced differentially by the environmental covariates . Table 2 show the results of the stepwise multiple linear regression between Plasmodium vivax cases and climate covariates ( temperature , humidity and rainfall ) and S6 Fig show the residuals of our best model for each region . The temporal variation in the low risk region responds predominantly to changes in temperature and humidity , whereas that of the high risk region is explained predominantly by humidity , although both parameters are seasonally associated with rainfall . Humidity would act in both the temporal and spatial dimensions , as the water table and associated ground moisture should be higher in the high risk region closer to the river , and this parameter is known to affect survival of the adult mosquitoes [35] . Thus , increased humidity ( through changes in the water table ) can lead to an earlier onset of the Plasmodium vivax season and to higher incidence S7 Fig . Finally , results for the general linear mixed model show that the best model includes the effects of temperature and humidity , and the random spatial variation in the data weighted by the distance to the river ( S4 Table and S5 Table ) . Interestingly , Plasmodium falciparum also exhibits stable regularities in risk levels within the city S8 Fig . We find the existence of two different regions with contrasting incidence , largely consistent with those for Plasmodium vivax Fig 3 , and a significant difference in socioeconomic level and environmental conditions for these two regions ( S6 Table and S7 Table ) . For a more dynamical perspective , we used an inhomogeneous Markov chain model that incorporates the effect of spatial and temporal variation on malaria risk . Results are also consistent for both parasites . Specifically , the comparison of the different models analyzed ( S3 Table and S8 Table ) , shows that the best model is the one that accounts for the effects of seasonality , neighbors , temperature , humidity , and includes the two different risk regions identified above ( model 5 Table 3 ) . To test the significance of the individual components of each model ( alternative hypothesis ) against a model including seasonality only ( model number 1 , null hypothesis ) we employed a likelihood ratio test . Improvements in likelihood for models 3 to 8 are significant , and so are the effects of seasonality , temperature , neighbors and the two regions ( p<0 . 05 ) . Improvements in likelihood for model 2 ( seasonal effect + neighborhood effect ) are not significant at p = 0 . 05 . Moreover , models that account for the effect of spatial heterogeneity , represented by including the two risk regions , tend to perform better than those that do not incorporate this effect ( S3 Table and S8 Table ) . Finally , Fig 4 shows the predictions for the epidemic months for Plasmodium vivax , generated with the cross-validation procedure described above and the best model , which includes seasonality , temperature , the state of the neighboring wards , and spatial heterogeneity . The predicted rate of cases for the peak of the season is coherent with the observed rate . Further explorations of the model show the importance of temperature and humidity in the prediction of malaria risk Fig 4B . For example , the pronounced dip in the cases for years 2009 and 2010 is captured by the dependencies of the model on temperature and humidity , as those two years are dryer , warmer and less humid Fig 4C and 4D . Although our best model exhibits a tendency to under-predict the size of the peaks which we discuss below , it is able to capture the seasonality and to a reasonable extent , the interannual variation in the data . The anomalous decrease in the number of cases in 2009 and 2010 seasons can be explained by the variation in the climate covariates . Because the model is stochastic and it considers discrete states ( no malaria , low and high ) , we simulated repeatedly , and from these ensemble of simulations computed the mean number of cases in a given month for a given ward . Our simulations generate realizations of the stochastic process and therefore , configurations of the discrete states . To convert the discrete states to cases , we used the mean number of cases for each class . The red line corresponds to the wards mean observed cases of the city . The blue dots show the median of the simulated values and the blue shaded regions correspond to the 5th– 95th percentile range over 5000 simulations .
Most transmission models of vector-borne diseases tend to aggregate the data at large scales and treat transmission homogeneously in space [27–29] . However , at local scales spatial heterogeneity can significantly influence the risk of infection . In particular , urban environments can exhibit pronounced heterogeneity from rapid and unplanned urbanization . Spatial heterogeneity in the environmental conditions such as temperature and humidity or in socioeconomic level can affect mosquito ecology , such as habitat distribution , vector longevity , biting rate or host finding ability [36] , and in factors related to human exposure and susceptibility respectively [8 , 16 , 37] . For Plasmodium vivax in Ahmedabad , we found defined heterogeneity in malaria risk that is slow-changing and therefore largely ‘stationary’ in time , relative to the characteristic temporal scales of the population dynamics of the disease . The presence of such stable pattern suggests strong and spatially-structured determinants of malaria risk . In particular , the existence of two main regions with different risk was shown to be associated with socioeconomic level . Higher risk is largely concentrated in the inner part of the city where socio-economic indicators reveal higher poverty on average ( Fig 3 and Table 1 ) . These results are consistent with the previously described negative associations between malaria and socioeconomic status [38–41] . Disease persistence decreases with increasing employment , literacy and income S1 Fig 9 , with poor people more vulnerable to ineffective diagnosis and treatment for financial and cultural reasons , and less able to access antimalarial and anti-mosquito protection [42] . Consistent results for Plasmodium falciparum emphasize the driven nature of the patterns . The two different risk regions within the city were also shown to exhibit differential temporal responses to climate forcing . This finding underscores the importance of humidity to malaria transmission , with a higher water table in the high risk region possibly increasing relative humidity and affecting vector ecology . The spread of malaria requires favorable conditions for the survival of both the mosquito and the parasite . Temperatures in the approximate range of 21°-32°C and a relative humidity of at least 60% are most conducive to transmission [38] . Malaria vectors need to live at least 8 days in order to transmit malaria , and higher humidity increases both survival rate and activity rate [43 , 44] . These relationships explain why Anopheles stephensi is more active and prefers feeding during the night when relative humidity is higher . The low risk region should have lower humidity ( ground moisture ) given its distance from the Sabarmati river and other water bodies . It has been established that if the average monthly relative humidity ( measured at 8 am ) is below 60% , then the lifespan of the mosquito is too short leading to very low or no malaria transmission [4 , 45 , 46] . The effect of humidity is also evident in the estimated regression coefficients ( Table 2 ) , where the value corresponding to humidity is one order of magnitude higher for the high than the low risk region . This means that a typical annual change of 10% in humidity with the rest of the covariates kept the same for both regions , will result in an increase of 403 to 1030 additional malaria cases ( or 15% to 39% relative to the mean ) in the high-risk region , and 70 to 411 additional cases ( or 7% to 33% relative to the mean ) in the low risk region . Rainfall itself , which is closely related to both humidity and temperature , is not retained in the regression as a significant explanatory variable probably because of its collinear effects with the other climatic factors . Our results from the mixed model indicate however that while climate factors ( temperature and humidity ) play an important role in disease transmission , it is their combined effect with spatial heterogeneity in the risk weighted by the distance to the Sabarmati river that better explains the spatio-temporal variation in the malaria incidence . More mechanistic models informed by time series data on disease incidence and vector abundance , will provide further insight on the role and interplay of climate covariates through their effects on parasite mortality , as well as on vector breeding and longevity . Social and economic elements such as the quality of housing can also favor the biological development of mosquitoes [47] . It is common in urban areas of India for water to be supplied irregularly; this leads to water storage within houses which creates multiple breeding sites for the mosquito in overhead tanks , cisterns and cement tanks [48] . Higher population density would result in higher water storage concentrations in close proximity to people . At the local scale spatial heterogeneity in urban malaria risk would follow from high density , especially where coexisting with poverty , as well as environmental microclimates from the proximity of water bodies . Our dynamical and stochastic model captures the seasonal pattern and the main trends in the interannual variation of the malaria cases . Interestingly , most of the models that incorporate spatial structure , namely the two regions , perform better than the models that do not . This conclusion is consistent with the results for diarrheal diseases in Dhaka , where consideration of different parts of the city also improved model performance [31 , 48] . Our best model identifies significant spatial effects at two different scales: ( 1 ) that of neighboring wards ( p<0 . 01 ) , where the probability of transitioning to a higher risk level depends of the level of the surrounding wards , and ( 2 ) that of the two regions ( p<0 . 001 ) influenced socioeconomic and demographic level . Although our model under-predicts the size of outbreaks , this tendency is expected from the discretization of the cases into a small number of levels , which preserves the rankings of the cases over time but tends to reduce the magnitude of the peaks . Predictions in that scale can still be useful when evaluating the risk of an outbreak larger than a given selected threshold [31] . Although additional classes could be incorporated in the formulation of the model , this would rapidly increase the number of parameters . The model is able to capture the interannual trends and in particular , the lower outbreaks of years like 2009 and 2010 , based on the effect of climate covariates . These two years exhibit anomalous high temperature and low humidity ( associated with low monsoon rainfall ) , only comparable to values in 2002 , another year with low incidence S10 Fig . The predictions of Fig 4B , for temperature and humidity kept at monthly averages , show that the model over-predicts the number of cases Fig 4 . Thus , the anomalous incidence is explained by the lower temperature and higher humidity . Fig 4C and 4D isolate the effect of each climate covariate on the prediction , showing a stronger effect of temperature on the temporal reduction of cases of those two years . Besides prediction , the phenomenological modeling framework applied here is also useful to address the spatial scale at which to aggregate the data to consider process-based epidemiological models , in a way that balances reducing the noise with representing dominant spatial heterogeneity . Questions on the spatial scale of aggregation are specifically relevant to addressing climate forcing in the context of socio-economic heterogeneity . Given that pronounced changes in urbanization will co-occur with those in climate , these are fundamental questions for infectious disease dynamics within cities . Although our approach is able to capture the interannual variation in the data and predict the peak of the epidemic , it could be improved in several directions . For example , one could incorporate in the model: ( 1 ) mobility fluxes derived from the spatial distribution of the population with movement models , to replace the near-neighbor effects on transition probabilities; ( 2 ) the explicit effect of population density on group-dependent parameters explicitly; ( 3 ) further analysis of the local effect of environmental heterogeneities such as river discharge and soil moisture on malaria incidence at higher resolution by increasing the number of groups in the model . Moreover , temporal changes of the city itself would be of interest , including changes in the local speed of urbanization , and their implications for mobility , population distribution and economic level . The region has experienced strong malaria interventions in the last three decades reflected in the pronounced negative trends in the number of reported cases from the 1980s and 19990s to the 2000s . From 2000 onwards , malaria prevalence in the city of Ahmedabad has remained however fairly stationary . Although the spatio-temporal variation in intervention efforts could influence the results of our models . This is unlikely given that the interventions within the city are largely homogeneous in space ( S11 Fig ) , with small differences between the high and low risk regions . Our probabilistic model tends to underpredict the size of outbreaks . This bias results from the transformation of the incidence data into discrete malaria levels , which smooth’s out extreme events . The effect of this bias can be assessed and corrected by lowering the threshold probability ( the proportion of simulations with large outbreaks ) below 50% using ROC ( Receiver-Operating-Curves ) ( Reiner et al . , 2012 ) . The number of discrete classes describing the malaria levels could also be increased or estimated to balance complexity and accuracy . At the limit , one could move to stochastic models that do not require such discretization , although their parameterization would present challenges related to model complexity . A better understanding of urbanization and malaria is needed , since urban environments can contribute to the persistence of the disease and frustrate elimination efforts more broadly at a regional level , by creating a reservoir for the disease in cities that contributes to transmission in rural areas . In India , the earlier National Eradication Programs focused on rural areas , with urban malaria contributing to the resurgence of disease in the 1970s [48] . Urbanization and growing populations also exacerbate inequalities in access to water , and in so doing introduce variation in another fundamental but poorly understood environmental factor for vector-transmitted infections . In concert these two dimensions , environmental and socio-economic , define the relevant spatial scales at which to address transmission dynamics . They also define the existence of spatial 'hotspots' of high disease risk in urban environments , especially those of the developing world where economic inequality and variation in disease vulnerability can be pronounced . Identifying these hotspots for targeted intervention in urban environments , can contribute to the control of vector-transmitted diseases , and eventually to the elimination of malaria in India . | Urbanization and environmental change are the main driving forces of ecological and social change around the globe , specifically in developing countries and for human health . Cities in developing countries exhibit rapid and unplanned urbanization which creates heterogeneous environmental and socio-economic conditions , which can in turn lead to different risks of infection . Here we address the role of urban spatial heterogeneity in infection risk by Plasmodium vivax with an extensive surveillance data set from Ahmedabad , India , spanning 12 years . This parasite is one of the four malaria species in humans , and is today the dominant cause of malaria in Indian cities . Our results show clear spatial structure in malaria incidence within the city , dependent on wealth , population density and the climate ( temperature and humidity ) . Because this pattern of spatial risk is largely stationary in time , it can be incorporated in a prediction model and it identifies target ‘hotspots’ for intervention . Similar conclusions apply to the reported cases of Plasmodium falciparum which reinforces the importance of spatial heterogeneity in urban malaria more generally . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"and",
"health",
"sciences",
"ecology",
"and",
"environmental",
"sciences",
"parasite",
"groups",
"plasmodium",
"atmospheric",
"science",
"tropical",
"diseases",
"geographical",
"locations",
"india",
"parasitic",
"diseases",
"parasitic",
"protozoans",
"parasitology",
"seasons",
"apicomplexa",
"plasmodium",
"vivax",
"protozoans",
"urban",
"environments",
"humidity",
"malarial",
"parasites",
"people",
"and",
"places",
"asia",
"meteorology",
"earth",
"sciences",
"biology",
"and",
"life",
"sciences",
"malaria",
"organisms",
"terrestrial",
"environments"
] | 2016 | Population Density, Climate Variables and Poverty Synergistically Structure Spatial Risk in Urban Malaria in India |
Investigation of macromolecular structure and dynamics is fundamental to understanding how macromolecules carry out their functions in the cell . Significant advances have been made toward this end in silico , with a growing number of computational methods proposed yearly to study and simulate various aspects of macromolecular structure and dynamics . This review aims to provide an overview of recent advances , focusing primarily on methods proposed for exploring the structure space of macromolecules in isolation and in assemblies for the purpose of characterizing equilibrium structure and dynamics . In addition to surveying recent applications that showcase current capabilities of computational methods , this review highlights state-of-the-art algorithmic techniques proposed to overcome challenges posed in silico by the disparate spatial and time scales accessed by dynamic macromolecules . This review is not meant to be exhaustive , as such an endeavor is impossible , but rather aims to balance breadth and depth of strategies for modeling macromolecular structure and dynamics for a broad audience of novices and experts .
A detailed understanding of how fundamental biological macromolecules , such as proteins and nucleic acids , carry out their biological functions is central to obtaining a detailed and complete picture of molecular mechanisms in the healthy and diseased cell . Furthering our understanding of macromolecules is central to understanding our own biology , as proteins and nucleic acids are central components of cellular organization and function . Many abnormalities involve macromolecules incapable of performing their biological function [1–4] , either due to external perturbations , such as environmental changes , or internal perturbations , such as mutations [5–10] , affecting their ability to assume specific function-carrying structures . It has long been known that the ability of a macromolecule to carry out its biological function is dependent on its ability to assume a specific three-dimensional structure ( in other words , structure carries function ) [11 , 12] . However , an increasing number of experimental , theoretical , and computational studies have demonstrated that function is the result of a complex yet precise relationship between macromolecular structure and dynamics [13–21] . Most notably , in proteins , the ability to access and switch between different structural states is key to biomolecular recognition and function modulation [22 , 23] . The intrinsic dynamic personality of macromolecules [18] is not surprising and can indeed be derived from first principles . Feynman highlighted the jiggling and wiggling of atoms well before wet-laboratory techniques provided evidence of macromolecular dynamics [24] . In the late 1970s and early 1980s , it became clear that treating macromolecules as thermodynamic systems and employing basic principles allowed anticipating and simulating their intrinsic state of perpetual motion [25 , 26] . The thermodynamic uncertainty principle was coined by Cooper in [26] to refer to the inherent uncertainty about the particular state a macromolecule is or will evolve to at any given time . Cooper was among the first to employ tools from statistical thermodynamics to show that macromolecular fluctuations are a direct result of thermal interaction with the environment and that any detailed description of macromolecular structure and dynamics entailed employing probability distributions . Further work by Wolynes and colleagues continued in this spirit , popularizing a statistical treatment of macromolecules with tools borrowed from statistical mechanics and culminating in the energy landscape view [5 , 13 , 27 , 28] . Great advances have been made in the wet laboratory to elucidate macromolecular structure and dynamics . Nowadays , techniques such as X-ray crystallography , Nuclear Magnetic Resonance ( NMR ) , and cryo-Electron Microscopy ( cryo-EM ) can resolve equilibrium structures and quantify equilibrium dynamics . Macroscopic measurements obtained in the wet laboratory are Boltzmann-weighted averages over microstates/structures populated by a macromolecule at equilibrium . Though in principle wet-laboratory techniques are limited in their description of equilibrium structures and dynamics to the time scales probed in the wet laboratory ( a problem also known as ensemble-averaging ) , much progress has been made [29–31] . The ensemble of structures contributing to macroscopic measurements obtained in the wet laboratory can be unraveled with complementary computational techniques [32–36] . In addition , wet-laboratory techniques , such as NMR spectroscopy , can on their own directly elucidate picosecond-millisecond long relaxation phenomena [37 , 38] . Indeed , recent single-molecule techniques have achieved great success at bypassing the ensemble averaging problem and elucidating equilibrium dynamics [31 , 39–47] . Transitions of a macromolecule between successive structural states can be captured in the wet laboratory [31 , 46 , 48–53] . Wet-laboratory techniques can resolve key well-populated intermediate structures along a transition [52 , 54] , but they are generally unable to span all the time scales involved in a transition and so fully account for a macromolecule’s equilibrium dynamics . A complete characterization of macromolecular dynamics remains elusive in the wet laboratory due to the disparate time scales that may be involved . Dwell times at successive states along a reaction may be too short to be detected in the wet laboratory . The actual time a macromolecule spends during a transition event can be short compared to its dwell time in any particular thermodynamically stable or meta-stable structural state . Indeed , neither wet- nor dry-laboratory techniques can , on their own , span all spatial and time scales involved in dynamic macromolecular processes [55] . Macromolecular modeling research in silico is driven by the need to complement wet-laboratory techniques and obtain a comprehensive and detailed characterization of equilibrium dynamics . Such a characterization poses outstanding challenges in silico . In principle , a full account of macromolecular dynamics requires a comprehensive characterization of both the structure space available to a macromolecule at equilibrium as well as the underlying free energy surface that governs accessibility of structures and transitions between structures . Early work on protein modeling focused on short protein chains and simplified representations models that laid out amino-acid chains on lattices . These distinct choices made it possible to perform interesting calculations revealing key properties of protein folding and unfolding [56] , as well as predict quantities of importance in protein stability and function , such as pKas of ionizable groups [57] . On-lattice models incidentally also allowed key theoretical findings on the computational complexity associated with computing lowest free-energy states in the context of ab initio ( now also known as de novo ) protein structure prediction [58–60] . The computational complexity of finding the global minimum energy conformation was shown to be NP-hard . These findings made the case that sophisticated algorithms would be needed to complement wet-laboratory characterizations of macromolecular structure and dynamics for the purpose of elucidating biological function . The advent of Molecular Dynamics ( MD ) simulations and the concept of an energy function promised to revolutionize macromolecular modeling , as in principle the entire equilibrium dynamics could be simulated by simply following the motions of the atoms constituting a macromolecule down the slope of the energy function . Research in this direction was made possible by a growing set of equilibrium structures resolved in the wet laboratory , from myoglobin [61 , 62] and lysozyme [63] by 1967 to more than a hundred thousand structures now freely available for anyone in the Protein Data Bank ( PDB ) [64] . Seminal work in the Karplus laboratory on the MD method and in the Lifson laboratory on the design of consistent energy functions and simplified molecular models set the stage for a computational revolution in structural biology . Commercialization of computers was critical to this revolution . MD simulations had been shown successful in reproducing equilibrium properties of argon [65] , but it was McCammon and Karplus who provided the earliest demonstration in 1977 of the power of MD-based modeling to simulate protein dynamics [25]: a short 9 . 2 picosecond-long trajectory was obtained showing in-vacuum , atomistic fluctuations of the bovine pancreatic trypsin inhibitor around its native , folded structure . Realizing the power of MD simulations to extract precious information on macromolecular structure and dynamics , the Karplus laboratory democratized modeling by offering the CHARMM program to the computational community [66] . Further work by Karplus and McCammon showed that significant features of protein dynamics would only emerge over longer time scales . The simulation in [67] reached 100 picoseconds , but it would soon become clear that MD-based probings of macromolecular structure and dynamics were in practice limited by both macromolecular size ( spatial scale ) and time of a phenomenon under investigation ( time scale ) . A significant body of complementary work in macromolecular structure and modeling investigated non-MD based methods . In fact , two years earlier to the 1977 MD simulation by Karplus of equilibrium fluctuations of the bovine pancreatic trypsin inhibitor , Levitt and Warshel had presented a computer simulation of the folding of the same inhibitor through a simplified ( now known as coarse-grained ) model , in which each residue was reduced to one pseudo-atom , and an algorithm based on steepest descent [68] . Reproducibility of this work has so far remained elusive . Further work by Levitt and Warshel , prompted by the visionary Lifson at the Weizmann Institute of Science , focused on the design of a consistent energy function for proteins [69] . The idea was to come up with a small number of consistent parameters that could be transferable from molecule to molecule and not depend on the local environment of an atom . Once such an energy function was implemented , simple algorithms could then be put together by making use of the function , its first derivative ( the force vector ) , and the second derivative ( the curvature of the energy surface ) . It is interesting to note that though Lifson and Warshel were the first to introduce a consistent energy function , they did so for small organic hydrocarbon molecules . It was Levitt who realized that their parameters could be used to carry out calculations on proteins . In 1969 , Levitt published the first non-MD , steepest descent algorithm on a simplified model encoding only heavy atoms of the X-ray structures of hemoglobin and lysozyme [70] . This work was seminal for Levitt and Warshel to claim the first simulation of protein folding [68] . The algorithm used in these simulations was quite sophisticated , changing torsion angles , as proposed by Scheraga [71] , and using normal modes to rapidly compute low-energy paths out of local minima [72] . Further work on coarse-grained and multiscale models built with the quantum mechanics ( QM ) /molecular mechanics ( MM ) method proposed by Warshel [73] was seminal in allowing simulation to reach longer spatial and time scales . Warshel , who had a background in quantum mechanics , realized that large molecular systems could be spatially divided into a region demanding quantum mechanical calculations ( e . g . , due to bonds being broken ) with the rest sufficiently represented by empirical force fields . This method remains the cornerstone of modern multiscale modeling [74–80] and , together with the idea of representing complex systems in different resolutions at different time and length scales [76] , has allowed simulations to elucidate structures , dynamics , and the biological activity of systems of increasing complexity , from enzymes [74 , 77 , 81] to complex molecular machines [82–91] . In tandem with these developments , a new method , Metropolis Monte Carlo ( MC ) [92 , 93] , made its debut in computational structural biology . In 1987 , important work in the Scheraga laboratory introduced an MC-based minimization method to simulate protein folding [94] . In 1996 , the Karplus laboratory demonstrated the ability of MC simulations on a cubic lattice to simulate the folding mechanism of a protein-like heteropolymer of 125 beads [95] . Following work in the Scheraga laboratory further made the case for the utility of MC-based methods in studies of macromolecular structure and dynamics [96–98] . Kinetic MC methods were designed to address the lack of kinetics in the classic MC framework [99] . In light of contributions that gave birth to computational structural biology [100] , it is no surprise that the Nobel 2013 prize in chemistry recognized computational scientists , namely , Karplus , Warshel , and Levitt for their seminal work in the development of multiscale models for complex chemical systems [101–103] . Improvements in hardware over the last forty years have been critical to extending the reach of MD- and MC-based modeling . For example , MD-based studies have expanded their scope , scale , and thus applicability due to specialized architectures , such as Anton [104 , 105] , Graphics Processing Units ( GPUs ) [106–109] , and petascale national supercomputers , such as BlueWaters , Titan , Mira , Stampede [110 , 111] . The pervasiveness of supercomputing has spurred great advances in algorithmic techniques to effectively parallelize MD . Typically , in parallel MD , the interacting particles are spatially divided into subdomains that are assigned to different processors . In this framework , load balancing becomes an issue for large-scale MD simulations now performed on thousands of processors and involving billions of particles [112] . Many techniques now exist for dynamic load balancing [113] . In addition , while each processor is responsible for advancing its own particles in time , processors need to exchange information; accurate force calculations require knowledge of neighbor particle positions . Work in [114] describes recent strategies for efficient neighbor searches in parallel MD . Other techniques that permit parallelization of MD address and optimize force splitting in the context of the particle-mesh Ewald algorithm [115] . It is worth noting that many of these techniques are now integrated in publicly-available parallel MD code , such as NAMD [116] . Important contributions in enhancing exploration capability have also been made from non-MD or non-MC frameworks but rather adaptations of stochastic optimization frameworks often designed for modeling other complex , non-biological systems . These frameworks , though less mature than MD and MC , are summarized here in the interest of introducing readers to interesting complementary ideas . Algorithmic advances , whether to extend the applicability of MD- and MC-based frameworks or adapt other frameworks for macromolecular modeling , now allow predicting native structures of given protein amino-acid sequences [117–120] , mapping equilibrium ensembles , structures spaces and underlying energy landscapes of macromolecules [6 , 8 , 121–126] , revealing detailed transitions between stable and meta-stable structures [127–134] , modeling binding and docking reactions [135–137] , revealing not only equilibrium structures of bound protein-ligand or protein-protein assemblies but also calculating association and disassociation rates [138 , 139] , and more . This review aims to provide an overview of such advances . Given the rapidly growing body of research in macromolecular modeling , aiming to provide an exhaustive review would be a task in futility . For instance , while the development of molecular force fields is recognized as crucial to accurate modeling [140 , 141] , this review does not focus on force field development . Other important contributions due to the development of ever-accurate coarse-grained representations of macromolecules , solvent models , and multiscaling techniques are acknowledged , but the reader is referred to existing comprehensive reviews on these topics [76 , 142–144] . Instead , this review focuses on sampling methods for the exploration of macromolecular structure spaces and underlying energy surfaces for the purpose of characterizing equilibrium structure and dynamics . This focus is warranted due to the recognition that sampling remains a problem [102 , 128 , 145] . The goal is to introduce a broad audience of researchers both to most recent and exciting research from an application point of view , as well as highlight important algorithmic contributions responsible for recent advancements in modeling macromolecular structure and dynamics .
Simulating interactions of macromolecules with other macromolecules or small molecules is important to understand the molecular basis of mechanisms in the healthy and diseased cell . Typically , three categories of interactions are of interest to researchers: those of a protein with a small ligand , those of a protein with another protein , and those of a protein with other molecular systems that include DNA , RNA , and membranes . These specific applications can be approached in two different ways . One considers simply the problem of predicting the three-dimensional native structure of the complexed system from knowledge of the structures of the unbound units , whereas the other additionally simulates the process of the units diffusing towards and then binding with one another . For the problem of structure prediction , non-MD based methods are currently the norm . They include algorithms enhancing MC or adapting other stochastic optimization frameworks under the umbrella of evolutionary computation . For the problem of actually simulating the dynamics of interacting units , MD-based studies provide more detail but typically require more computational resources or algorithmic enhancements in order to surpass the long time scale often needed for a complexation ( binding ) event to occur . One of the challenges with modeling and simulating macromolecular interactions with other small molecules or macromolecules is the possibility of induced fit . Induced fit , introduced by Koshland in [154] , refers to the mechanism of an initially loose complex that induces a conformational change in either one or all loosely bound units , which then triggers a cascade of rearrangements ultimately resulting in a tighter-bound complex . The induced fit mechanism seems to question the idea that structure-guided studies can focus on shape complementarity first , but many wet-laboratory studies , as well as the success of complementarity-driven methods , have demonstrated that induced fit cannot describe all binding events [155] . In response , inspired by the free energy landscape view presented by Frauenfelder and Wolynes [13 , 27] , Nussinov and colleagues proposed a new concept to explain binding events , that of conformational selection , also known as population shift [156–158] . Conformational selection refers to the idea that all conformational states of an unbound unit are present and accessible by the bound unit . The binding or docking event causes a shift in the populations observed in the unbound ensembles towards the specific bound conformational state . Though Nussinov and colleagues were inspired by the free energy landscape view of Frauenfelder and Wolynes , it is worth noting that the conformational selection model is a generalization of a much earlier model , the Monod-Wyman-Changeaux ( MWC ) model [159] . The MWC model , also known as the concerted or symmetry model , proposed the idea that regulated proteins exist in different interconvertible states in the absence of any regulator , and that the ratio of the different states is determined by the thermal equilibrium . The MWC model has been credited with introducing the concept of conformational equilibrium and selection by ligand binding , though in its original formulation the model was restricted to two distinct symmetric states and to proteins made up of identical subunits . The review in [23] summarizes many studies that observe conformational selection for protein-ligand , protein-protein , protein-DNA , protein-RNA and RNA-ligand interactions . We highlight work in [160] , where unfolded structures of uncomplexed ubiquitin in explicit solvent were subjected simultaneously to restraints from NMR Nuclear Overhauser Effect ( NOE ) and Residual Dipolar Coupling ( RDC ) data comprising solution dynamics up to microseconds . The obtained ensemble of structures covered the structural homogeneity observed in 46 crystal structures of ubiquitin at the time; the majority of the crystal structures were in complex with other proteins . These results suggest that conformational selection rather than induced fit suffices to explain the molecular recognition dynamics of ubiquitin . While at face value the concepts of induced fit and conformational selection appear mutually exclusive , studies have shown that versions of each are indeed observed; for instance , conformational selection is usually followed by slight conformational adjustments . In 2010 , Nussinov and colleagues presented an extended view of binding events where conformational selection and induced fit were seen as complementary to each other [161] . In many cases , following conformational selection , minor adjustments of side chains and backbone are observed to take place to optimize interactions [161] . Based on such observations , extended models have been proposed that combine conformational selection , induced fit , and the classical lock-and-key mechanisms [162] . A better understanding of contributions of each of these three mechanisms has contributed over the years to several effective methods for modeling and simulating binding and docking events . A detailed review in the context of protein-ligand binding for structure-based drug discovery is presented in [163] . The overview below summarizes methods based on the lock-and-key mechanism , as well as methods based on the induced-fit and conformational selection mechanisms . While the lock-and-key mechanism allows disregarding flexibility , the other mechanisms clearly make the case for modeling the flexibility of the units participating in the complexation event . While the induced-fit mechanism seems to suggest that only MD-based methods can describe a complexation event , the conformational selection mechanism has inspired many non-MD methods to integrate flexibility during or prior to complexation , thus contributing to a rich and still growing literature . In the following we provide an overview of this work , guided by applications on protein-ligand binding , protein-protein docking , and protein-DNA docking . Modeling the structural flexibility of uncomplexed proteins is key not only to allow application of methods such as ensemble docking to the protein-ligand and protein-protein docking problems , but also to obtain detailed information on the role of protein sequence on structure , dynamics , and function . While it is in principle very difficult to map the entire conformation space and underlying energy landscape of a protein sequence , many methods are dedicated to specialized sub-problems . For instance , literature is rich in methods that obtain a sample-based representation of the equilibrium conformation ensemble of a protein . Other methods extend this characterization to proteins that exhibit not only local fluctuations around an average , wet-laboratory , equilibrium structure but indeed are characterized by multi-basin landscapes where distinct structural states have comparable Boltzmann probabilities . Many methods focus on such proteins and particularly on modeling transitions between similarly stable structural states as a way to obtain information on function modulation and changes to function upon sequence mutations . Other methods are dedicated to capturing allosteric regulation and identifying coupled motions not in the vicinity of binding sites . Yet others focus on obtaining detailed structural characterizations of meta-stable states and other states present at low populations , even in natively unfolded proteins , as a way to understand aggregation , misfunction , and other disorders . In the following we provide an overview of these applications , highlighting selected ones to showcase current capabilities . Many proteins undergo large conformational changes that allow them to tune their biological function by transitioning between different structural states , effectively acting as dynamic molecular machines [325] . Since it is generally difficult for wet-laboratory techniques to elucidate a transition in terms of intermediate conformations ( though successful examples exist [326] ) , computational techniques provide an alternative approach [327] . However , transition trajectories may span multiple length and time scales , connecting structural states more than 100Å apart . This length scale is up to two orders of magnitude larger than a typical interatomic distance of 2 Å . Transitions can also demand micro-millisecond time scales , which is six to 12 orders of magnitude larger than typical atomic oscillations of the femto-pico second time scale . Typically , three types of methods are applied to model structural transitions , MD-based methods , morphing-based methods , and robotics-inspired methods . MD-based methods typically have to employ powerful algorithmic enhancements to surpass high-energy barriers in structural transitions . However , cases exist when classic MD methods have been able to capture spontaneous transitions of allosteric proteins by monitoring the structural relaxation upon removal of the bound molecule from the binding pocket [328 , 329] . These works further highlight the utility of the conformational selection or population shift principle , as removal of the bound molecule prompts spontaneous movement towards a new equilibrium state . In cases of high-energy barriers , biased or targeted MD methods are useful to expedite transitions between given structures [127 , 330] , but the concern with such methods is that the transition trajectory may not correspond to the true one , as these methods modify the underlying energy landscape; the order of events in transition paths computed via targeted MD methods depends on the direction in which the MD simulations are performed . For example , an application of biased MD to capture transitions of Ras between its active and inactive structures resulted in unrealistic , high-energy structures [330] . It is worth noting , however , that recent work in [331] has proposed a technique to remove the length-scale bias from targeted MD simulations . Essentially , the technique formulates local restraints , each acting on a small connected portion of the protein sequence , resulting in a number of potentials that are then used in targeted MD simulations . The technique has been demonstrated effective on an application to the open ↔ closed transition in the protein calmodulin . The free energy barriers associated with the computed paths have been shown comparable to those obtained with a finite-temperature string method . In contrast to biased MD methods , accelerated MD methods do not change the entire landscape but only the relative height of the basins corresponding to the structures that need connecting with intermediate conformations [332] . Accelerated MD has been applied to several proteins to capture the transition of H-Ras between the inactive and active structural states [10] , map the structural and dynamical features of kinesin motor domains [91] , compute domain opening and dynamic coupling in alpha subunit of heterotrimeric G proteins [333] , and more . Representative results on an application of accelerated MD for capturing the dynamics of the Eg5 kinesin motor domain are shown in Fig 2 . Even accelerated MD methods are limited in their ability to elucidate transition trajectories that cross high energy barriers [10] . In contrast , the dynamic importance sampling ( DIMS ) MD method [334 , 335] is more effective at simulating macromolecular transitions with energy barriers . In DIMS , the next conformational state sampled to obtain a transition from a state A to a state B will be chosen to satisfy the most productive movement to B and cross the energy barrier . The productive movement is indicated by a robust progress variable , the instantaneous RMSD over heavy atoms between a conformation and the target structure . DIMS is integrated in CHARMM and has been tested on several systems [336] , including modeling of slow transitions in AdK [334] , folding of protein A and protein G , and conformational changes in the calcium sensor S100A6 , the glucose–galactose-binding protein , maltodextrin , and lactoferrin , showing good agreement between sampled intermediates and experimental data [336] . In particular , in [334] , DIMS is applied to sample the ensemble of open-to-closed transitions for AdK . AdK is an enzyme that regulates the concentration of free adenylate nucleotides in the cell by catalyzing the conversion of ATP and AMP into two ADP molecules . The enzyme undergoes a large conformational change in its transition between an open and a closed structural states , and this change has been observed even in the absence of a substrate . As a result , AdK is one of the few proteins for which wet-laboratory studies have been able to capture a great number of intermediate structures populated during the open-to-closed transition . For this reason , AdK is a poster system to measure the capability of computational methods to reproduce transitions in great structural detail . Work in [334] is one of the few to provide atomistic detail , as well as reproduce and map with great accuracy the location of known intermediate structures along the transition . Representative results are shown in Fig 3 . Morphing- and string-based methods provide an alternative way to compute transition trajectories . Morphing-based methods include MolMov [337] , FATCAT [338] , NOMAD-Ref [339] , MinAction [130] , Climber [340] , and more . In Climber , the interresidue distances in a given start structure are pulled towards distances in the goal structure , using harmonic restraints incorporated in a pseudo-energy function . MolMov and FATCAT interpolate linearly in Cartesian space or over rigid-body motions . NOMAD-Ref uses elastic normal modes and interpolates interresidue distances per the elastic network algorithm in [341] . MinAction solves action minimization equations at each of the provided structures assuming a harmonic potential at them . Other methods include those based on elastic network models ( ENMs ) [131 , 341] , the nudged elastic band , zero- and finite-temperature string methods [340 , 342–347] . In particular , the string-based methods make use of the committor function to account for not generally knowing the collective variables underlying the transition [343] , whereas methods based on ENMs show the ability of coarse-grained models at capturing allosteric transitions in supramolecular systems on the order of megadaltons [131] . In general , while efficient , all these methods tend to reproduce similar conformational paths in independent runs rather than provide a possibly heterogeneous ensemble of conformational paths realizing the transition . Work in [348 , 349] tackles this issue of possibly high inter-run path correlations with the weighted ensemble method ( WEM ) . WEM , originally proposed in [350] , has been shown a useful enhanced sampling method for off-equilibrium and equilibrium processes . WEM uses a multiple-trajectory strategy where MC trajectories spawn new ones upon reaching new regions of the conformation space . One of the first applications of WEM to path sampling was on a 72-residue domain of the calmodulin protein . Coupled with a united residue model , WEM was able to capture the transition between the calcium-bound and calcium-free structural states and compare well with brute force simulations in a fraction of brute-force simulation time . In [349] , WEM is used to investigate the mechanism of the conformational change that the 5HIR benzylhydantoin transporter Mhp1 undergoes from a state poised to bind extracellular substrates to a state that is competent to deliver substrate to the cytoplasm . WEM reveals a heterogeneous ensemble of outward-to-inward conformational paths and identifies two distinct modes of transport . Robotics-inspired methods have also been applied to model structural transitions . They rely on deep analogies between robot motion planning and macromolecular motion simulation . In particular , the T-RRT [351] and PDST [352] methods , adapted from tree-based robot motion planning frameworks , have focused on the problem of computing conformational changes connecting two given structures in small and large proteins . While T-RRT has been shown to connect known low-energy states of the dialanine peptide ( two amino acids long ) [351] , the PDST method has been shown to produce credible information on the order of conformational changes connecting stable structural states of large proteins ( 200–500 amino acids long ) [352] . Both methods control the dimensionality of the conformation space by either focusing on systems with few amino acids [351] or by employing very coarse-grained representations to limit the number of modeled parameters in large proteins [352] . Work in [353] extends the capability of these frameworks to address large conformational changes in proteins , such as calmodulin and AdK , while providing high-resolution intermediate conformations by employing fragment-based moves . Other work detaches the sampling of the structure space from analysis of motions [354] . MSM-based analysis of sampled conformations is conducted to compute average properties of interest , such as expected number of transitions connecting two given structural states in lieu of direct time-scale information . Protein folding and structure prediction are often treated as two sides of the same coin . Protein folding , however , focuses on uncovering the detailed series of conformational changes that a protein goes through from a denatured , unfolded state to its long-lived , equilibrium , folded state . The folded or native structure is the end-result of this process , but not the only goal . Indeed , there are many protein folding algorithms that employ information about the native structure in order to expedite the search for the folding mechanism . Structure prediction algorithms focus more on the end result; that is , the goal is to uncover the native , folded structure even if the process by which these methods do so does not resemble the physical folding one . In its broadest context , the protein folding problem aims to shed light on the physical code by which a protein amino-acid sequence determines the native structure , the speed with which proteins fold , and the design of effective algorithms for predicting the native structure from sequence . An extensive review of protein folding is presented in [355] . The credit with introducing the problem to the computational biology community goes to Kendrew and coworkers , who published the first structure of a globular protein , myoglobin and showed the complexity and lack of symmetry or regularity in protein native structures [61] . Since then , a general mechanism for folding has been elusive . Various paradigms have been proposed , evolving from the early days when folding was thought to proceed deterministically , through a unique series of conformations for a protein at hand , to the free energy landscape view founded upon description of an inherently stochastic but biased process . The latter emerged from polymer statistical thermodynamics and built evidence that protein folding energy landscapes are funnel-like , narrower at the bottom , as the freedom of the protein to populate low-energy regions is gradually restricted [5 , 28 , 356] . While the energy landscape view has inspired many folding and structure prediction algorithms , in itself there is no suggestion of a mechanism that can be followed to efficiently fold proteins in silico . Application of MD simulations to observe the rare transition of a protein from an unfolded state to a folded state have come a long way in both the size of the proteins that can be handled and the time scales that can be modeled . Hardware advances , improvements in force fields , coarse-grained models , multiscaling techniques , and novel enhanced sampling techniques for MD have been crucial to surpassing spatial and time scales . Atomistic MD simulations can now be afforded [357] , with supercomputers such as ANTON allowing running folding simulations of proteins of 50–100 amino acids for milliseconds [358] , and software such as GROMACS [359] , NAMD [116] , and AMBER [360] becoming more accessible and easy to use to many researchers . In the following we elect to highlight recent work that showcases the state of protein folding . We then proceed with an overview of complementary work in de novo structure prediction . Protein folding . Some of the most striking advances in protein folding with atomistic , equilibrium MD simulations in the presence of water molecules have come from the Pande group , particularly through the Folding@Home project [148 , 361–364] . In 2005 , van der Spoel and colleagues provided the first folding simulation that also predicted the native structure of a peptide based on the Gibbs energy landscape [365] . In 2010 , Shaw and colleagues successfully modeled the folding of a 35-residue protein in explicit solvent [147] . Soon afterward , Lindorff-Larsen and colleagues in the Shaw group managed to fold 12 fast-folding proteins of length up to 80 amino acids and diverse native topologies with atomistic detail and in explicit solvent [105] . Some striking observations were made from analysis of the folding trajectories of these small proteins , which generated much discussion in the protein folding community [366] . In addition to matching folding rates measured in the wet laboratory , work in [105] demonstrated that the folding trajectories contained discrete transitions between native and unfolded states , in agreement with barrier-limited cooperative folding . Pathway heterogeneity was shown to be minimal for nine of the 12 proteins , with pathways sharing more than 60% of the native contacts . These results naturally suggested that the pathways observed in simulation were variations of a single underlying folding pathway . The conclusions in [105] were also supported by wet-laboratory work in [367] , which detected a limited set of pathways and only four intermediates for the folding of the calmodulin . Moreover , in [105] it was observed that long-range contacts locking in place the native fold formed early along , together with a significant amount of secondary structures and surface burial . This was confirmed in other folding simulations , as well [368] . While the amount of residual structure is questioned by wet-laboratory studies and may possibly be the result of the bias of current force fields [366] , the observations in [105] build the case for sequential stabilization as a mechanism for the folding of small , fast-folding proteins . The term sequential stabilization , coined in [369] , refers to the fact that folding may not be completely cooperative but is characterized by small-scale events that add secondary structure elements named foldons [370] in a stepwise manner . Because foldons are intrinsically unstable , low-energy paths are likely to involve foldons building on top of existing structures , thus resulting in sequential stabilization . Demonstration of the contribution and role of long-range native contacts early on in folding provided further justification for the use of Gō-models and other coarse-grained models that assume native contacts are the only ones that are kinetically-relevant [143] . However , while the wet-laboratory study of the folding of calmodulin in [367] demonstrated the presence of non-native intermediates in larger , more complex proteins , which is certainly observed in de novo structure prediction algorithms in the richness of non-native local minima . It is worth noting that a growing body of wet-laboratory studies are adding to the list of proteins known to fold through distinct native-like intermediates [371] . From a methodological point of view , a significant body of recent work in protein folding employs long , equilibrium , atomistic MD simulations in explicit solvent to observe multiple , spontaneous folding and unfolding events and reliably measure thermodynamic and kinetic quantities , such as folding rates , free energies , folding enthalpies , heat capacities , ϕ-values , and temperature-jump relaxation profiles [104 , 105 , 368] . While generally short , off-equilibrium MD simulations can at best sufficiently capture a single folding event , recent work that embeds many short off-equilibrium runs in coarse-grained kinetic models , such as MSMs , is able to approximate well the underlying folding dynamics [123 , 133 , 372] . Methods that embed many short simulations ( MD or other stochastic optimization methods ) in MSMs for the calculation of system dynamics is gaining ground in diverse applications , from folding , to structural transitions , to binding [128 , 132 , 354 , 373–375] . Lately , increasing attention is paid to the problem of characterizing the structure and dynamics of intrinsically-disordered proteins ( IDPs ) [394–396] . There are now growing databases of IDPS and intrinsically-disordered protein regions ( IDPRs ) , such as pE-DB , DisProt and IDEAL [397–399] . CECAM now regularly includes a workshop dedicated to promoting the development of new modeling methods and better understanding IDPs [400] . Since 2002 , even CASP provides an independent assessment of methods for IDPS [396] . Several reviews discuss the fundamental principles of disorder in the biological function of IDPs/IDPRSs biological functions , including the role of disorder in cancer , neurodegeneration , genetic forms of Parkinson’s disease , and cardiovascular diseases [401–405] . IDPs/IDPRs pose unique challenges in silico . They do not have stable tertiary structures but still demonstrate biological activity . This phenomenon challenges the fundamental structure-function relationship and is an extreme case of the exception to the lock-and-key model [395] . IDPs/IDPRs are not random coils . They exhibit different degrees of disorder , from molten globules to coils , but even coil-like structures exhibit residual structure [402 , 405] . A recent replica exchange MD simulation study revealed the structural contents of intrinsically disordered tau proteins . Tau proteins were discovered to be able to catalyze self-acetylation , which may promote pathological aggregation . The work characterized the atomic structures of two truncated tau constructs , K18 and K19 , providing structural insights into tau’s paradox [406] . IDPRs sequences are very different from those of ordered proteins , poor in hydrophobic amino acids and rich in charged amino aids . Disorder-promoting amino acids have now been identified , and they include Ala , Arg , Gly , Gln , Ser , Glu , Lys , and Pro [404 , 405] . Based on sequence information alone , tools now exist to estimate the propensity of a sequence for disorder [407] . There are many methods for disorder analysis and prediction of the location of disordered regions [124 , 408–411] . Computational methods are being designed to characterize structures and dynamics of IDPs/IDPRs . With specifically designed force fields , some methods have shown promise in this regard [412 , 413] . Treatment of IDPRs is now included in Rosetta [414] . Two main groups of methods focus on IDPs/IDPRs . The first group consists of wet-laboratory techniques based on NMR Chemical Shifts and RDCs [415] . The second consists of MD-based methods [152 , 153 , 408 , 416–418] . Both unrestrained MD [416] and long-range correlated MD [417] for well-characterized disordered proteins demonstrate good agreement with wet-laboratory data . The replica exchange with guided annealing method has also been shown suitable for IDPs [418] . The method escapes nonspecific compact states more efficiently and speeds up the generation of correct ensembles compared to classic replica exchange simulations . Work in [153] additionally shows the effectiveness of MD and MSMs for IDP modeling . Other methods combine NMR-based knowledge and MD simulations [6 , 302 , 314 , 413] . While NMR ensembles are better suited to characterize local conformational states of IDPs [415] , MD simulations allow calculating kinetics and elucidating meta-stable states and barriers between states [314] . Given their unique characteristics , computational methods are expected to continue their treatment of IDPS to better understand the connection between disorder and biological function and misfunction . The protein design problem is that of finding an amino-acid sequence whose global free energy minimum state corresponds to a desired , target structure or contains a structural motif associated with a desired function [419] . Also known as inverse folding or inverse structure prediction , this problem is now at the crux of protein engineering , with applications in medicine , biotechnology , synthetic biology , nanotechnology , biomimetics , and more [420] . Stated as an optimization problem , protein design is amenable to algorithmic frameworks employed for structure prediction . Computational approaches to protein design can be categorized into forward design , explicit negative design , and heuristic negative design [419] . In forward design , the sequence and target fold/structure of a protein are known , and the goal is to optimize the sequence so that the target structure reaches such a low energy that will make any other non-target structures less energetically favored . No explicit non-target structures are considered . A successful application of forward design has yielded a very stable protein , Top7 [421] , whose native structure was later shown identical to the determined X-ray structure . In explicit negative design , alternative structures are explicitly considered . The sequence is optimized so that the target native structure is lower in energy than all the alternative structures . Explicit negative design has been used to design specific coiled coils and DNA-binding and -cleaving enzymes [422–425] . The limitation of explicit negative design regarding prior knowledge and enumeration of non-favored alternative states has motivated heuristic negative design . In heuristic negative design , the goal is not to disfavor specific alternative structures; instead , the sequence is optimized through features that are likely to increase the energy of most undesired structures . Features follow closely strategies employed by nature to achieve the energy gap between the native structure and other structures that seems to be required for thermodynamic stability and function [419] . It is worth noting that conclusions regarding energy gaps between native and non-native structures when employing scoring functions need to be taken with a grain of salt . Work in [365] relates gaps in Gibbs free energy to structure deviations ( from NMR data ) . Compared to the other two strategies summarized above , heuristic negative design seems particularly important for biomolecular interactions [426 , 427] . Heuristic negative design also seems to be employed by nature for IDPs and by pathogens to fend off the host immune system [419] . Successful cases of designing proteins with novel functions abound [428–430] and are made possible by considerable advances in methods for de novo protein design . The current predominant computational approach is based on the ( inverse folding ) paradigm proposed in [431] , which assumes a fixed backbone and searches over discrete low-energy configurations/rotamers of side chains for rotameric combinations that result in a lowest-energy all-atom tertiary structure [432] . In the interest of tractability , energy models are limited to pairwise energy functions . State-of-the-art functions for protein design are knowledge-based , relying on statistical parameters derived from databases of known protein properties [433–437] . Even with such energy models , the design problem with a rigid backbone and a discrete set of rotamers has been proven to be NP-hard [438] . Two types of methods have been proposed to address the combinatorial optimization problem of finding rotameric combinations . The first are based on exact optimization and seek completeness; that is , finding the global minimum energy conformation . The second forego completeness and are based on heuristic optimization . Exact optimization methods include dead-end elimination [439] , branch-and-bound algorithms [440–442] , integer linear programming [443 , 444] , dynamic programming [445] , or cost function networks [446] . These exact methods are efficient and they limit inaccuracies to the inadequacy of the energy model , but their focus on one single assignment is highly subjective to possible artifacts in the energy function , known and lamented in [447] . Moreover , the solution provided by such methods may be overly stabilized ( effectively residing in a narrow basin ) , that it lacks the structural flexibility for the protein to operate the sought biological function under physiological conditions [448] . It is worth noting that unlike discrete rotamer assignments , work by Donald and colleagues pursues continuous rotamers and is able to reach lower-energy conformations [449] . This functionality is integrated in the popular OSPREY software [450] . It is expected that the design of a smoothed backbone-dependent rotamer library in [451] , which allows evaluating rotamer characteristics as smooth and continuous functions of the ϕ , ψ angles will lead to more advances in taking into account side-chain flexibility in de novo design . An illustration of the capability of protein design algorithms is provided in Fig 4 . Heuristic optimization methods for de novo design build on stochastic optimization or meta-heuristics , such as MC-SA [433 , 453] , Genetic Algorithms [454 , 455] , and other stochastic optimization methods [442 , 456 , 457] . Methods based on stochastic optimization , best represented by RosettaDesign [458] , currently dominate , mainly due to the ability to provide an ensemble of near-optimal solutions through their sampling-based approach . The backbone is kept fixed , and rotameric states are sampled systematically or in a sampling-based manner [433 , 453] over pre-built rotamer libraries [435 , 459] . All-atom energy minimization of the entire resulting all-atom conformation is often carried out [460 , 461] . It is here , in the minimization stage to which all constructed conformations and sequences are subjected , that localized backbone fluctuations are allowed . The extent of these fluctuations is small , limited to backrub motions [462–465] . Larger motions are allowed , but only on loop regions , made possible by efficient inverse kinematics techniques like Cyclic Coordinate Descent [320 , 466] . The importance of allowing backbone flexibility in the design process cannot be overestimated . The simple model of the backrub motion consists of a small dipeptide rotation about the C-Cα axis . Recent studies suggest that integrating backrub motions in the design process leads to improved designs of protein-protein interaction interfaces and more realistic templates with improved fit between simulated side-chain dynamics and NMR data [462 , 464 , 467] . Additionally , work in [468] has demonstrated that taking into account backrub motions expands sequence diversity during search and allows new residue interactions that rigid-backbone approaches cannot accommodate . This leads to better designs with lower energies and has been confirmed in other studies , as well [469 , 470] . Finally , an important highlight in protein design is the fact that , despite the absence of evolutionary history in newly-designed proteins , evolutionary information can be accommodated in the design process . Work in [470] reveals strong correlations between residue covariance in naturally-occurring protein sequences and sequences optimized for the same structures by computational protein design . Covariance has been demonstrated for complementary changes in residue size , residue charge , and hydrogen bonding [471–475] . These findings suggest that structural restrains on co-evolving residues in contact can lead to further improvements both in de novo protein design and structure prediction .
In the classic MD setting , Newton’s equation of motion is iteratively solved on a finely discretized time scale to observe collective movements of the atoms comprising a molecular system through successive conformations terminating at a local minimum conformation in the system’s energy surface . The ensemble of conformations obtained at equilibrium conditions observes the Boltzmann distribution . A distinct advantage of employing MD to simulate the equilibrium dynamics of a macromolecule is the ability to obtain great detail on individual and correlated motions of specific atoms and specific sites on a macromolecule , as well as correlated motions between macromolecular units of a complex . A disadvantage of the classic MD simulation setting is the inability to sample rare events that occur on long time scales . In particular , in the presence of high energetic barriers separating local minima in the energy surface , a classic MD simulation may be trapped and never escape within the time scale of the simulation . Limited sampling of the conformation space is a fundamental issue in classical MD , and algorithmic enhancements are proposed on a regular basis to enhance sampling capability . These include replica exchange , accelerated MD , umbrella sampling , biased or steered MD , importance sampling , activation relaxation , local elevation , conformational flooding , jump walking , multicanonical ensemble , MSM-driven MD , discrete timestep MD , swarm methods , and others [8 , 149 , 203–206 , 334 , 476–489] . Recent reviews of advanced MD-based methods and outstanding issues are discussed in [124 , 490–495] . A comprehensive list of commonly used MD packages for biomolecular simulation is presented in [493] . Examples of MD applications on proteins with large conformational changes that occur on long time scales , such as G-proteins , Ras-proteins , kinases , signaling proteins , and others can be found in [121 , 122] . In the following , we highlight some of the algorithmic enhancements to the classical MD setting that are responsible for surpassing traditional MD time scales and characterizing the dynamics of complex systems . While a significant portion of research on macromolecular structure and dynamics employs MD-based methods , a just as significant portion employs MC sampling . In MC , the evolution of one conformation into another is not guided by Newton’s equation of motion but instead a programmed move or step designed to introduce a small or large conformational change . The end result of the move is only accepted according to the Metropolis criterion in order to promote the trajectory of consecutive conformations to converge to the global minimum while allowing some non-zero probability of escaping a current minimum . MC-based methods employ the notion of effective temperature to regulate the height of energy barriers that can be crossed . While generally regarded to have higher sampling capability than MD , MC methods also are prone to convergence to local minima and forego any direct information of time scales and kinetics . Many of the enhancement strategies for MD can be applied to MC-based methods . In the following we highlight two such enhancements . An important group of methods to address optimization-related problems in macromolecular modeling consists of evolutionary algorithms ( EAs ) . EAs approach stochastic optimization under the umbrella of evolutionary computation , where the main idea is for computation to mimic the process of evolution and natural selection to find local optima of a complex objective/fitness function . The realization that the potential energy landscape of a macromolecule can be non-linear and multimodal , and that many structure-centric macromolecular modeling problems can be cast as optimization problems makes EAs highly appealing for macromolecular modeling . Though EAs are highly customizable algorithms , they all follow a simple template . A population of samples of a configuration space ( generally referred to as individuals ) is evolved over a number of generations . An initialization mechanism specifies the initial population , which can consist of random samples or include configurations known to be local optima ( for instance , experimentally-available structures may play this role ) . The population evolves either over a fixed , user-defined number of generations or until a different termination criterion is reached . In each generation , individuals with high fitness are repeatedly selected and varied upon . The selection mechanism specifies which individuals to select as parents for reproduction . The improvement mechanism consists of reproductive or variation operators , which can be asexual , introducing a mutation on a parent , or sexual , combining the material of two parents at one or more crossover points to generate offspring . A survival mechanism determines which individuals survive to the next generation . In non-overlapping or generational survival mechanisms , the offspring replace the parents . In overlapping ones , a subset of individuals from the combined parent and offspring pool are selected for survival onto the next generation . A comprehensive review of EAs can be found in [647] . EAs are very rich algorithmic frameworks , as different design decisions in the initialization , variation , selection , and survival mechanisms can lead to very different behaviors . The decision on how to represent individuals is key both to the effectiveness and ease with which variation operators can be designed to produce good-quality individuals . EAs that employ crossover in addition to the asexual ( mutation ) operator are referred to as genetic algorithms ( GAs ) . EAs that additionally incorporate a meme , which is a local improvement operator to improve an offspring and effectively map it to a nearby optimum , are referred to as hybrid or memetic EAs ( MAs ) . The employment of multiple , independent objective functions as opposed to a single fitness function results in multi-objective EAs ( MO-EAs ) . Specific variants that build over GA are respectively referred to as MGAs and MO-GAs . One of the first EAs for macromolecular structure modeling was a GA , proposed in [648] for the de novo protein structure prediction problem . Work in [648] also demonstrated that EAs are better able to escape local minima of a protein energy function than MC [648] . This result is not surprising , considering that the algorithm able to compute Lennard-Jones optima of atomic clusters in [649] was in fact an EA . Referred to as Basin Hopping , the algorithm was a 1+1 MA , which refers to an MA that has only one parent and one offspring . In a 1+1 MA , the population evolving over generations has size 1 , and the offspring competes with the parent . We recall that MA refers to an EA where the offspring is subjected to a local improvement operator ( energetic minimization ) . In Basin Hopping , the offspring replaces the parent with a probability resembling the Metropolis criterion . An MC search can also be viewed as an EA , specifically , a 1+1 EA , and all MC-based methods can be conceptualized as EAs employing highly specific insight about the optimization problem at hand . Given the early work in [648] , EAs have a long history in de novo protein structure prediction . Customized EAs for this problem contain many evolutionary strategies and meta-heuristics , including the employment of a hall of fame to preserve “good” individuals ( decoys ) , tabu search to improve the performance of a meme , co-evolving memes , niching , crowding , twin removal for population diversification , structuring of the solution space to facilitate distributed implementations capable of exploiting parallel computing architectures , and more . The main focus of algorithmic research on EAs is what mechanisms avoid premature convergence and allow finding the global optimum in overly rugged fitness landscapes . This is of particular interest on applications of EAs for different structure-centric problems in macromolecular modeling [650] . A comprehensive review of EAs for de novo protein structure prediction can be found in [651] . Though they have a long history in de novo structure prediction , EAs are not considered among the top performers in this problem for proteins no longer than 200 amino acids . On long protein chains , where off-lattice models result in impractical computational demands , on-lattice EAs are by now the only viable algorithms [652 , 653] . However , on shorter chains , where off-lattice models can be afforded , the injection of specialized operators ( moves ) , such as molecular fragment replacement , and sophisticated hybrid potential energy functions have allowed rather simple MC-based algorithms to outperform non-customized EAs . Of note here are the Rosetta and Quark methods that often dominate the leader board in the CASP competition [118–120] . Even though EAs have yet to become state of the art in the de novo structure prediction setting , much progress has been made in recent years [390 , 391 , 654] . Recently , EAs have incorporated state-of-the-art , off-lattice representations and energy functions to become competitive with MC-based methods such as Rosetta [390 , 391] . The additional recasting of the structure prediction problem as a multi-objective optimization one has resulted in higher exploration capability and conformation quality over single-objective optimization approaches such as Rosetta [392 , 655] . EAs are also employed to address protein folding [656] . While there is still much work to be done to demonstrate EAs as the state-of-the-art approaches for de novo structure prediction , there are three domains in macromolecular structure modeling where EAs are by now the best performers: protein-ligand binding , multimeric protein-protein docking , and cryo-EM reconstruction; In protein-ligand binding , some of the top algorithms are EAs . For instance , Autodock now employs a Lamarckian GA , which has been demonstrated to result in better-quality receptor-ligand bound configurations over the MC-SA algorithm employed in earlier releases [180] . In particular , work in [180] demonstrates that both the Lamarckian GA and a traditional GA can handle ligands of more degrees of freedom than MC-SA , and that the Lamarckian GA outperforms the traditional GA . The latter is due to the fact that in a Lamarckian GA , contrary to the Darwinian model of evolution , where only genetic traits are inheritable , an offspring is replaced with the result of the local improvement operator to which it is subjected . This results in essentially introducing phenotypic traits in the genotypic pool ( improvements are passed onto the next generation ) , per Jean Baptiste Lamarck’s now discredited claim that phenotypic characteristics acquired during an individual’s lifetime can be become inheritable traits; ( epigenetics is bringing more credibility , however , to Lamarck’s claims ) . It is worth pointing out that many MAs ( for instance , even Basin Hopping ) are Lamarckian EAs . MAs that are not Lamarckian choose not to replace the offspring with the result of the local improvement operator to which it is subjected but use the improved fitness in the survival mechanism; this is known as the Baldwin effect [657] . A domain where EAs are showing promise is in structure prediction for asymmetric , heteromeric assemblies . Currently , the only algorithm that has been shown capable of producing native asymmetric structures of heteromeric assemblies in the absence of wet-laboratory data is Multi-LZerD [292] . Multi-LZerD is a GA that represents multimeric conformations through spanning trees . The nodes in the tree represent the units , and the edges encode the presence of a direct interaction . As presented , Multi-LZerD proceeds over 3 , 000 generations . While promising , the algorithm incurs a high computational cost to be practical in its current form for multimeric assemblies of more than 6 units . Another domain where EAs are shown to be highly successful is the simultaneous registration problem in cryo-EM microscopy reconstruction . One issue with cryo-EM is that low-resolution maps are often obtained for large asymmetric and/or dynamic macromolecular assemblies . In such cases , an important problem is how to simultaneously fit known structures of the units in the given map . A GA with specialized variation operators and tabu search has been proposed in [658] to successfully address this problem . This GA has also been used in later work in [659] to trace α helices in low- to mid-resolution cryo-EM maps . While most of the work on EAs in the evolutionary computation community is driven by algorithmic design and analysis of the exploration capability rather than data quality , key ideas and strategies on evolutionary search are proving powerful in enhancing exploration capability in macromolecular structure modeling problems . For instance , several algorithmic decisions on how to select which parents for reproduction , generate offspring , and setup the competition for survival are key for balancing the breadth ( exploration ) and depth ( exploitation ) issue in exploration [647] . Lately , interesting ideas from multi-objective optimization are being incorporated in EAs for conformation sampling in de novo protein structure prediction . Namely , instead of pursuing the global minimum of an aggregate energy score , EA-based methods are proposed to obtain conformations that optimize specific sub-groupings of interatomic interactions [392] . EA-based methods are also showing promise in mapping energy landscapes of proteins with large conformational changes [324 , 660] . Due to the ongoing work in the evolutionary computation community on powerful and effective algorithmic strategies for obtaining solutions of complex objective functions and the realization of outstanding sampling bottlenecks in de novo structure prediction [661] , adoption of EAs holds great promise for macromolecular structure modeling . Since simulation of dynamics is the limiting factor in dynamics-based methods , efficiency concerns can be addressed by foregoing or at least delaying dynamics until credible conformational paths have been obtained . A different class of methods focuses not on producing transition trajectories but rather computing a sequence of conformations ( a conformational path ) with a credible energy profile . The working assumption is that , once obtained , credible conformational paths can then be locally deformed with techniques that consider dynamics to obtain actual transition trajectories . Such methods adapt sampling-based algorithms developed to address the robot motion-planning problem and are thus known as robotics-inspired methods . The objective in robot motion planning is to obtain paths that take a robot from a start to a goal configuration . The robot motion planning problem bears mechanistic analogies to the problem of computing conformations along a transition trajectory; in both problems the goal is to uncover what of the underlying conformation or configuration space is employed in motions of a mechanical or biological system from a start to a goal conformation or configuration . Analogies between molecular bonds and robot links and atoms and robot joints are made to perform fast molecular kinematics . Robotics-inspired methods are tree-based or roadmap-based [662] . Tree-based methods grow a tree in conformation space from a given , start to a given , goal conformation representing the structures bridged by the sought transition . The growth of the tree is biased so the goal conformation can be reached in reasonable computational time . As a result , tree-based methods are efficient but limited in their sampling . They are known as single-query methods , as they can only answer one start-to-goal query at a time; that is , only one path of consecutive conformations that connect the start to the goal can be extracted from the tree . Running them multiple times to sample an ensemble of conformational paths for the same query results in an ensemble with high inter-path correlations due to the biasing of the conformation tree . Roadmap-based methods adapt the Probabilistic Road Map ( PRM ) framework [663] . These methods support multiple queries . Rather than grow a tree in conformation space , these methods detach the sampling of conformations from the structure that encodes neighborhood relationships among conformations in the conformation space . Typically , a sampling stage first provides a discrete representation of the conformation space of interest , and then a roadmap building stage embeds sampled conformations in a graph/roadmap by connecting each one to its nearest neighbors . Roadmap-based methods bring their own unique set of challenges . Randomly-sampled conformations have very low probability of being in the region of interest for the transition . In particular , for long chains with many degrees of freedom ( hundreds of backbone angles in small-to-medium protein chains ) , a protein conformation sampled at random is very unlikely to be physically realistic . Biased sampling techniques can be used to remedy this issue [664 , 665] , but it is hard to know which ones will focus sampling to regions of interest for the transition . In addition , both roadmap- and tree-based methods rely on local planners or local deformation techniques to connect two neighboring conformations . It is hard to find reasonable local planners for protein conformations . A linear interpolation is often carried over the employed parameters , typically backbone angles , but this can produce unrealistic conformations , and a lot of time can be spent energetically refining these conformations . Recent work is considering complex local planners that are not based on interpolation but are instead re-formulations of the motion computation problem . Recent work in [666] introduces a prioritized path sampling scheme to address the computational demands of complex local planners in roadmap-based methods for protein motion computation . Roadmap-based methods have been employed to model unfolding of small proteins [665 , 667] . Tree-based methods have been employed to model conformational changes and flexibility , predict the native structure , and compute conformational paths connecting given structural states [351 , 352 , 387 , 668–670] . In particular , the T-RRT method described in [351] and the PDST method described in [352] have focused on the problem of computing conformational paths connecting two given structures . While T-RRT has been shown to connect known low-energy states of the dialanine peptide ( two amino acids long ) [351] , the PDST method has been shown to produce credible information on the order of conformational changes connecting stable states of large proteins ( 200–500 amino acids long ) [352] . Both methods control the dimensionality of the conformation space by either focusing on systems with few amino acids [351] or by employing coarse-grained representations to reduce the number of modeled parameters in large proteins [352] . The tree-based method in [353] employs the fragment replacement technique to reduce the dimensionality of the conformation space and sample conformational paths connecting two given structural states of proteins ranging from from a few dozen to a few hundred amino acids . At each iteration , a conformation in the tree is selected for expansion . The expansion employs molecular fragment replacement and the Metropolis criterion to bias the tree towards low-energy conformations over time . The selection penalizes the tree from growing towards regions of the conformation space that have been oversampled , thus resulting in enhanced sampling of the conformation space .
This review has highlighted the breadth and depth of research in macromolecular modeling and simulation . A plethora of computational methods have been developed to study a wide spectrum of molecular events . QM methods are used to study molecular electronic structures and obtain detailed and accurate electronic structure calculations . Work in [671] employs such calculations to correlate quantum descriptors and the biological activity of 13 quinoxaline drug compounds and then suggest effective compounds against drug-resistant Mycobacterium tuberculosis . Recent efforts in quantum chemistry are devoted to circumventing computational bottlenecks of large-scale electronic structure calculations and extending applicability to molecular systems composed of hundreds of atoms [672] . At present , QM methods have too high a computational cost to be a competitive alternative to MD or MC methods and their variants . For this reason , the focus of this review has been on MM methods , such as MD and variations , which are the methods of choice to study macromolecular structure and dynamics . It should be noted that hybrid , QM/MM methods exist and are the methods of choice for modeling reactions in biomolecular systems [673] . One of the major themes in MM-based macromolecular modeling is the choice of resolution or detail . As this review has summarized , atomistic , explicit solvent MD simulations are becoming more affordable , both due to improvements in hardware and techniques that allow aggressive parallelization . Despite the challenges posed by the disparate spatial and time scales employed by macromolecules flexing their structures and interacting with their environment , significant algorithmic and hardware advances have allowed breaking the millisecond barrier [147] . Dynamical processes that involve millions of atoms can now be characterized . For example , work in [674] tracks via MD simulations the microsecond-long atomic motions of 1 . 2 million particles to study the dissolution of the capsid of the satellite tobacco necrosis virus . MD and non-MD methods that employ reduced , coarse-grained macromolecular models are often regarded as “cheaper” albeit less accurate alternatives to atomistic MD methods . Such cheaper methods currently complement or facilitate atomistic MD-based studies . For example , protein docking methods are routinely employed to assist cryo-EM in resolving structures of molecular assemblies . Once such methods narrow down the possible conformation space , subsequent atomistic MD simulations are employed to make final predictions by examining stability and dynamics [111] . In some settings , these cheaper methods provide the only practical approach . Even with various accelerated MD simulations , mapping of protein energy landscapes remains challenging . For example , work in [10] shows that the sampling capability of accelerated MD greatly depends on the structure used to initiate a trajectory [10] . In our own laboratories , we have been able to compare the cheaper methods to published atomistic MD simulations of H-Ras [660] . In particular , on H-Ras , the evolutionary algorithm in [660] is able to map the energy landscape of H-Ras wildtype and selected variants in atomistic detail better than what can currently be achieved via known MD methods . In MD-based research , two different directions seem to be pursued by researchers at the moment . The first involves the employment of very long MD simulations , made possible by complex MD-customized architectures , like Anton . Thermodynamic and kinetic quantities can be readily extracted from such simulations . The second involves the employment of several short , off-equilibrium MD simulations , which allows the employment of parallel architectures but necessitates the employment of statistical models , such as Markov state models , to collect and organize the simulations to describe the long-time behavior of a system . Both directions are exciting and complementary . In particular , the second direction is leading to advances in the combination of continuous and discrete models for expediting modeling of long-time scale phenomena and is likely to lead to further algorithmic advancements . Within each of these directions , several open questions remain for researchers to pursue . A combination of both directions , dedicated architectures and continuous and discrete models promises to push the spatial and time scales that can be observed in silico even further . As summarized in this review , many non-MD algorithmic frameworks are being pursued to model different aspects of macromolecular structure and dynamics . Often , these frameworks are inspired or initiated from diverse communities of researchers . Of note here are evolutionary algorithms and robotics-inspired algorithms . While components of these algorithms are often investigated in detail in each of the corresponding communities , the focus in these communities has traditionally been on often on computational performance rather than quality of findings . Broad employment of these algorithms as tools complementary to MD is currently challenged by an inability to demonstrate utility on a broad class of macromolecular systems and validate findings with existing wet-laboratory or MD-based studies . Nonetheless , a growing body of researchers within each of these communities is introducing treatments focused on both computational performance and data quality . This review has summarized the current state of the art in diverse application areas . An emerging theme is the need to characterize in detail the structural flexibility of a macromolecular system under specific conditions . While great progress is being made , computing a conformation ensemble consistent with explicit or implicit constraints is likely to motivate the development of novel algorithms for years to come . Many other directions of research in macromolecular modeling and simulation could not be described in detail here . These include the development of accurate and sensitive molecular force fields [140 , 141] for macromolecular simulation , the development of increasingly accurate coarse-grained representations of macromolecules , solvent models , and multiscaling techniques [76 , 142–144] , decoy/model selection algorithms [675] in de novo structure prediction , as well as the development of algorithmic tools to assist structure resolution in the wet laboratory [676 , 677] . Additionally , while this review highlights some of the unique challenges posed by intrinsically disordered proteins and regions , it does not provide an overview of similar challenges posed by membrane proteins . The reader is referred to work in [678] for a review of such challenges and algorithmic advancements . Expected advances in each of the reviewed application areas promise to provide us with a more comprehensive and detailed understanding of our biology . In particular , unraveling the behavior of macromolecules in isolation and assembly will help us understand the molecular basis of mechanisms in the healthy and diseased cell . A truly synergistic employment of in-silico and wet-lab research to unravel molecular mechanisms also promises to lead to better therapeutics for combating cancer , neurodegenerative disorders , infections , and other important human disorders of our time . The journey into the future of computational structural biology promises to be exciting , and we hope that this review has inspired a few more researchers to join us on this journey . | This paper provides an overview of recent advancements in computational methods for modeling macromolecular structure and dynamics . The focus is on methods aimed at providing efficient representations of macromolecular structure spaces for the purpose of characterizing equilibrium dynamics . The overview is meant to provide a summary of state-of-the-art capabilities of these methods from an application point of view , as well as highlight important algorithmic contributions responsible for recent advances in macromolecular structure and dynamics modeling . | [
"Abstract",
"Introduction",
"Recent",
"Applications",
"Made",
"Possible",
"by",
"Hardware",
"and",
"Algorithmic",
"Advancements",
"Categorization",
"by",
"Algorithmic",
"Frameworks",
"Conclusions"
] | [
"applied",
"mathematics",
"simulation",
"and",
"modeling",
"algorithms",
"review",
"mathematics",
"protein",
"structure",
"prediction",
"protein",
"structure",
"macromolecules",
"thermodynamics",
"research",
"and",
"analysis",
"methods",
"polymer",
"chemistry",
"proteins",
"chemistry",
"biophysics",
"molecular",
"biology",
"free",
"energy",
"physics",
"biochemistry",
"biochemical",
"simulations",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"computational",
"biology",
"macromolecular",
"structure",
"analysis",
"biophysical",
"simulations"
] | 2016 | Principles and Overview of Sampling Methods for Modeling Macromolecular Structure and Dynamics |
Human T-cell leukemia virus type 1 ( HTLV-1 ) is the etiological agent of adult T-cell leukemia/lymphoma ( ATLL ) , an aggressive malignant proliferation of activated CD4+ T lymphocytes . The viral Tax oncoprotein is critically involved in both HTLV-1-replication and T-cell proliferation , a prerequisite to the development of ATLL . In this study , we investigated the in vivo contribution of the Tax PDZ domain-binding motif ( PBM ) to the lymphoproliferative process . To that aim , we examined T-cell proliferation in humanized mice ( hu-mice ) carrying a human hemato-lymphoid system infected with either a wild type ( WT ) or a Tax PBM-deleted ( ΔPBM ) provirus . We observed that the frequency of CD4+ activated T-cells in the peripheral blood and in the spleen was significantly higher in WT than in ΔPBM hu-mice . Likewise , human T-cells collected from WT hu-mice and cultivated in vitro in presence of interleukin-2 were proliferating at a higher level than those from ΔPBM animals . We next examined the association of Tax with the Scribble PDZ protein , a prominent regulator of T-cell polarity , in human T-cells analyzed either after ex vivo isolation or after in vitro culture . We confirmed the interaction of Tax with Scribble only in T-cells from the WT hu-mice . This association correlated with the presence of both proteins in aggregates at the leading edge of the cells and with the formation of long actin filopods . Finally , data from a comparative genome-wide transcriptomic analysis suggested that the PBM-PDZ association is implicated in the expression of genes regulating proliferation , apoptosis and cytoskeletal organization . Collectively , our findings suggest that the Tax PBM is an auxiliary motif that contributes to the sustained growth of HTLV-1 infected T-cells in vivo and in vitro and is essential to T-cell immortalization .
HTLV-1 ( Human T-cell leukemia virus , type 1 ) is the etiological agent of adult T-cell leukemia/lymphoma ( ATLL ) , an aggressive and fatal form of leukemia characterized by the malignant expansion of activated CD4+ T-cells [1] . Among several non-structural regulatory proteins encoded by HTLV-1 , Tax , a crucial transcriptional activator of the viral life cycle , exerts pleiotropic effects during the initial stages of the multistep leukemic process [2] . This viral protein modulates the expression of cellular genes leading to the deregulation of T-cell proliferation , perturbing the integrity of cell cycle checkpoints , the DNA damage response and apoptosis pathways [3–6] . Like other viral oncoproteins such as human adenovirus E4-ORF1 and human papillomavirus ( HPV ) E6 , Tax encodes a carboxyl-terminal ( ETEV amino acids 350–353 ) PDZ domain-Binding Motif ( PBM ) that mediates interactions with a particular group of cellular proteins containing one or several PDZ ( PSD95/DLG/ZO-1 ) domain ( s ) [7–9] . Many of these PDZ proteins are involved in processes that control cell attachment , cell proliferation , cell polarity and cell signaling [10 , 11] . Previous studies have indicated that the interaction of viral oncoproteins with PDZ proteins may play a critical role in the development of malignancies by perturbing the function of these cellular proteins [12 , 13] . The HTLV-1 Tax PBM has been shown to associate with several PDZ cellular proteins such as DLG1 ( Discs large 1 ) , Scribble , Erbin , TIP-1 ( Tax-interacting protein-1 ) or MAGI-3 ( Membrane-associated guanylate kinase-3 ) in in vitro studies [14–16] . One of them , Scribble that acts as a tumor suppressor and a regulator of cell polarity , is highly expressed in activated T-lymphocytes [17 , 18] . Interestingly , the absence of this motif in the Tax of HTLV-2 , a non-leukemic strain of HTLV , has led to the assumption that the HTLV-1 Tax PBM fulfills an essential function in the leukemic process [19 , 20] . Previous studies have shown that the deletion or mutation of the Tax PBM decreases IL2-independent growth of CTLL-2 cells and the Tax transforming activity in a rat fibroblast cell line [19] . More interestingly , Xie et al have reported that PBM is required for virus-mediated T-cell proliferation and genetic instability in vitro and for viral persistence in a rabbit infection model [21] . These observations strongly support the hypothesis that the Tax PBM is critically involved in supporting the infectious process , prompting us to in vivo evaluate the implication of the PBM in T-cell proliferation . To that aim , we used immunodeficient mice , which display a human hemato-lymphoid system , therein referred to as hu-mice . These hu-mice provide a powerful model for investigating the pathogenesis associated with infection by human lymphotropic viruses [22 , 23] . Several studies have previously demonstrated that infection of hu-mice with HTLV-1 recapitulates certain features of ATLL [24–26] . More specifically , our group has demonstrated that HTLV-1 is able to perturb early αβT-cell development in humanized BALB/c Rag2-/-γc-/- ( BRG ) mice [27] . We showed that HTLV-1 infection propelled thymic human T-cell development towards the mature stages and that this effect was dependent on Tax expression . In this study , we addressed the role of the Tax PBM in hu-mice infected with irradiated cells producing either a wild-type virus ( HTLV-1 WT ) or a virus characterized by a Tax PBM-deleted ( HTLV-1 ΔPBM ) . In the peripheral blood of WT hu-mice , the proliferation of activated CD4+CD25+ T-cells was significantly higher than in the peripheral blood of ΔPBM hu-mice . Likewise , human T-cells collected from WT hu-mice and cultivated in vitro in presence of interleukin-2 were proliferating at a higher level than those from ΔPBM animals . We then showed that the PDZ Scribble protein interacts with Tax in ex vivo or in vitro T-cells from WT hu-mice , but not in cells from ΔPBM hu-mice . These results underline that the PBM-PDZ association is critical for sustaining HTLV-1-induced T-cell proliferation . Finally , a genome-wide transcriptomic analysis of T-cells from infected hu-mice suggests that this association is involved in the regulation of host genes implicated in cell proliferation , inhibition of apoptosis , cell polarity and cytoskeletal changes .
To evaluate the role of the Tax PBM in vivo , a total of 32 hu-mice were used for this study . Thirteen hu-mice were inoculated with X-irradiated 293T cells previously transfected with either ACH-WT or ACH-ΔPBM molecular clones . Three hu-mice were inoculated with either X-irradiated 293T cells transfected with ACH-M22 ( that displays a Tax PBM , but is unable to activate the NF-κB pathway ) or untransfected ( mock infected ) . These hu-mice were daily monitored for apparent suffering signs , such as weight loss , back arches and prostrated behavior . Furthermore , a small volume of peripheral blood was eye-harvested from each infected hu-mouse every two weeks starting from one week post-infection ( Fig 1A ) and cytometry analysis was immediately performed to follow the presence of activated human CD25+ T-cells . Accordingly , 8 WT and 5 ΔPBM hu-mice with suffering signs between 3 and 5 weeks after infection were sacrificed . At 7 weeks , the 5 remaining WT hu-mice and 4 ΔPBM hu-mice that exhibited more than 10% of circulating CD25+ T-cells were sacrificed . The 4 surviving ΔPBM hu-mice that did not show any suffering signs were sacrificed two days later together with the 3 ACH M22 and the 3 mock infected animals ( Fig 1B ) . First , contrary to mock and M22 infected hu-mice , the percentage of circulating hu-CD3+ T-cells of WT and ΔPBM hu-mice increased gradually up to 7 weeks ( Fig 1C ) . This correlated with a significant increase of the frequency of hu-CD45 cells in both WT and ΔPBM infected hu-mice compared to the M22 and mock infected mice ( S3 Table ) . Among the hu-CD3+ T-cells , the frequency of activated CD4+CD25+ T-cells increased gradually up to 5 weeks after infection in the peripheral blood of both WT and ΔPBM hu-mice , while at 7 weeks this percentage was significantly higher in WT than in ΔPBM hu-mice ( Fig 1D ) . A low frequency of CD4+CD25+ T-cells was observed in the peripheral blood of ACH-M22 hu-mice as well as in that of mock infected hu-mice ( S3 Table ) . We did not observe a significant proliferation of the CD8+ T-cells in the peripheral blood of either group of mice ( S3 Table ) . It is important to note that the Tax transcriptional activity mediated by the WT and ΔPBM proviruses through both CREB/ATF and NF-κB signalling pathways is independent of the Tax PBM ( S1 Fig ) . In addition , it is evident that the PBM is operational only when the NF-κB pathway is functional . Collectively , these data suggest that the PBM is endowed with a sustaining activity of T-cell proliferation . Examination of sacrificed mice revealed that enlargement of the spleen was the most frequently observed pathological symptom . Splenomegaly was observed in all , but five WT and six ΔPBM hu-mice respectively ( S4 Table ) . Spleen was collected as well as bone marrow and when possible mesenteric lymph nodes . There was no significant difference in the spleen weight between groups ( mean of 0 . 255±0 . 199 g for WT vs 0 . 232±0 . 126 g for ΔPBM hu-mice compared to 0 . 103±0 . 037 g for mock and M22 infected hu-mice ) ( S4 Table ) . Sequence analysis of genomic DNA prepared from splenocytes of infected hu-mice confirmed the original sequence of the ACH-WT or ACH-ΔPBM molecular clones used for infection ( S5 Table ) . PVL of both WT and ΔPBM hu-mice splenocytes was between 0 . 1 to 1 copy/cell with no significant difference between groups ( Fig 2A; S4 Table ) . Using high throughput sequencing ( HTS ) -based mapping of HTLV-1 integration sites , we did not observe significant differences in the number of unique insertion sites ( UIS ) corresponding to the number of independent clones between WT and ΔPBM hu-mice ( S2 Fig ) . Similar levels of Tax mRNA and protein were detected in CD3+ T-cells from splenocytes of both WT and ΔPBM infected hu-mice ( Fig 2B and 2C and S3 Fig ) . There was no correlation between Tax mRNA levels and the weight of the spleens . As shown in Fig 3 and S6 Table , the number and the percentage of CD3+ T-cells collected from the spleen of WT and ΔPBM infected hu-mice were similar and higher than those observed in ACH M22 and mock infected hu-mice . Remarkably , a comparative analysis of the T-cell subpopulations ( DN: CD4-CD8-; DP: CD4+CD8+ , and SP: CD4+ or CD8+ ) among these CD3+ T cells in both WT and ΔPBM infected hu-mice revealed a low percentage of the DN T-cells and a high percentage of the SP T-cell population ( among which CD4+ T-cells dominated over CD8+ T-cells ) , in sharp contrast with the distribution profile of these subpopulations in ACH M22 and mock infected hu-mice ( Fig 3C and S4 Fig ) . Interestingly , even if similar degrees of splenomegaly ( S4 Table ) that correlated with similar numbers of CD3+ T-cells in the spleen ( Fig 3A ) were detected , it remains that the number as well as the percentage of CD4+CD25+ T-cells among human splenocytes were higher in WT than in ΔPBM hu-mice ( Fig 3B–3D ) . Likewise , lymph nodes and bone marrow collected from WT hu-mice showed a higher frequency of CD4+ CD25+ T-cells than those from ΔPBM-hu mice . As indicated above , such a difference also observed in the peripheral blood of animals sacrificed at 7 weeks suggests that the seeding of the periphery by CD4+CD25+ T-cells homing from lymphoid organs is more efficient for WT hu-mice than for ΔPBM hu-mice . Altogether , these results indicate that hu-mice are providing an appropriate environment for the proliferation of human T-cells infected with either HTLV-1 WT or ΔPBM proviruses . Overall , they underline that the HTLV-1 Tax PBM is acting as an in vivo auxiliary motif in the HTLV-1-induced proliferation of infected human T-cells . As introduced above , several studies have documented that the Tax PBM mediates interactions with a select group of PDZ-containing proteins [8 , 19 , 21 , 28] . In the present study , we focused our attention on one of them , the Scribble protein , that under physiological conditions is differentially localized throughout polarized T-cells and acts as a tumor suppressor [18] . Indeed , Scribble has been shown to undergo mislocalization in cultured HTLV-1 infected T-cells [14 , 15] . We first examined the interaction of Tax with Scribble and the localization of the two proteins in ex vivo splenocytes collected immediately from WT and ΔPBM infected hu-mice after their sacrifice , by using the in situ Proximity Ligation Assay ( PLA ) technology . PLA is a reliable readout of the molecular proximity of two endogenous proteins , thereby facilitating the direct observation of individual protein complexes in situ [29] . Thus , the presence of at least 4 dots per cell in about 60% of splenocytes from WT hu-mice clearly revealed a direct contact between Tax and Scribble ( Fig 4A , panel 1 and 4 ) . In contrast , we did not observe similar interactions in ΔPBM hu-mice ( Fig 4A , panel 2 ) . With regards to the intracellular localization of Scribble and Tax in these ex vivo splenocytes , immunofluorescent-staining ( IF ) assays clearly indicated that Scribble was preferentially detected in large polarized aggregates in the cytoplasm of cells from WT mice . In contrast , it was diffusely localized in the cytoplasm and at the plasma membrane of cells from ΔPBM hu-mice ( Fig 4B ) . Concomitantly , Tax was found to be mostly localized in the cytoplasm and also visible in condensed aggregates at the plasma membrane of cells from WT hu-mice ( Fig 4C ) . In contrast , Tax was detected in the nucleus and in the cytoplasm of cells from ΔPBM hu-mice . These data strongly suggest that PBM is associated with the sequestration of Scribble into polarized aggregates of T-cells from WT hu-mice . As mislocalization of Scribble might interfere with its tumor suppressor function , one can postulate that the Tax PBM is implicated in the enhanced T-cell proliferation observed in vivo in WT hu-mice . Such a possibility was further investigated in assaying the proliferation of T-cells collected from WT or ΔPBM hu-mice and in vitro seeded in growth medium supplemented with IL2 . We periodically verified that these T-cells contained integrated copies of the provirus used at infection ( S5 Table ) . Interestingly , the patterns of Tax/Scribble interaction ( Fig 5A ) and subcellular localization of both Tax and Scribble ( Fig 5B and 5C ) in these cultured human T-cells were identical to those observed in ex vivo splenocytes . We also observed that they expressed similar amounts of Tax ( Fig 6A ) . Likewise , periodic FACS analyses of both types of cells revealed the presence of a majority of CD25+ , GITR+ , CCR4+ and CADM-1+ T-cells ( Fig 6B ) . Interestingly , cell enumeration performed during several weeks showed that WT T-cells were actively proliferating , in contrast to ΔPBM T-cells that displayed a restrained growth ( Fig 6C ) . Overall , it is important to note that , contrary to ΔPBM T-cells , the proliferation of which was regularly in crisis , WT T-cells constantly proliferated and became immortalized . Taken together , these observations proposed that the Tax-PBM is enhancing HTLV-1-mediated T-cell proliferation and is necessary to immortalize T-cells isolated from hu-mice . It is well known that T-cells infected by HTLV-1 are forming large clumps when cultivated in vitro in presence of IL2 . We observed that the clumps of WT T-cells were regular and concentric whereas those of ΔPBM T-cells were irregular and eccentric ( Fig 6D ) . These observations are reminiscent of the FACS data showing differences in the size and granularity of ex vivo CD4+CD25+ T-cells isolated from the spleens of either WT hu-mice or ΔPBM hu-mice ( S5 Fig ) . Furthermore , the WT T-cells displayed long protrusions ( filopods ) of actin while ΔPBM T-cells showed shorter actin filopods , suggesting that PBM might be involved in cell migration ( Fig 6D ) . We also observed that in WT cells , the nuclei were oval ( ratio L/l = 1 . 8 ) while they were spherical in ΔPBM cells , suggesting that the cytoskeleton in WT cells exerts a distortion force on the nuclei ( Fig 6D and 6E ) . In summary , in vivo and in vitro data indicate that Tax PBM mislocalizes Scribble leading to morphological and cytoskeletal modifications that correlate with a sustained proliferation of WT T-cells . As PDZ proteins have been directly linked to the control of processes such as cytoskeletal organization , cell polarity and signal transduction pathways , we next investigated global transcriptional pathways that might be dysregulated by the PBM-PDZ interaction [17] . This was achieved by analyzing the transcriptome of cytoplasmic mRNA levels of both WT- and ΔPBM T-cells by RNA-sequencing ( RNA-seq ) , at 5 months in vitro culture . Differential gene expression analysis using the DESeq2 package ( adjusted P-value <0 . 01 ) resulted in the identification of 629 transcripts downregulated in WT T-cells , 503 of them displaying a fold change of at least 5 . 6 ( Log2 fold-change <-2 . 5 ) , whereas 400 transcripts were found to be upregulated , 337 of them displaying a fold change of at least 5 . 6 ( Log2 fold-change >2 . 5 ) compared to the ΔPBM T-cells ( Fig 7A , 7B and 7C ) . We looked for the transcriptional expression of genes coding for the PDZ proteins known to be involved in T-cell homeostasis , such as Scribble , MAGI-1 , MAGI-3 and DLG1 . We did not observe a significant difference in their expression levels , indicating that expression of these PDZ proteins is PBM-independent . In addition , the number of HTLV-1-related reads was identical under both conditions suggesting that the viral expression was not impaired in ΔPBM cells . Finally , we performed a gene ontology analysis of differentially expressed mRNAs ( adjusted P-value <0 . 01; log2 fold-change of 2 . 5 ) . Among the genes upregulated in WT cells we identified genes involved in cell proliferation such as IL9 ( fold change of 57 ) and cell activation such as LCK ( fold change of 172 ) ( Figs 7D and S6B ) . In contrast , genes downregulated in WT T-cells consisted of genes involved in inhibition of cell proliferation such as CD9 ( fold change of 129 ) and in apoptotic processes such as RHOB ( fold change of 27 ) . Furthermore , genes related to cytoskeleton organization were also identified as dependent on the Tax/PDZ interactions , some of them upregulated such as CDC42BPA while others showed decreased expression such as FLNB . Interestingly , the expression of class I regulatory PIK3R6 and PIK3CD subunits was upregulated in WT cells ( fold change of 32 and 5 . 7 respectively ) . These proteins are implicated in the activation of the Akt/mTOR pathway , involved in cell proliferation and survival . We also identified a gene involved in the non-canonical Wnt pathway ( WNT5B; increased expression; fold change of 16 . 7 ) and two genes associated with the canonical Wnt pathway ( WNT2 and WNT1; decreased expression; fold change of 16 . 7 and 147 respectively . This indicates that both canonical and non-canonical Wnt pathways might be modified by the interaction of Tax with PDZ proteins . In conclusion , these results suggest that the Tax PBM is involved in the transcriptional regulation of multiple genes implicated in T-cell proliferation , in the inhibition of apoptosis , as well as in cell polarity , in cytoskeletal and in morphological changes . Taken together , these in vivo and in vitro findings underline that PBM/PDZ recognition may be required for sustaining HTLV-1 mediated T-cell proliferation , for inducing cell polarity and cytoskeletal modifications and for triggering the immortalization of T-cells in hu-mice .
The generation of hu-mouse models , capable of multi-lineage human hematopoiesis has paved the way for the in vivo study of infection by human specific pathogens . Several human viruses and among them lymphotropic viruses have been extensively used in these models [22 , 23 , 30] . Thus , hu-mice have been used to approach the pathogenic activity of HTLV-1 Tax in a more biological model than cultured cells [25–27 , 31] . Here we have investigated the role played by the Tax PBM in the lymphoproliferation triggered by HTLV-1 infection of hu-mice . To achieve this objective , we validated a new procedure to infect these animals with cloned proviruses . Thus , hu-mice were infected either by WT virus or by ΔPBM virus carrying a PBM-deleted genome , both produced after transfection of 293T cells with the corresponding provirus . After examining the response of hu-mice to infection by WT or ΔPBM HTLV-1 , we have characterized T-cells either freshly isolated from the spleen of these infected hu-mice or in vitro cultured . We observed an increased number and frequency of activated CD4+CD25+ T-cells in the peripheral blood of WT HTLV-1 hu-mice , albeit at a lower level in ΔPBM HTLV-1 infected mice . Splenomegaly , which was observed in both infected hu-mice , is mainly caused by a similar accumulation of CD3+ T-cells . But , once again , the number and frequency of CD4+CD25+ T-cells is significantly higher in WT than in ΔPBM hu-mice . Proviral loads and clonality in both types were similar , indicating that the PBM does not impact any of these parameters . Furthermore , we report that T-cell proliferation was severely impaired in hu-mice infected with the ACH-M22 provirus that carries the PBM , but unable to activate the NF-κB pathway . These data indicate for the first time that in vivo the PBM alone is unable to induce HTLV-1-mediated proliferation , but is only able to sustain the NF-κB-mediated proliferation . We next analyzed the proliferation of cells collected from the spleen of infected hu-mice and in vitro cultivated in presence of IL2 . Short-term assays underlined that WT T-cells constantly proliferated over at least more than one year and were therefore considered as immortalized . In contrast , ΔPBM T-cells showed a restrained growth and experienced several death crisis suggesting that they were not immortalized . These data provided from infected hu-mice underline that the Tax-PBM is enhancing HTLV-1-mediated T-cell proliferation and is required for T-cell immortalization . A previous study has reported that the Tax PBM significantly increased HTLV-1-induced T-cell proliferation after in vitro cocultivation of human PBMCs ( Peripheral Blood Mononuclear Cells ) with irradiated cell lines producing WT or ΔPBM HTLV-1 [21] . These authors report that Tax PBM promotes HTLV-1-induced proliferation of human PBMCs , but that it is not required for virus-mediated immortalization of these cells , as they did not detect any difference in immortalization potential between WT Tax and ΔPBM Tax . As this conclusion concerning the implication of the Tax PBM in immortalization is differing from ours , it is tempting to speculate that such a difference may be related to the specific experimental approach of each study . Previous studies have shown that the Tax PBM induced the mislocalization of Scribble [14 , 15] . However , most if not all of the reported observations have been obtained through the use of transfected cell lines or HTLV-1 T-cell lines . Our study was performed with T-cells from hu-mice clearly indicates that the PBM/PDZ interactions are involved in the distribution pattern of Tax . Among the PDZ proteins , we focused our attention on the Scribble protein known to interact with Tax and to be involved in cell polarity , in T-cell development and proliferation . Based on the results of IF and PLA assays , a comparative analysis of ex vivo ( immediately after their collection from the spleen ) and in vitro ( cultured in the presence of IL2 ) WT and ΔPBM T-cells underline that the Tax PBM plays a prominent role in the re-localization of both proteins . In WT cells , both Tax and Scribble were observed in large aggregates , mainly at the cell membrane . In contrast , such aggregates were not detected in ΔPBM cells , confirming the ability of Tax PBM to sequester Scribble . The mislocalization of Scribble upon binding to the Tax PBM may be linked to morphological changes that differentially affect the actin cytoskeleton and the polarity of these WT and ΔPBM cells . In addition , the binding of the Tax PBM to Scribble may be responsible for the sustained proliferation of WT infected T-cells by negatively interfering with the tumor suppressor property of that PDZ protein . In contrast , the lack of interaction of Tax with Scribble may result in the decreased proliferation of ΔPBM T-cells . Further studies in hu-mice will aim at characterizing other PDZ proteins that interact with the Tax PBM , such as DLG-1 , to unravel the possible link between these interactions and T-cell proliferation . It will also be of interest to test whether the acetylation of Tax lysine K10 ( amino acid 346 ) located immediately upstream of the PBM could have an impact on the PBM/PDZ interactions , and finally on the localization of Tax and its function in T-cell proliferation . It has been demonstrated that overexpression of Scribble attenuated NFAT reporter activity in anti-CD3/anti-CD28-stimulated Jurkat cells . By interacting with Scribble , Tax could counteract this negative effect on NFAT activation and thus stimulate T-cell proliferation [14] . Consequently Tax PBM association with PDZ proteins represents an essential event during the development and maintenance of the lymphoproliferative process . Moreover , this association appears to be linked to the exclusive HTLV-1-induced genetic instability observed in human PBMCs infected with WT-HTLV-1 [21] . In that context , to further explore the differences between cells expressing either Tax WT or Tax ΔPBM , we performed a comparative transcriptomic analysis that enabled the identification of a set of genes that are differentially expressed in either type of cells . Genes coding for PDZ proteins normally expressed in T lymphocytes such as Scribble , MAGI-I , MAGI-3 and DLG1 were found to be similarly expressed in WT and ΔPBM T-cells . In contrast , we noticed the deregulated transcription of genes involved in T-cell signaling and proliferation , in apoptosis induction and in cytoskeleton organization . The expression of PIK3 subunits retained our attention as these kinases activate Akt [32 , 33] . We observed that the expression of PIK3 subunits is significantly higher in WT cells than in ΔPBM cells , suggesting a direct effect on Akt activation . It has been reported that Tax by binding to the PDZ DLG-1 protein counteracts the negative effects of the PTEN and PHLPP phosphatases and thus activates Akt [16] . Further studies will be needed to determine whether Scribble by interacting with Tax will have such an effect on Akt activation . Finally , results from the genome wide transcriptome analysis , by supporting our in vivo and in vitro observations , propose that the Tax PBM plays an eminent role in the pathogenicity of HTLV-1 . Further work is needed to establish that such an activity is dependent on PBM-PDZ interactions and to precisely determine which PDZ protein deregulates which set of genes . As stated in the Introduction , the presence of the PBM was identified not only in Tax but also in several viral oncoproteins such as in the protein E4-ORF1 of adenovirus type 9 and in E6 proteins of several human papilloma viruses [7 , 34] . It is worth noting that the presence of a PBM is linked to the ability of HPV-16 and HPV-18 to induce malignant tumors and that such a motif is absent in HPV subtypes that induce benign tumors ( for example HPV-9 or-11 ) [35] . Our data converge to a similar conclusion and stress that the PBM/PDZ recognition is perturbing the regulation of processes such as cytoskeletal organization , cell polarity , cell proliferation . Consequently , the PBM represents a Tax domain endowed with an auxiliary activity essential in the induction and the maintenance of the HTLV-1-induced T-cell proliferation leading to a malignant proliferation . Thus targeting the PBM/PDZ nodes by small peptides is offering a novel strategy to slowdown the T-cell proliferation in HTLV-1 infected hu-mice [36] .
The human embryonic kidney ( HEK ) 293T cells ( American Type Culture Collection CRL-3216 ) were grown in Dulbecco’s modified Eagle’s medium ( DMEM ) supplemented with 10% fetal calf serum ( FCS ) ( Sigma-Aldrich , France ) , 50 μg/ml streptomycin , 100 U/ml penicillin ( Invitrogen , France ) . The ACH plasmid is an infectious molecular clone of HTLV-1 [37] . The HTLV-1 provirus deleted from the PBM of Tax ( ACH-ΔPBM ) was constructed by introducing a TAA stop codon instead of the GAA codon in the Tax C-terminus resulting in loss of the last four amino-acids of Tax ( ETEV: consensus PBM ) . This mutation was confirmed by DNA sequencing . The ACH-M22 plasmid encoding for a mutated Tax protein that do not activate the NF-κB pathway is a kind gift of Dr . F . Bex ( Belgium ) . 293T cells were plated at 5x105 cells in a 6-well plate the day prior to transfection . Plasmid DNA ( 3 . 3 μg ) was applied to the cells as calcium phosphate coprecipitates . Medium was changed 6h after transfection . One day later , supernatants were collected and analyzed for HTLV p19 antigen content by using the Retrotek ELISA-kit ( ZeptoMetrix Corp . , USA ) . Anonymized human umbilical cord samples from the Maternity Ward of Hôpital Femme-Mère-Enfant ( Bron , France ) were obtained from healthy full-term newborns with written parental informed consent according to the guidelines of the medical and ethical committees of Hospices Civils de Lyon and of Agence de la Biomédecine , Paris , France . Experiments using cord blood were approved by both committees and were performed in full compliance with French law . Animal experimentation was performed in strict accordance with the French “Comité National de Réflexion Ethique sur l’Expérimentation Animale , n°15” and the ethical guidelines for the care of these mice of the Plateau de Biologie Expérimentale de la Souris ( PBES , UMS 3444 ) at Ecole Normale Supérieure ( ENS ) de Lyon . This protocol has been approved by the Committee on the Ethics of Animal Experiments of ENS de Lyon ( approval number: ENS_2014_043 ) . All efforts were made to minimize animal suffering . After density gradient centrifugation of human cord blood , CD34+ cells were enriched twice using immunomagnetic beads according to the manufacturer instructions ( CD34+ MicroBead Kit , Miltenyi Biotec , Bergisch-Gladbach , Germany ) . Purity ( ≥ 95% ) was evaluated by FACS analysis using human PE-CD34 antibody ( Miltenyi Biotec ) . Cells were frozen before the transplantion when newborn mice were available . NOD . Cg-Prkdcscid Il2rγtm1Wjl Tg ( HLA . A2 . 1 ) 1Enge/SzJ ( NSG-HLA-A2/HDD ) were obtained from the Jackson Laboratory ( Bar Harbor , USA ) and bred and maintained under pathogen-free conditions at the PBES . Newborn males and females NSG mice ( 2 to 5 days old ) were sub-lethally irradiated with 1 . 1 Gray ( 320 kV , 12 . 5 mA ) from a X-ray irradiator ( XRad-320 , PXI Precision XRay , France ) and intra-hepatically injected with 2x105 human CD34+ hematopoietic stem cells isolated from cord blood samples [38] . After 10 weeks , humanized mice ( ≥ 30% hu-CD45+ cells in peripheral blood ) were infected by HTLV-1 . Infection and mice monitoring were performed in a Biosafety Level 3 Laboratory in accordance with the PBES guidelines . Lethally irradiated 293T cells ( 50 Gray , 320kV , 12 . 5 mA ) transfected with full length or truncated HTLV-1 molecular clone were intra-peritoneally injected: the amount of irradiated cells to inject per mouse corresponds to the number of cells producing 70 ng of p19 in 24h-culture . Mock infected mice were injected with the same amount of irradiated non-transfected 293T . Hu-mice were daily monitored for signs of obvious suffering , such as weight loss , back arches and prostrated behavior . Peripheral blood was collected from the retro-orbital venous sinus under Isoflurane anesthesia . When mice were either suffering or displaying more than 10% circulating CD25+ T-cells , they were sacrificed after anesthesia . Tissue specimens ( spleen , mesenteric lymph node and tibia bone ) were collected and gently minced in PBS to obtain a single-cell suspension and immediately frozen in FCS containing 10% DMSO and kept at -80°C . Spleens from infected hu-mice were harvested and analyzed at indicated time points following infection . To obtain a single-cell suspension , spleens were minced and passed through a nylon mesh . Red cell lysis was performed in red cell lysis buffer ( Sigma , France ) for 10 min . Cells were then washed and enumerated . For flow cytometry , single-cell suspensions were stained with the appropriate monoclonal antibody ( S1 Table ) or the respective isotype control antibody for 30 min in the dark at 4°C . Human lymphocytes first gated as hCD45+ cells were then defined as CD3+ T-cells containing the following subsets: DN , CD4-CD8-; DP , CD4+CD8+; and SP , CD4+ CD8- or CD4-CD8+ . Absolute numbers of cells were determined by multiplying the number of nucleated cells by the percentage of positive cells for the indicated cell surface marker ( s ) . For CADM-1 expression analysis , cells were stained with the primary antibody ( Rabbit polyclonal antibody H-300 Santa Cruz Biotechnology , USA ) for 2h at 4°C in PBS containing 5% FCS then washed twice and incubated 30 min at 4°C in the dark with a secondary fluorescent antibody anti-rabbit ( A-11008 Molecular Probes , France ) . After two wash steps in FACS-buffer , fluorescence was measured on a flow cytometer ( FACSCanto II , BD , San Jose , CA , USA ) . Cells were always gated to exclude doublet . Compensations were realized using Miltenyi MACS Comp Beads . Data were evaluated with BD Diva software ( Becton Dickinson Immunocytometry Systems , Mountain View , CA ) and analyzed using FlowJo software ( Treestar , Ashland , USA ) . Results are expressed as the mean of % positivity of surface expression ± SEM or as the Mean Fluorescence Intensity ( MFI ) . To obtain cell lines derived from infected hu-mice , single cell suspensions were cultured in complete RPMI 1640 medium supplemented with 10% FCS ( Sigma-Aldrich , France ) . Recombinant IL2 ( 20 U/ml ) ( Peprotech , France ) was added to the cultures every 3 days . After one month of culture , selected cell lines were tested for the relevant proviral sequence and weekly monitored for their proliferation . Spleen samples were fixed with 4% paraformaldehyde in PBS , embedded in paraffin , sectioned and stained with H&E solution . An indirect immunoperoxidase technique with commercially available monoclonal antibodies to CD3 and rabbit polyclonal antibody to Tax ( kind gift of Dr . B . Cullen ) was applied to the tissue sections as previously described [27] . Pictures were analyzed with ImageJ software . Slides or ibiTreat μ-dishes ( IBIDI ) were pre-coated with poly-L-lysine ( Sigma-Aldrich , France ) 20 μg/ml for IF and 10 μg/ml for PLA . To detect Tax and Scribble proteins , T-cells isolated from spleens were added to the slides before being fixed with 3% paraformaldehyde in PBS for 10 min at room temperature . Cells were then permeabilized with TritonX-100 or methanol at -20°C . For IF staining , cells were blocked with 5% FCS then stained with primary and secondary antibodies and mounting medium with DAPI ( Duolink , Sigma , France ) . Polyclonal goat antibodies to Scribble ( C-20; Santa Cruz Biotechnology ) , rabbit antibodies to Tax ( kind gift of Dr . B . Cullen ) and appropriate controls were used . Phalloidin , fluorescein isothyocyanate labeled ( Sigma-Aldrich ) interacts with polymeric actin . PLA was carried out with Duolink In Situ-Fluorescence Red kit ( Sigma-Aldrich ) according to the manufacturer’s instructions . Negative controls were performed on T-cells in the absence of antibodies to Scribble . Microscopic examination was performed using a Zeiss LSM710 confocal microscope ( Carl Zeiss Jena , inc , Germany ) or an Axioimager Z1 epifluorescence microscope ( Carl Zeiss Jena , inc , Germany ) . Images were analyzed using ImageJ software . Statistical tests were performed using GraphPad PRISM software . When n≤ 5 in one or both groups , they are one-tailed and the non-parametric Wilcoxon-Mann-Whitney test was performed . When n> 5 in one or both groups , the parametric Student t-test was performed if variance are equal ( F-test with P-value > 0 . 05 ) . If not ( F-test with P-value < 0 . 05 ) , the non-parametric Wilcoxon-Mann-Whitney U test was performed . Statistical analysis of hu-CD3 subpopulations was performed with one-way ANOVA test . The results were considered statistically significant when P-value < 0 . 05 . Cells were lysed with sample buffer ( 10 mM HEPES , pH 7 . 9 , 500 mM NaCl , 3 mM MgCl2 , 1 mM DTT , 1 mM PMSF , and 0 . 5% Triton X-100 supplemented with protease inhibitors ) . After incubation on ice for 60 min , whole cell lysates were centrifuged at 15 , 000 g for 10 min at 4°C to remove the debris . Protein concentration of the cleared lysates was determined using the Bradford assay . Cell lysates ( 15μg ) were size-separated by electrophoresis on a 12% SDS-polyacrylamide gel ( 3h migration at 20 mA ) and transferred onto PVDF membranes . The blot was blocked in PBS-5% milk and incubated with anti-Tax antibodies ( 1:1 , 000; kind gift of B . Cullen ) , anti-β-actin ( 1:5 , 000 ) obtained from Sigma-Aldrich ( clone AC-15 ) . After several washes , horseradish peroxidase ( HRP ) -conjugated anti-rabbit IgG ( 1:10 , 000; Cell Signaling , The Netherlands ) or HRP-linked anti-mouse IgG ( 1:10 , 000; GE Healthcare , France ) were added to the membranes which were washed again several times and subsequently incubated with the Western Lightning ECL solution ( Thermo Fisher Scientific , France ) . Images were captured using a ChemiDoc Imaging system ( Biorad , France ) . Genomic DNA was extracted from the single cell suspension using the Nucleospin Blood kit ( Macherey-Nagel , Düren , Germany ) according to the manufacturer’s instructions . PVL was measured by qPCR with HTLV-1 tax-specific primers . The PVL was calculated as previously described [27] and expressed as the number of pX copies per 100 human cells . RNA was extracted from the single cell suspension using RNAzol RT ( Sigma Aldrich , France ) according to the manufacturer’s instructions , and resuspended in 10 μl of RNAse-free water and treated with 10 U of RNase-free DNase I ( Qiagen , Hilden , Germany ) for 15 min at 30°C and then for 15 min at 60°C . 500 ng of total RNA were then retro-transcribed at 42°C during 50 min in a total volume of 20 μl reaction buffer containing 100 U of SuperScript II reverse transcriptase ( RT; Invitrogen , CA , USA ) . A reaction without RT was performed as a control for genomic DNA contamination . The mRNA levels were normalized using 3 different housekeeping genes ( ACTB , RSP11 and RSP14 ) chosen to be the most stable in our model with BestKeeper and NormFinder algorithms . Quantitative real-time PCR ( qPCR ) was performed using the FastStart Universal SYBR Green Master ( Roche , Mannheim , Germany ) on a StepOnePlus system ( Applied Biosystem , CA , USA ) . The initial denaturation step at 95°C for 10 min was followed by 40 cycles with one cycle consisting of 10s at 95°C , 30s at 60°C , and 15s at 72°C . The nucleotide sequences of the primers were used for RT-PCR , proviral load measurement and DNA sequencing of the mutation of Tax gene are shown in S2 Table . To determine the number and abundance of HTLV-1 infected clones in humanized mice , we used an improved quantitative HTS method to map the proviral integration sites in the human genome [39] . Libraries were prepared starting from 500 ng DNA and sequenced on an Illumina MiSeq instrument . 150 bp paired-end reads were acquired and sequencing reads that supported either the 5’ or the 3’ LTR-host junctions were retained . The number of unique integration sites ( UIS ) was determined as previously described [40] . T lymphocytes isolated from the spleen of WT or ΔPBM infected hu-mice were cultured in complete RPMI medium containing IL2 . Cells ( 5x106 ) were then washed with ice-cold PBS , centrifuged at 500 g for 5 min at 4°C and lysed in 1 ml of lysis buffer ( 10mM Tris-HCl pH7 . 5 , 5mM MgCl2 , 100mM KCl , 2mM DTT , protease inhibitor EDTA-free ( Roche , Mannheim , Germany ) , 1% Triton X-100 ) . Lysates were gently homogenized and incubated at 4°C for 10 min , centrifuged at 1 , 300 g for 10 min at 4°C and the supernatant was recovered . Total RNA was extracted from cellular extracts using Trizol and subjected to cDNA library construction using the smartseq2 protocol [41] . Reads were first split with respect to their 5′-barcode sequence . After this , 5′-barcode and 3′-adaptor sequences were removed from reads . Reads were then aligned to a custom set of sequences corresponding to ribosomal RNA and tRNA sequences using Bowtie [42] in order to remove contaminants . Remaining reads were aligned to the human genome and transcriptome ( hg19 assembly ) using TopHat2 [43] . The alignment files obtained from TopHat2 were used to count reads mapping to the 5’UTR coding sequence and 3’UTR of human transcripts using HTSeq [44] and the UCSC hg19 gene annotation file . Differential gene expression was performed using the R package DESeq2 [45] . Only genes with an adjusted P-value ≤ 0 . 01 and a log2 foldchange superior to 2 . 5 were selected . GeneOntology was done using the GeneRanker tool of Genomatix software . | The viral Tax oncoprotein is a critical contributor to the development of adult T-cell leukemia/lymphoma , an aggressive malignant proliferation of T lymphocytes . Tax contains a PDZ domain-binding motif ( PBM ) that favors the interaction with several cellular PDZ proteins . Here , we compare the in vivo involvement of the Tax PBM in humanized mice infected with either a full-length provirus or a Tax PBM-deleted provirus . We observe that the establishment of the sustained lymphoproliferation in the peripheral blood of infected mice is dependent on the Tax PBM . Furthermore , binding of the Tax PBM to the PDZ Scribble protein correlated with perturbations of cytoskeletal organization and cell polarity . In addition , genome-wide transcriptomic analyses strongly suggest that the association of Tax PBM with cellular PDZ proteins results in the expression of several genes involved in proliferation , apoptosis and cytoskeletal organization . Collectively , these results indicate that the Tax PBM is an auxiliary motif that contributes to the growth of HTLV-1 infected T-cells . As a consequence , targeting the PBM/PDZ nodes using small peptides may have the potential to antagonize the Tax-induced lymphoproliferation , offering a novel strategy for the treatment of this disease . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"blood",
"cells",
"cell",
"physiology",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"immune",
"physiology",
"pathology",
"and",
"laboratory",
"medicine",
"body",
"fluids",
"spleen",
"pathogens",
"293t",
"cells",
"immunology",
"biological",
"cultures",
"microbiology",
"rna",
"extraction",
"retroviruses",
"viruses",
"cell",
"polarity",
"rna",
"viruses",
"cellular",
"structures",
"and",
"organelles",
"extraction",
"techniques",
"cytoskeleton",
"research",
"and",
"analysis",
"methods",
"white",
"blood",
"cells",
"animal",
"cells",
"medical",
"microbiology",
"htlv-1",
"t",
"cells",
"microbial",
"pathogens",
"cell",
"lines",
"blood",
"cell",
"biology",
"anatomy",
"viral",
"pathogens",
"physiology",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"organisms"
] | 2018 | PDZ domain-binding motif of Tax sustains T-cell proliferation in HTLV-1-infected humanized mice |
Genome-wide association study ( GWAS ) entails examining a large number of single nucleotide polymorphisms ( SNPs ) in a limited sample with hundreds of individuals , implying a variable selection problem in the high dimensional dataset . Although many single-locus GWAS approaches under polygenic background and population structure controls have been widely used , some significant loci fail to be detected . In this study , we used an iterative modified-sure independence screening ( ISIS ) approach in reducing the number of SNPs to a moderate size . Expectation-Maximization ( EM ) -Bayesian least absolute shrinkage and selection operator ( BLASSO ) was used to estimate all the selected SNP effects for true quantitative trait nucleotide ( QTN ) detection . This method is referred to as ISIS EM-BLASSO algorithm . Monte Carlo simulation studies validated the new method , which has the highest empirical power in QTN detection and the highest accuracy in QTN effect estimation , and it is the fastest , as compared with efficient mixed-model association ( EMMA ) , smoothly clipped absolute deviation ( SCAD ) , fixed and random model circulating probability unification ( FarmCPU ) , and multi-locus random-SNP-effect mixed linear model ( mrMLM ) . To further demonstrate the new method , six flowering time traits in Arabidopsis thaliana were re-analyzed by four methods ( New method , EMMA , FarmCPU , and mrMLM ) . As a result , the new method identified most previously reported genes . Therefore , the new method is a good alternative for multi-locus GWAS .
Genome-wide association study ( GWAS ) focuses on associations between single nucleotide polymorphism ( SNP ) and traits of interest in order to investigate the genetic foundation of these traits [1 , 2] . In GWAS , hundreds of thousands of SNPs are genotyped for several hundreds of individuals . In this case , statistical estimation and detection of the relationship between these SNPs and the traits become challenging . Although the single variant analysis in standard GWAS methods has succeeded in identifying thousands of genetic variants associated with hundreds of various traits , this approach fails to consider the joint effect of multiple genetic markers on traits . Another problem with this approach is the issue of multiple test correction for the threshold value of significance test . The Bonferroni correction is too stringent , and many relevant loci are missed out . In genetics , only a small subset of SNPs is associated with the phenotype of a trait . This is an example of a variable selection problem for high-dimensional data , where the number of SNPs ( p ) is several times larger than the number of individuals ( n ) [3] . To solve this issue , many penalization methods have been developed in statistics , for example , bridge [4] , nonnegative garotte [5] , least absolute shrinkage and selection operator ( LASSO ) [6] , smoothly clipped absolute deviation ( SCAD ) [7] , elastic net [8] , fused LASSO [9] , adaptive LASSO [10] and minimax concave penalty [11] . Among these methods , Bayesian LASSO [12] , penalized logistic regression [13] , sure independence screening [14] , adaptive mixed LASSO [15] , elastic net [16] , LASSO [17] , empirical Bayes [18] and empirical Bayes LASSO [19] have been adopted in GWAS . These methods are multi-locus in nature , hence a less stringent significance criterion can be adopted [20] . Despite these methods being able to shrink some markers to zero , they will fail when the number of markers is several times larger than the sample size . In this case , the solution lies in reducing the number of markers before employing a shrinkage method in the multi-locus genetic model . For example , a Bayesian sparse linear mixed model [21] and Bayesian mixture models [22] . However , the computing time becomes a major concern for these Bayesian approaches . An alternative is to integrate single-marker scanning with the multi-locus models , such as a model-free approach [23] , multi-locus random-SNP-effect mixed linear model ( mrMLM ) [20] , and fixed and random model circulating probability unification ( FarmCPU ) [24] . In this study , we developed an approach that reduced the number of markers , p , via correlation learning ( i . e . Iterative modified-Sure Independence Screening ) to a moderate number . A moderate-scale variable selection method , SCAD , was then employed to select variables in the reduced model . We chose SCAD because of its nice oracle property . Conditional on the selected variables , we repeated the screening procedure and chose another set of variables . All the effects of the above-selected variables were estimated by Expectation-Maximization ( EM ) -Bayesian LASSO algorithm [25] and tested by likelihood ratio statistic for true quantitative trait nucleotide ( QTN ) detection . We call this approach Iterative modified-Sure Independence Screening EM-Bayesian LASSO ( ISIS EM-BLASSO ) algorithm . A series of simulated and real datasets were used to validate this new method . We compared our new method with single-locus methods: efficient mixed-model association ( EMMA ) [26] and FarmCPU [24] , multi-locus methods: SCAD [7] and mrMLM [20] . Several reasons guided the choice of these comparison methods: EMMA [26] has been a standard gold method for GWAS , FarmCPU [24] reduces the number of markers used in GWAS just like ISIS EM-BLASSO , SCAD [7] is used in the screening method of ISIS EM-BLASSO hence the need to compare it independently and lastly , mrMLM [20] integrates single locus with multi-locus approach .
Three Monte-Carlo simulation experiments were carried out to measure the effectiveness of our new method . We used statistical power to evaluate the effectiveness of ISIS EM-BLASSO method alongside the three methods for comparison purposes . For each QTN , we defined its statistical power as the fraction of the samples in which the QTN was detected ( see Methods for significant testing ) . Fig 1A , 1B and 1C represents the results of the statistical power of detecting each QTN from the three simulation experiments respectively . In the first simulation experiment in which no polygenic variance was simulated , ISIS EM-BLASSO method has the highest power for detecting almost all the six simulated QTNs except QTN two . mrMLM has a high power of detecting the second QTN . EMMA and FarmCPU have a low power of detecting the second and fourth simulated QTNs ( Fig 1A and S1 Table ) . Indeed , Bonferroni correction is too stringent , and it may cause many significant loci to be missed out . SCAD has a moderately higher power than EMMA and FarmCPU for the second , third and sixth QTNs . SCAD lacks consistency in detecting the simulated QTNs hence it cannot be relied upon especially when the QTN size is small . The same trends were observed in the second simulation experiment ( Fig 1B and S2 Table ) when an additive polygenic variance was added to the polygenic background . In the third simulation experiment ( Fig 1C and S3 Table ) where three pairs of epistatic effects ( collectively contributing 0 . 15 to the phenotypic variance ) were added to the genetic background , ISIS EM-BLASSO is still powerful in detecting almost all the six simulated QTNs . We presented a paired t- test for the differences in statistical power ( Table 1 ) . We observe that there are significance differences ( at the 0 . 05 level ) in statistical power between ISIS EM-BLASSO and the other three methods ( SCAD , EMMA , FarmCPU ) . There are no significance differences in statistical power between ISIS EM-BLASSO and mrMLM except in the third simulation ( at the 0 . 1 level ) . Based on these findings , it implies that the simulated QTNs are mostly likely to be identified when ISIS EM-BLASSO method is used . Mean squared error ( MSE ) was used to measure the accuracy of each estimated QTN effect for all the methods . We evaluated the accuracies of all the six simulated QTNs effects in the three simulation experiments . The results of all methods considered are presented in Fig 2 and S1 , S2 and S3 Tables . The ISIS EM-BLASSO method is consistently more accurate in estimating the QTN effects than the other methods ( EMMA , SCAD , and FarmCPU ) . From these results , EMMA has the highest MSEs for each of six simulated QTNs , implying it is inaccurate in estimating the QTN effect . ISIS EM-BLASSO has lower MSEs than mrMLM for simulated QTNs 4 , 5 and 6 in all the three simulations . This implies that it is reliable for estimating the QTN effects . At the 0 . 05 significance level , the differences of MSEs between ISIS EM-BLASSO and other methods ( EMMA and SCAD ) were significant ( Table 1 ) . Applying EM-Bayesian LASSO to SNPs selected from the iterative procedure will not only remove unimportant SNPs but also improves the effect estimation . Type 1 errors for all the methods in the three simulation experiments were calculated ( Fig 3 ) . The values for Type 1 errors for each method in each simulation are recorded in S1 , S2 and S3 Tables . Despite having high power in the detection of QTNs , ISIS EM-BLASSO had slightly higher Type 1 errors compared with SCAD , EMMA , FarmCPU , and mrMLM . Note that all the Type 1 errors were less than 0 . 04% . Bonferroni correction eliminates many un-associated loci hence reducing the false positive rate in FarmCPU and EMMA at the expense of some associated SNPs . Conversely , this study reveals that ISIS EM-BLASSO method may slightly include some un-important SNPs in the model than SCAD , EMMA , FarmCPU and mrMLM though with a higher power in detecting associated QTNs . Receiver operating characteristic ( ROC ) curve is used to compare different methods for their efficiencies in the detection of significant effects . ROC curve is a plot of the statistical power against the controlled Type 1 error . A method with the highest ROC curve is deemed the best . We simulated various 67 probability levels of significance between 1e-8 to 1e-2 , with these values we calculated the corresponding powers in the first simulation experiment . Fig 4 shows a comparison of the ROC curves from the four methods for the first experiment for each of the six QTNs simulated . ISIS EM-BLASSO performs best among all the other methods considered for all the simulated QTNs . Comparing the computing times ( Fig 5 and S1 , S2 and S3 Tables ) of these methods in the three simulation experiments respectively , we observed that ISIS EM-BLASSO has the lowest computing time whereas EMMA takes a longer computing time ( Intel Core i5-4570 CPU 3 . 20GHz , Memory 7 . 88G ) . ISIS EM-BLASSO is computationally efficient and can be used in GWAS in a few hours to obtain the associated genes . ISIS EM-BLASSO in itself reduces the number of SNPs to only those that are significantly correlated with the phenotype and hence this reduces the problem to a moderate high dimensional data setting problem saving on the computational time . Six Arabidopsis flowering time traits in Atwell et al . [27] have been re-analyzed by ISIS EM-BLASSO , EMMA , FarmCPU , and mrMLM . These traits are LD , LDV , SD , 0W , 2W , and 4W . ISIS EM-BLASSO detected 14 , 11 , 23 , 21 , 9 and 11 SNPs to be significantly associated respectively with the six traits above . These detected SNPs for each trait were used to conduct a multiple linear regression analysis , and the corresponding AIC and BIC were calculated . The ISIS EM-BLASSO method showed low AIC and BIC values for nearly all traits ( S4 Table ) . The only method that compares almost equally with ISIS EM-BLASSO is mrMLM . This indicates that SNPs detected by ISIS EM-BLASSO fit the data better than the other methods . The numbers of known genes in the proximity of SNPs for the above six traits were in total 67 genes from ISIS EM-BLASSO , 22 genes from mrMLM , 15 genes from FarmCPU , and 13 genes from EMMA . ISIS EM-BLASSO detected more known genes than the other methods ( S5 Table ) . Among these known genes , 50 were identified only by ISIS EM-BLASSO ( Table 2 ) . It is interesting to note that for the trait SD , EMMA was not able to determine any significant gene whereas the new approach identified 21 genes . A similar trend is also observed when we considered trait 0W where FarmCPU did not detect any gene . Based on these results , we observe that the new approach can capture the genes associated with the trait under study . We also noted that some genes tested were found significant in nearly all the traits considered . For example , gene DOGI was discovered to be associated with LDV , SD , 0W , and 4W , gene SVP was found to be related to LD and LDV , gene ETC3 was found to be associated with LD , 0W , and 2W , and gene ABCB19 was discovered to be related to SD , 2W , and 4W . These results are consistent with previous studies related to these traits as outlined in the references presented in S5 Table . Based on the results obtained from this study , we observe that correlation learning can be used as a screening tool to reduce the number of markers in GWAS study . As already noted , most methods used in variable selection in linear regression fail in high and ultra-high dimensional settings . This study has presented a simple yet powerful tool to solve this problem . Therefore , the Arabidopsis thaliana GWAS results of this study are reliable .
The single locus tests in standard GWAS methods have been used successfully in identifying thousands of genetic variants associated with hundreds of various traits . As noted by Segura et al . [28] , when we carry out single-locus tests of association , we risk using the wrong model unless the trait is actually due to a single locus . Single locus tests fail to consider the joint effect of multiple genetic markers on traits . They also suffer from the issue of multiple test correction for the threshold value of significance test . The Bonferroni correction is too stringent , and many significant loci are missed out . This calls for multi-locus testing in GWAS . Because only a subset of SNPs is associated with the phenotype , penalized variable selection methods are appropriate in GWAS . Several penalized methods have been used in solving this problem although they fail when the number of variables , p , is several times larger than the sample size ( n ) . In this article , we developed an algorithm , ISIS EM-BLASSO that screens and significantly reduces the number of SNPs to a moderate number . A moderate-scale variable selection method was used to select variables in the reduced set . We chose SCAD since it is asymptotically oracle efficient . Parameter estimation and significance testing were done in the last stage by applying EM-Bayesian LASSO [25] and likelihood ratio test . This algorithm is based on correlation learning in the screening stage . Our approach lays emphasis on the significance of the correlation between SNPs and the trait of interest in the screening stage . It is only the SNPs that are significantly correlated with the trait that are selected in the screening stage . Hence we do not need to subjectively fix the number of SNPs in the screening stage as in Fan and Lv [14] method . ISIS EM-BLASSO differs with the original sure independence screening method by how screening is done . Secondly , it integrates EM- Bayesian LASSO algorithm [25] in the final stage to select and estimate effects . We compared the results of our new approach with results obtained from EMMA [26] , FarmCPU [24] and SCAD [7] methods . The reasons are varied: EMMA [26] has been a standard gold method for GWAS , FarmCPU [24] just like ISIS EM-BLASSO also reduces the number of markers used in GWAS , SCAD [7] is actually used in the screening method of ISIS EM-BLASSO hence the need to compare it independently and mrMLM [20] integrates single locus with multi-locus approach . ISIS EM-BLASSO is the fastest as compared to these other methods . ISIS EM-BLASSO only takes 3% of the computing time needed by the EMMA method , 16% of the time taken by mrMLM , 20% of the time taken by SCAD and 50% of the time taken by FarmCPU methods . Screening stage reduces the number of SNPs to an average number hence less time taken in the overall process . More importantly , ISIS EM-BLASSO performs generally the best; it has high statistical power , low Type 1 error and low MSE of estimated QTN effects ( S1 , S2 and S3 Tables ) . This algorithm improves the estimation of parameters because even in the last stage , EM-Bayesian LASSO still performs variable selection shrinking other SNPs to zero hence significantly improving the estimates and empirical power . Notice that EM-Bayesian LASSO [25] performs effectively in the last stage because the number of SNPs has been reduced considerably . Just like SCAD , EM-Bayesian LASSO will fail when the number of SNPs runs into hundreds of thousands . For this reason we did not perform tests for EM-Bayesian LASSO independently . Combining sure independence screening , SCAD , and EM-Bayesian LASSO improves the results regarding of empirical power , accuracy , and computational efficiency . In the screening stage of ISIS EM-BLASSO , iterative sure independence screening can be performed several times . In the article , we only performed a single iteration since we chose a high level of significance , 0 . 01 , for identifying predictors that are significantly correlated with the response . At this level of significance , we expect to have moderately more variables in the screening stage so that SCAD can be applied to shrink some of these variables to zero . A lower level of significance and/or multiple test correction in the screening stage might be too stringent and may result in missing significant SNPs at this juncture . Most shrinkage methods will still perform effectively even when the number of variables , p , is moderately greater than or equal to the number of observations , n . Therefore the choice of a level of significance of 0 . 01 without any multiple test correction is justified . Note that if several iterations are performed , then many unimportant variables are selected . Even so , this algorithm is still valid because the extension of this method with EM-Bayesian Lasso [25] in the last stage eliminates un-associated . Other multi-locus GWAS approaches such as multi-locus mixed model ( MLMM ) [28] have been studied before . MLMM approach of Segura et al . [28] is inadequate since its greedy forward-backward method inclusion of SNPs is clearly limited in exploring the huge model space . ISIS EM-BLASSO considers the joint effect of all SNPs that passes the screening criterion . Unlike EMMA and FarmCPU , we do not apply the Bonferroni correction for multiple testing hence ISIS EM-BLASSO performs better than these methods regarding statistical power . Bonferroni correction is indeed too stringent hence it removes some significant SNPs in the final results when EMMA and FarmCPU are used . Although we still used a slightly stringent criterion of LOD value 3 in our final stage , ISIS EM-BLASSO still has high statistical power and low false positive rate , indicating a better performance of the new algorithm over EMMA and FarmCPU . mrMLM compares almost equally with ISIS EM-BLASSO regarding power and MSE ( S1 , S2 and S3 Tables ) . mrMLM is also a two stage method hence the similarity . Nevertheless , ISIS EM-BLASSO is still faster than mrMLM . The results obtained here also demonstrate that many penalized methods fail when the number of SNPs is many times larger than sample size as seen from results obtained when SCAD is used to analyze data . ISIS EM-BLASSO was used to analyze six flowering time traits in this study . As a result , 67 known genes were detected . Among these known genes , 50 were identified only by ISIS EM-BLASSO ( Table 2 ) . Many genes obtained by our approach are in the neighborhood of the 89 SNPs detected ( S5 Table ) . These results are consistent with those previously reported , as shown in the database ( http://www . arabidopsis . org/ ) the work of Atwell et al . [27] and other related references in S5 Table . Atwell et al . [27] listed many significantly associated SNPs to these traits though some of them were not significant at the 0 . 05/m criterion . Therefore , the Arabidopsis thaliana GWAS results presented by our algorithm in this study are reliable .
We considered an iterative modified sure independence screening ( ISIS ) and extended it by applying EM Bayesian LASSO algorithm herein referred to as ISIS EM-BLASSO . This approach was used to identify relevant genes in GWAS study in real data . The new approach is simple , fast and shows high statistical power of detecting relevant SNPs on simulated data . Mean squared errors of the estimated effects are also minimal . We recommend this approach as an accurate and fast alternative for carrying out multi-locus GWAS study especially in high dimensional settings . ISIS EM-BLASSO reduces the search to a moderate number of SNPs that are significantly correlated with the trait of interest . As a result , we reduce the computing time for GWAS , and also ensure that GWAS can be carried out on a small computer . Our extension by EM-Bayesian LASSO ensures that the parameter estimates are reliable .
In this study , we considered the regression model , y=μ+∑j=1qQjαj+∑i=1pXiβi+ε ( 1 ) with y being a n×1 vector of phenotypic quantitative trait , μ is the overall average , Qj is the jth fixed effect which must be included in the model , for example , the population structure , Xi is a n×1 vector of the ith SNP values and ε∼MVNn ( 0 , σ2I ) is the residual error . Given the model in Eq ( 1 ) we corrected y for the fixed effects Qj , j = 1 , 2 , … , q before applying the screening procedure . The effects α^j are obtained by ordinary least squares . Eq ( 1 ) then becomes , yc=y−∑j=1qQjαj=μ+∑i=1pXiβi+ε and without loss of generality , this can just be denoted as y=μ+∑i=1pXiβi+ε ( 2 ) ISIS-EM-BLASSO gives the marker effects estimates , γk which must be tested for their significance to the phenotype under study . We propose that all γ^k≠0 k = 1 , 2 , ⋯ , p could be viewed to be associated with the trait under study . With the model ( 10 ) above at hand , and all γ^k≠0 let’s say of size q , we consider the model’s likelihood functions . Let L0=L ( θ^−k ) and L1=L ( θ^1 ) be short expressions of the natural logarithms of the likelihood functions under the null model and the full model , respectively , with θ^−k={γ ( 1 ) , ⋯ , γ ( k−1 ) , γ ( k+1 ) , γ ( q ) } and θ^1={γ ( 1 ) , ⋯ , γ ( q ) } . In essence , we test the null hypothesis , H0:γ ( k ) = 0 that there is no QTN linked to the marker k . We use log of odds ( LOD ) score as the test statistic . The original likelihood functions ( before taking the natural log ) are l0=eL0 and l1=eL1 , respectively . The LOD score is defined as LOD=log10 ( l0l1 ) =−2 ( L0−L1 ) 4 . 6052 ( 11 ) The LOD score is easy to interpret because of base 10 . We select all markers with a score LOD ≥ 3 and regard them as significant . Note that LOD value 3 is a slightly stringent criterion and is equivalent to p-value: Pr ( χ2 > 3×4 . 6052 ) = 0 . 0002 . If the null hypothesis is true , LOD × 4 . 6052 follows a Chi-square distribution with one degree of freedom . A similar procedure is used to test the significance of the estimates obtained from SCAD and mrMLM . It is important to point out that , for the EMMA and FarmCPU methods we select significant markers based on Bonferroni correction for multiple tests by setting a threshold for P-value at 0 . 05/m , where m is the number of markers . | Genome-wide association study is concerned with the associations between markers and traits of interest so as to identify all the significantly associated markers . In genome-wide association studies , hundreds of thousands of markers are genotyped for several hundreds of individuals . Usually , only a very minor subset of these markers is associated with the trait . Most penalization methods fail when the number of markers is much larger than the sample size . Based on this fact , we have developed an algorithm that proceeds in two stages . In the first stage ( screening ) , we reduced the number of markers via correlation learning to a moderate size . We then used a moderate-scale variable selection method to select variables in the reduced model . Conditional on the selected variables , we repeated the screening procedure and chose another set of variables . In the second stage ( estimation ) , all the above-selected variables are accurately estimated in a multi-locus model . Our approach is simple , accurate in estimation , fast and shows high statistical power of detecting relevant markers on simulated data . We have also used this method to identify relevant genes in real data analysis . We recommend our approach for conducting a multi-locus genome-wide association study . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Conclusion",
"Methods"
] | [
"genome-wide",
"association",
"studies",
"applied",
"mathematics",
"brassica",
"quantitative",
"traits",
"simulation",
"and",
"modeling",
"algorithms",
"model",
"organisms",
"mathematics",
"statistics",
"(mathematics)",
"genome",
"analysis",
"experimental",
"organism",
"systems",
"plants",
"arabidopsis",
"thaliana",
"research",
"and",
"analysis",
"methods",
"mathematical",
"and",
"statistical",
"techniques",
"statistical",
"methods",
"monte",
"carlo",
"method",
"genetic",
"loci",
"plant",
"and",
"algal",
"models",
"phenotypes",
"heredity",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"genomics",
"computational",
"biology",
"organisms",
"human",
"genetics"
] | 2017 | Iterative sure independence screening EM-Bayesian LASSO algorithm for multi-locus genome-wide association studies |
The current understanding of mammalian kidney development is largely based on mouse models . Recent landmark studies revealed pervasive differences in renal embryogenesis between mouse and human . The scarcity of detailed gene expression data in humans therefore hampers a thorough understanding of human kidney development and the possible developmental origin of kidney diseases . In this paper , we present a single-cell transcriptomics study of the human fetal kidney . We identified 22 cell types and a host of marker genes . Comparison of samples from different developmental ages revealed continuous gene expression changes in podocytes . To demonstrate the usefulness of our data set , we explored the heterogeneity of the nephrogenic niche , localized podocyte precursors , and confirmed disease-associated marker genes . With close to 18 , 000 renal cells from five different developmental ages , this study provides a rich resource for the elucidation of human kidney development , easily accessible through an interactive web application .
Mammalian kidney development initiates in the intermediate mesoderm through crosstalk between the metanephric mesenchyme ( MM ) and the ureteric bud ( UB ) . The UB originates from the nephric duct , invades the MM , and starts to subdivide progressively into multiple ramifications . The UB tip cells , which make the first contact with the MM , become enveloped by an assembly of mesenchymal cells , the cap mesenchyme ( CM ) ( Fig 1A and S1 Fig ) . The CM contains nephron progenitor cells ( NPCs ) , which give rise to the whole nephron epithelium through tightly regulated morphogenic transformations [1] . Self-renewal of ( mouse ) NPCs is governed by key transcription factors , such as Six2 and Meox1 , which mark the nephrogenic zone of the kidney [2] . Signaling between UB tip cells and NPCs regulates the balance between self-renewal and differentiation of NPCs [3] . In humans , about 1 million nephrons are produced before the NPC population is irrevocably exhausted a few weeks before birth [4] . During nephrogenesis , NPCs undergo mesenchymal-epithelial transition and differentiate into a succession of intermediate structures: the pretubular aggregate ( PTA ) , renal vesicle ( RV ) , and comma- and s-shaped body ( CSB and SSB , respectively ) . Then , via the capillary loop stage , mature and functional glomerular and tubular structures are eventually formed . In contrast to the nephron epithelium , the collecting duct system originates from the UB . Glial cell-derived neurotrophic factor/ RET proto-oncogene ( GDNF/RET ) signaling between the UB and CM critically regulates proliferation of UB tip cells and branching morphogenesis of the UB [5] . Stromal cells—such as interstitial cells ( ICs ) , mesangial cells , juxtaglomerular cells , smooth muscle cells , fibroblasts , and pericytes—derive from a common interstitial progenitor [6 , 7] . Finally , vascular endothelial cells and the highly specified glomerular endothelium originate from the MM [8] , and leukocytes and erythrocytes enter with the blood stream . The current understanding of mammalian kidney development is largely based on mouse studies , although it is clear that human and mouse kidneys are morphologically different . Three recent landmark studies have revealed , in great detail , a significant divergence between mouse and human renal embryogenesis in terms of morphology as well as gene expression [9–11] . These studies underline that the prevailing lack of data on human kidney development severely hinders the detailed understanding of human kidney development and possible developmental origins of kidney disease . In the study described here , we used single-cell RNA sequencing ( scRNA-seq ) to study gene expression dynamics in human fetal kidney development . Analysis of a fetal kidney from week 16 ( w16 ) of gestation revealed 22 cell types , which we identified by known marker genes . Pseudotime analysis clarified their temporal relationship . We further defined specifically expressed cell type marker genes of which many have not been implied in kidney development . Comparison to four additional samples ( from w9 , w11 , w13 , and w18 ) suggested that most cell types have a constant expression pattern , with the notable exception of podocytes . To highlight two ways in which our data set can be interrogated , we then explored the nephrogenic niche and the development of podocytes . Gene expression differences between four NPC clusters were related to spatial heterogeneity by immunostaining and single-molecule fluorescence in-situ hybridization ( smFISH ) . Expression of the disease-associated gene UNCX was localized to NPCs and their early derivatives . Finally , we focused on podocyte development , which proceeds via a distinct precursor state . By immunostaining and smFISH , we localized these precursors in situ and confirmed the disease-associated gene OLFM3 as a marker .
We performed single-cell transcriptomics on a human fetal kidney from w16 of gestation , equivalent to 14 weeks of development ( Fig 1B ) . After data pruning and stringent removal of cells affected by stress ( S2 Fig , Methods ) , 6 , 602 cells were retained for further analysis . Clusters of cells were identified by hierarchical clustering after k-nearest neighbor smoothing [12] . We assigned cell types to these clusters by expression of marker genes from the literature on mouse kidney development . The studies that linked the genes of this literature set to particular cell types are referenced in S1 Table . After merging similar clusters ( S3 Fig , Methods ) , we obtained 22 cell types ( S4 Fig ) and visualized the single-cell transcriptomes in a two-dimensional t-distributed stochastic neighbor embedding ( tSNE ) map [13] ( Fig 1C ) . The mean expression levels of the literature set genes showed clear differences between cell types ( Fig 2A ) . NPCs , which were distributed over four distinct clusters ( NPCa–d ) , were marked by the established markers SIX2 ( Fig 1C ) , CITED1 , MEOX1 , and EYA1 . Expression of these progenitor markers was highest in NPCa , which we hence considered “bona fide” self-renewing NPCs . NPCb showed lower levels of CITED1 and SALL1 and higher levels of GDNF and HES1 compared to the other NPC clusters . HES1 , a transcription factor downstream of NOTCH signaling , is important for further renal cell differentiation . Compared to the other NPC clusters , NPCc showed higher expression of CRABP2 , which is related to retinoic acid signaling [14] . NPCd exhibited low OSR1 , CITED1 , and MEOX1 expression and increased levels of LEF1 , a known indicator of NPC induction towards differentiation . Compared to the other NPC subtypes , NPCd were also marked by a larger fraction of cells in G2/M-phase of the cell cycle ( Fig 2B and 2C ) and a higher expression of proliferation markers ( Fig 2D ) , which indicated faster proliferation . We will discuss the relationship between the various NPC clusters in more detail below ( see Heterogeneity in the nephrogenic niche ) . Nephrogenesis continues with the creation of the pre-tubular aggregate ( PTA ) cells , which in turn develops into the RV and CSB . In our data , PTA cells were identified based on high expression of LHX1 ( Fig 1C ) , JAG1 , WNT4 , and CCND1 . Because RV and CSB are mainly distinguishable by morphology , cells belonging to these two structures were grouped in our analysis ( RVCSB ) . RVCSB cells were marked by the same genes as PTA cells , but they appeared to proliferate less ( Fig 2B–2D ) . Furthermore , they expressed markers reflecting more advanced regional patterning , which allowed us to discriminate between two subtypes ( a and b ) . RVCSBa had a higher expression of genes that were recently associated with the distal RV ( SFRP2 , DLL1 , LHX1 ) , whereas RVCSBb expressed genes that indicate the proximal RV ( CDH6 , FOXC2 , MAFB , CLDN1 , WT1 ) . The next step in development is the formation of the SSB . In our data set , this structure was represented by three clusters , named according to the part of tubule and glomerular epithelium they are known to give rise to—SSBpr , proximal tubule; SSBm/d , medial/distal; SSBpod , podocytes . SSBpr were identified in our data by markers of the early proximal tubule ( ErPrT ) ( such as HNF1A , CDH6 , AMN ) , as well as low expression of SLC3A1 , LRP2 , and SLC13A1 , which are known to be found in more mature proximal tubule cells . Therefore , SSBpr were likely precursors of the ErPrT cells , which expressed higher levels of early proximal markers together with CLDN2 ( Fig 1C ) , ANPEP , and SLC34A1 . Another cluster ( SSBm/d ) accounted for the precursor cells of the loop of Henle and the distal tubule in the SSB . This cluster could be identified by the presence of IRX1 , IRX2 , SIM2 , SOX9 , POU3F3 , and HNF1B , together with low expression of PAPPA2 and MAL and the absence of CDH6 and HNF1A . Cell types that are known to develop from the SSBm/d were found together in one cluster ( distal tubule/loop of Henle [DTLH] ) . This cluster showed high expression of the distal markers MAL , CLCN5 , SLC12A3 , and POU3F3 , which are specific to the distal tubule , as well as SLC12A1 , PAPPA2 , and UMOD , which are found in the loop of Henle . Finally , cells that likely gave rise to podocytes , SSBpod , clustered separately . These cells showed high expression of MAFB and FOXC2 , both transcription factors necessary for the development of podocyte identity , and low levels of the mature podocytes markers CLIC5 ( Fig 1C ) , PTPRO , NPHS1 , and NPHS2 . This cluster also showed the highest expression of OLFM3 , previously identified as a specific marker of podocyte precursors residing in the visceral part of the proximal segment of the SSB . In contrast to SSBpod , podocytes ( Pods ) showed higher expression of mature podocyte markers and lower levels of MAFB . Differences between SSBpod and Pods will be studied in more detail below ( see Podocyte development ) . Because of the high similarity in gene expression between SSB and capillary loop stage , we could not exclude that the SSB clusters also contained cells from the capillary loop stage . Cells of the connecting tubule ( CnT ) , which connects the distal tubule to the collecting duct , could also be identified in the data . They shared markers with the collecting duct ( such as ALDH1A1 , TACSTD2 , and CDH1 ) , distal tubule ( SOX9 , POU3F3 ) , and UB ( RET , KRT8 , KRT18 , MMP7 ) . Cells of the UB and collecting duct ( UBCD ) were strongly marked by well-known genes like AQP2 ( Fig 1C ) , CALB1 , KRT8 , KRT18 , RET , and GATA2 , found in the collecting duct as well as the stalk and tip of the UB . The developing nephrons are surrounded by interstitial tissue , a separate lineage that originates in interstitial progenitor cells ( IPCs ) . We identified IPCs by coexpression of FOXD1 and GDNF . These cells also expressed lower levels of markers known to be found in more mature cells like PDGFRA for ICs or PDGFRB ( Fig 1C ) and ACTA2 for mesangial cells . We identified two subtypes of ICs ( a and b ) , which were similar in their marker gene profile . Compared to IPCs , they lacked FOXD1 and expressed less ( ICa ) or no ( ICb ) GDNF . ICa showed high levels of FGF7 , which has been localized to the renal fibroblasts or stroma surrounding the ureter and the collecting system . ICa also showed high levels of TPM2 and ACTA2 , markers of smooth muscle-like cells . ICb , on the other hand , expressed genes like DCN ( Fig 1C ) , DES , SERPINE2 , and COL3A1 , which are known to mark cortical stromal cells . Endothelial cells were identified by markers such as KDR and TEK , whereas leukocytes showed many specifically expressed genes , such as CD37 or CD48 . Finally , one cluster of cells ( Prolif ) had a higher expression of proliferation markers compared to most other cell types ( Fig 2D ) but lacked discernible cell type markers . The literature-based analysis of the found clusters seemed to suggest that cells cluster by developmental progression ( e . g . , NPCs versus PTA cells ) , as well as location ( e . g . , RVCSBa , distal , versus RVCSBb , proximal ) . Because the interpretation of clusters is sometimes based on genes that are expressed in multiple developmental stages , we wanted to retrieve the developmental flow with an independent method . We used Monocle 2 [17] to learn a graph that represents the developmental hierarchy of the cell types from the PTA on ( Fig 3A ) . Subsequently , cells were placed on a pseudotime scale rooted in the PTA ( Fig 3B and 3C ) . This analysis showed that PTA cells were followed by RVCSBa and RVCSBb , the SSB clusters , and finally the clusters identified as more mature types ( DTLH , ErPrT , Pod ) . Therefore , the clustering was strongly driven by developmental progression . RVCSBa cells were distributed over a fairly broad period of pseudotime and already occurred before branch point 1 , which separates proximal from distal cell fates ( Fig 3A ) . This might indicate that some of these cells preceded the RVCSBb , whereas others were primed to develop into distal fates . RVCSBb cells , however , only appeared after branch point 1 , which confirmed that they were likely progenitors of proximal cell fates . On three separate branches , SSBm/d preceded DTLH , SSBpr preceded ErPrT , and SSBpod preceded Pods , which confirmed the identity of the SSB clusters . The temporal relationship of the NPC subtypes will be discussed in detail below ( see Heterogeneity in the nephrogenic niche ) . To further confirm the interpretation of the cell clusters , we wanted to compare our data with an existing single-cell transcriptomics study of a w17 fetal kidney by Lindström and colleagues [19] . To that end , we first corrected for batch effects , using a method based on matching mutual nearest neighbors in the two data sets [20] . After correction , the two data sets showed a large degree of overlap ( S5A Fig ) . This allowed us to use the cell types found by Lindström and colleagues to classify the cell clusters found here , using a k-nearest neighbors approach ( see Methods ) . NPCa–c were also classified as NPC by Lindström and colleagues , whereas NPCb were considered “primed NPC , ” which supports the notion that NPCb were primed to differentiate . The NPCd cluster was classified as “proliferating cells . ” This classification is in agreement with our observation that NPCd seemed to proliferate more than other NPC subtypes ( Fig 2B–2D ) . Because NPCd expressed low levels of NPC markers ( such as SIX2 and CITED1 ) , these cells were likely in a transition state between NPCs and PTA cells . Whereas the majority of PTA cells identified here were considered “PTA/RV I” by Lindström and colleagues , RVCSBa cells were spread over multiple cell types . This spread was likely due to the fact that transitory cell types are transcriptionally similar , and their clustering is therefore less robust . Nevertheless , the “PTA/RV II” cluster received most of the RVCSBa cells . RVCSBb cells were called “podocyte precursors” in the Lindström data set , whereas SSBpod as well as Pods were classified as “podocytes . ” In our data set , RVCSBb directly preceded SSBpod ( Fig 3A ) , so they could indeed be considered podocyte progenitors . Below , we will show that SSBpod did form a cell state separate from Pods and should not be grouped with them ( see Podocyte development ) . In agreement with our analysis , the majority of SSBpr were classified as “proximal precursor” or “proximal tubule , ” and all ErPrT were considered “proximal tubule” by Lindström and colleagues . CnT and DTLH were both classified as “distal/loop of Henle ( LOH ) precursor . ” The fact that two cell types in the study by Lindström and colleagues ( “podocytes” and “distal/LOH precursor” ) were split in multiple subclusters in our study likely reflects differences in sample preparation . Whereas Lindström and colleagues preferentially released single cells from the nephrogenic niche , here , the whole kidney was used . Consequently , the Lindström data set has a finer resolution of NPCs and early , proliferating cell types , whereas our data set allowed us to resolve more mature cell types . The two data sets therefore complement each other . To confirm the inferred cell types and also identify novel markers , we pursued two complementary strategies . First , we determined a set of marker genes based on their usefulness as classifiers for individual cell types: for each gene , the performance of a binary classifier was evaluated by the area under the receiver operating characteristic ( AUROC ) and combined with expression level filtering ( see Methods ) . This resulted in 88 marker genes ( S6A Fig , marker set , S3 Table ) . Only 11 of these markers overlapped with the 89 genes in the literature set ( S6C Fig ) . To our knowledge , many of the remaining markers had not been associated with kidney development in previous studies . As an independent approach , we used the KeyGenes algorithm [21] to identify classifier genes among the 500 most highly variable genes ( HVGs ) , using two-thirds of all cells as a training set . Based on the classifier genes determined by KeyGenes , we next predicted the cell types of the remaining one-third of the cells ( test set ) . Cell types could be predicted with an average certainty ( id score ) of 0 . 59; 24% of the cells in the test set obtained an id score higher than 0 . 8 . Of the 95 classifier genes ( S6B Fig , KeyGenes set , S3 Table ) , 24 were the same as in the marker set , and 14 were common with the literature set ( S6C Fig ) . Because the interpretation of the found cell clusters was largely based on markers identified in mouse development , we were wondering whether the new markers identified here were informative for the classification of cell types in the mouse kidney . Using a scRNA-seq measurement of cells from a whole P1 mouse kidney [22] , we plotted the expression of the newly identified marker genes in single cells ( S7 Fig ) . In many cases , markers that were found to label a particular cell type in the human fetal kidney were coexpressed in the same subset of mouse cells . A few markers , however , were either ubiquitously expressed or almost completely absent . This might be due to interspecies differences . By establishing the identity of cell clusters at w16 , we obtained a snapshot of cell type diversity in the fetal kidney . To explore whether the identified expression patterns change dynamically throughout development , we analyzed four additional samples from different developmental ages ( w9 , w11 , w13 , and w18 ) , which together contained 11 , 359 usable cells . Using , again , batch correction based on mutual nearest neighbors [20] , we visualized all samples in a common tSNE map ( Fig 4 , S8 Fig ) . Overall , gene expression in the different samples was largely overlapping for the majority of cell types . For example , proximal tubules cells ( ErPrT ) appeared at the same positions in the tSNE map in all samples ( Fig 4B ) . The position of Pods , however , shifted systematically across different ages , which corresponds to a continuing change in expression pattern ( Fig 4C ) . This observation might suggest that Pods further matured in terms of their expression pattern after being specified . Differential expression analysis of Pods of different ages revealed 109 differentially expressed genes ( fold change > 2 in any comparison , false discovery rate ( FDR ) < 0 . 05 , S4 Table ) . Functional annotation analysis of these genes showed significant enrichment of two gene ontology ( GO ) terms—“proteinaceous extracellular matrix” ( adjusted p-value = 1 . 9 × 10−3 , including SPON2 , BGN , COL1A2 , and CTGF ) and “extracellular exosomes” ( adjusted p-value = 1 . 4 × 10−3 , including NPNT , S100A10 , ANXA1 , and EPCAM ) . Some of the differentially expressed genes have been shown to be important for kidney development . For example , NPNT and DCN showed increasing expression from w11 to w18 . Knockout of the extracellular matrix protein NPNT in mice decreases the invasion of the UB and causes agenesis or hypoplasia [23] . NPNT was further shown to be expressed in the glomerular basement membrane and to be necessary for podocyte adhesion in mice [24] . Ablation of this gene in mice causes podocyte effacement . As in the case of NPNT , DCN has been reported to be part of the glomerular basement membrane proteins [25] . This gene appeared strongly up-regulated in podocytes between w11 and w13 or w18 ( fold changes of 3 . 25 and 4 . 6 , respectively ) . The increase of NPNT and DCN expression over time in our data set could reflect an increase in adhesion between podocytes and glomerular basement membrane . Pods further showed significant differential expressions of genes related to stress , like HSPA1A and HSPA1B or NFKB genes ( NFKB2 , NFKBIA , and REL ) , with the highest levels at w18 . This might suggest that dissociation-related stress increases with age for podocytes , maybe related to stronger adhesion of the cells , or that stress-related genes have another , physiological role in development . We would like to emphasize that the observed gene expression changes with age should be considered with caution because they might be related to the differences in genotype between the samples . A much larger number of samples would be necessary to rule out such interindividual differences as a cause . Having established the identity of the cell clusters , we next wanted to demonstrate how the data set can be used to explore different aspects of kidney development . We specifically focused on the nephrogenic niche , which showed pronounced heterogeneity , and the development of podocytes , which progressed via a distinct , intermediate cell state ( SSBpod ) . The formation of the nephron epithelium starts with the NPCs that differentiate and form the PTA , RV , and CSB . Studies in the mouse suggest that cells in the NPC compartment are not biased towards a particular lineage and patterning is first detectable in the PTA [26] . Nevertheless , the w16 scRNA-seq data indicated the presence of several nephron progenitor subpopulations , NPCa–d . To clarify the temporal relationship of these clusters , we employed Monocle 2 again to arrange them together with the PTA cells on a pseudotime scale ( Fig 5A ) . NPCa clearly preceded NPCb and c , which seemed to appear around the same pseudotime . NPCd cells followed NPCb and c and preceded PTA . This analysis suggested that NPCa are the “bona fide” NPCs and give rise to NPCb and c . NPCd , which were likely more proliferative than the other NPCs ( Fig 2B–2D ) , seemed to be an intermediate state between ( slowly cycling ) NPCa–c and the PTA . To localize the NPC clusters in the tissue , we made use of the fact that they expressed various levels of CITED1 and SIX2 ( Fig 2A ) : whereas NPCa and NPCc exhibited roughly similar levels of these markers , NPCb and NPCd had lost CITED1 almost completely , while retaining some SIX2 expression . In an immunostaining of a w15 kidney , CITED1 and SIX2 appeared overlapping in a subset of cells ( Fig 5B–5D ) . Quantification of the fluorescence signal ( Methods ) revealed clear differences between their expression patterns . Whereas SIX2 expression was approximately constant throughout the CM , CITED1 expression decreased , relative to SIX2 , with increasing ( radial ) distance from the UB ( Fig 5D ) . A marked drop of CITED1 was visible between 10 and 20 μm from the UB , which approximately corresponds to the first layer of cells . To exclude that the observed difference between SIX2 and CITED1 expression was due to the different fluorophores on the secondary antibodies , we repeated the experiment with swapped fluorophores . This measurement produced a very similar expression gradient ( S9A Fig ) . To exclude that the observed effect was influenced by PTA found in the CM towards the stalk of the UB , the analysis was repeated , taking only the 20% of CM cells closest to the edge of the cortex into account . A similar expression gradient was observed ( S9C Fig ) . This result implies the existence of a CITED1 low/SIX2 high subpopulation of cells , which are not in contact with the UB . Secondly , we observed that CITED1 decreased relative to SIX2 towards the interface with the PTA and the stalk of the UB ( Fig 5D ) . A similar observation was made when the experiment was repeated with swapped fluorophores ( S9B Fig ) . Taken together , these results suggested that NPCa and NPCc were located closer to the surface of the UB and closer to the tip of the UB compared to the other NPC subtypes . Additionally , we also observed differences in subcellular localization of CITED1 protein within the CITED1 high compartment . Whereas for the majority of cells CITED1 was found in the cytoplasm , in several cells it was concentrated in the nucleus ( right inset in Fig 5C ) . In contrast , SIX2 was always found restricted to the nucleus ( left inset in Fig 5C ) . This observation might indicate that CITED1 was only active in a small population of cells , which would constitute another layer of cell–cell heterogeneity . In addition to the observed heterogeneity in CITED1 and SIX2 , differences between the NPC clusters could also be gleaned from the set of novel markers ( marker set , S3 Table ) . TMEM100 and RSPO3 specifically marked NPCa . RSPO3 is an activator of the canonical WNT signaling pathway [27] , suggesting a role of WNT either in NPC self-renewal or UB branching morphogenesis . Notably , all markers of NPCb ( CACYBP , MRPL18 , ZFAND2A , DNAJB1 ) were related to the stress response in some form . The markers of NPCc ( CRABP2 , HAS2 , MDK , HOXC6 ) , which were also expressed in the other NPC types , are all either targets of retinoic acid or binding it [14 , 28 , 29] . MDK has been shown to be expressed in the CM of the developing rat kidney , and its neutralization reduced the number of formed nephrons in vitro [30 , 31] . Finally , the NPCd markers CENPF , HMGB2 , CCNB2 , and NUSAP1 all have a role in cell cycle regulation or proliferation [16] . HMGB2 was recently implied in the activation of quiescent adult neural stem cells [32] . The observation that markers of NPCb were related to the stress response seemed to suggest that this cluster was created as an artefact of cell dissociation [33] , despite our best efforts to remove stressed cells ( see Methods ) . On the other hand , the vast majority of NPCb cells were classified as “primed NPC II” in the Lindström data set ( S5 Fig ) . The fact that NPCb cells were only detected in the w16 and w18 kidneys is consistent with single-cell dissociation becoming increasingly difficult with fetal age , or alternatively , with a progenitor cell aging phenomenon . To explore the differences between NPCb and the other NPC clusters further , we immunostained HSPA1A and NR4A1 , both known stress-response genes , in w15 kidney sections ( Fig 5E ) . HSPA1A was identified as a marker of NPCb ( S9D Fig , S3 Table ) , whereas NR4A1 was expressed in multiple NPC clusters but highest in NPCb ( S9D Fig ) . Furthermore , NR4A1 was also identified in the study by Adam and colleagues [22] to be up-regulated in response to elevated temperatures during enzymatic dissociation . HSPA1A and NR4A1 were both observable in the nephrogenic niche at the level of the stalk of the UB and at the transition to the PTA or RV . Additionally , we studied the expression of EGR1 , another stress-related gene that marked NPCb , with smFISH ( S9E Fig ) . EGR1 was mainly found toward the stalk of the UB and in a few cells around the tip of the UB , whereas SIX2 and CITED1 transcripts were visible throughout the CM ( S9F Fig ) . Because results obtained in fixed tissue sections are not confounded by dissociation-related artifacts , these immunostainings and smFISH measurements supported the existence of NPCb cells in the fetal kidney . The fourth NPC cluster , NPCd , was clearly distinguished by proliferation markers ( Fig 2B–2D ) . To locate NPCd in the tissue , we immunostained the cell cycle regulator and NPCd marker CKS2 ( S2G Fig ) . CKS2 signal could be observed around the stalk of the UB and in RVs ( Fig 5E ) . This result supported the interpretation that NPCd were a proliferating transitory state between NPCa–c and PTA . Given the crucial role of the nephrogenic niche in the development of nephrons , it is likely that misexpression or mutation of genes that are specifically expressed in NPCs affect kidney function . Mining a database of genome wide association studies ( GWAS ) revealed that genes that were differentially expressed in NPCs were significantly enriched for association with kidney disease ( p = 1 . 7 × 10−3 , one-sided Fisher’s exact test ) . No enrichment was found for lung diseases ( p = 0 . 21 , one-sided Fisher’s exact test ) ( see Methods , S10 Fig , S4 Table ) . Unsurprisingly , several of the disease-associated genes are known regulators of kidney development , such as SALL1 , SOX11 , and HAS2 [34–37] . The other identified genes had not been previously associated with kidney development . For example , DDX1 , which was differentially expressed in NPCs as well as SSBpod , is an RNA helicase that promotes micro-RNA maturation [38] . UNCX , which was broadly expressed in all NPC clusters , is a homeobox transcription factor involved in somitogenesis and neurogenesis [39] and has also been found to be up-regulated in the induced mouse nephrogenic mesenchyme in culture [40] . It was recently associated with renal-function–related traits [41] as well as glomerular filtration rate [42–44] . In our data , the expression profile of UNCX was similar to that of CITED1 ( S9H Fig ) . Immunostaining of UNCX confirmed the scRNA-seq results and showed expression of UNCX in the nephrogenic zone , as marked by CITED1 ( Fig 5F and S9I Fig ) . These findings suggested UNCX as a novel potential regulator of early nephrogenesis . Another cell type of high relevance for kidney function is the podocyte . This cell type is critical for filtration and is implied in several forms of kidney disease [45] . Podocytes wrap around the glomerular basement membrane ( capillary bed ) inside Bowman’s capsule ( Fig 6A ) . Clustering ( Fig 1C ) and pseudotime analysis ( Fig 3 ) of the w16 kidney data set had indicated that development into Pods occurs via a distinct intermediate state that we dubbed SSBpod here . This cell state was likely related or even identical to previously discovered podocyte precursors [19 , 46] . In the Lindström data set [19] , SSBpod and Pods were both classified as “podocytes , ” and the RVCSBb were considered “podocyte precursors” ( S5B Fig ) . To show that SSBpod cells were indeed a localizable cell state distinct from Pods , we further investigated their expression pattern ( Fig 6B ) , focusing on known literature markers ( literature set ) and the marker set ( S3 Table ) . Compared to RVCSBb , SSBpod showed higher expression of MAFB and FOXC2 , which are necessary for the determination of podocyte identity [47 , 48] . On the other hand , compared to Pods , they exhibited lower expression of genes typically associated with more mature podocytes , like CLIC5 , PODXL , and PTPRO . Filtration function-related genes like NPHS1 , NPHS2 , and PTPRO were expressed at intermediate levels in SSBpod compared to RVCSBb , where they were absent , and Pods , where they are highly expressed . A similar pattern could be observed for genes associated with podocyte polarization or structural organization as well as pedicel growth and patterning . Finally , Pods showed the expression of genes that negatively regulate the cell cycle and support long term survival , consistent with their postmitotic nature [49] . In contrast , SSBpod specifically expressed ORC4 , which has a function in DNA replication . However , proliferation markers were lowly expressed in both SSBpod and Pods ( Fig 2D ) , which suggested low proliferative potential in both cell types . In contrast to NPCs , association with kidney disease was not significantly enriched among genes differentially expressed in SSBpod ( p-value = 0 . 1 , one-sided Fisher’s exact test ) . One of the disease-associated genes was OLFM3 , which has been associated with glomerular filtration rate [50] . OLFM3 , a secreted glycoprotein , has a known function in brain and retina development [51] and has been identified as a marker for podocyte precursors in two independent studies [19 , 46] . In our data set , it was specifically expressed in SSBpod ( Fig 6C ) and was a marker for this cell type in the marker set and KeyGenes set ( S3 Table ) . In order to localize SSBpod , Pods and mesangial cells in situ , we immunostained w15 kidney sections with antibodies for MAFB , PODXL , and ACTA2 ( Fig 6D ) . As expected , PODXL and MAFB were found in Pods at the capillary loop stage and in more mature glomeruli . MAFB staining extended to the proximal segment of the SSB , which indicated that SSBpod may be part of this structure . To locate the SSBpod cells more precisely , we performed smFISH on CLIC5 and MAFB , expressed both in Pods and SSBpod ( Fig 6E ) . We observed a subpopulation of MAFB+/CLIC5− cells outside the glomeruli , which we identified as the SSBpod . These cells could be found predominantly in the visceral part of the proximal segment of the SSB but also at the capillary loop stage . This result supported the notion that SSBpod were transient cells that preceded ( mature ) Pods . Having localized the SSBpod , we next wanted to confirm OLFM3 as a marker of this cell type . smFISH of OLFM3 , MAFB , and CLIC5 showed OLFM3 to be coexpressed with MAFB but absent in cells that were positive for CLIC5 ( Fig 6E ) , a marker that persists in podocytes in the adult kidney . Quantification of the density of smFISH signals ( Fig 6F ) showed that OLFM3 was absent in glomeruli but could be detected in the subpopulation we identified as SSBpod ( MAFB+/CLIC5− ) . In summary , these results supported OLFM3 as a robust marker of podocyte precursors . Finally , we were wondering whether our data set would also allow us to identify candidate mechanisms that drive development from SSBpod to Pods . Differential expression analysis revealed 228 genes that had a significant , bigger than 2-fold changes between SSBpod and Pods ( Fig 6G , S4 Table ) . Among these we found factors belonging to multiple signaling pathways , such as FGF1 , VEGFA , HES1 , and EGF1 . Vegfa and Fgf1 are known to have a homeostatic function in podocytes [52–54] , whereas Hes1 , a target of the NOTCH signaling pathway , seems to be necessary for the synthesis of extracellular matrix proteins in these cells [55] . Binding sites for the transcription factor AP-1 were strongly enriched in this set of genes ( 145 out of 228 genes , adjusted p-value = 1 . 3 × 10−5 ) . AP-1 would therefore be an interesting target for perturbation studies in mouse models . All in all , the results presented here complement other , recent , single-cell transcriptomics studies of the fetal kidney . We demonstrated how the data can be interrogated to find expression patterns that will improve our understanding of human kidney development .
Building an organ during development requires the careful balance between two fundamental processes—growth and the creation of structure . In many organs , these two functions are reconciled by self-renewing progenitor cells that can be induced to differentiate . In kidney development , NPCs give rise to the epithelium of the nephron , the functional unit of the kidney . To balance growth with patterning , self-renewal and differentiation of NPCs have to be tightly controlled . It is well established that the niche of the NPC plays an important role in this control , but the precise mechanisms are not well understood . In particular , it is not clear how the position and movement of NPCs in the niche might impact the induction towards differentiation . Heterogeneity in the nephrogenic niche was brought to light first by Mugford and colleagues in 2009 [26] and has been confirmed by multiple recent studies [10 , 19 , 46 , 56 , 57] . Mugford and colleagues used in situ hybridization to study the localization of transcriptional regulators in E15 . 5 mouse kidney . Three distinct compartments were defined in the CM—inner capping mesenchyme ( which lies closest to the cleft of the UB ) , outer capping mesenchyme ( at the tip of the UB ) , and induced mesenchyme ( at the level of the stalk of the UB ) . Although all compartments expressed Six2 , only inner and outer capping mesenchyme expressed Cited1 . The induced CM was distinguished by Wnt pathway activity , as evidenced by Wnt4 expression . Several recent scRNA-seq studies confirmed heterogeneity in the nephrogenic niche . Brunskill and colleagues studied an E12 . 5 mouse kidney and found two subpopulations in the CM , which they classified as uninduced ( Six2 positive , Cited1 positive ) and induced ( Six2 positive , Cited1 negative ) [56] . Among the hundreds of genes that were differentially expressed between these two populations , they found genes related to the Wnt signaling pathway as well as protein vesicular trafficking and degradation . Wang and colleagues also found two subclusters in the CM of the human fetal kidney [57] . They interpreted one subcluster as the self-renewing compartment due to higher expression of markers for cell division . The other subcluster , which showed gene expression related to NOTCH signaling ( HES1 , HEY1 ) , was considered induced . Two studies by Lindström and colleagues [10 , 19] also explored NPC heterogeneity . The first study [10] identified four NPC clusters ( self-renewing , primed , differentiating , and proliferating ) , whereas the second [19] revealed four clusters of NPCs ( I–IV ) , two clusters of primed NPCs ( I–II ) , as well as several clusters of proliferating cells . In the data set presented here , we identified four clusters of NPCs . Among these , NPCa are most likely the self-renewing compartment . In agreement with the studies by Lindström and colleagues [10 , 19] , they expressed the highest levels of CITED1 and TMEM100 compared with the other NPCs . Furthermore , they preceded all other NPC clusters in pseudotime analysis . NPCb showed expression of several genes that modulate NOTCH , bone morphogenetic protein ( BMP ) , and transforming growth factor beta ( TGF-beta ) pathway activity , as well as low levels of LEF1 , which has been shown to indicate induction towards differentiation [10 , 19] . The classification of NPCb as “primed NPC” by comparison to the Lindström data set supported the interpretation of NPCb as a state distinct from NPCa that is primed to differentiate . The fact that we detected NPCb only at w16 and w18 leads us to speculate that NPCb could be the result of continuous changes in the nephrogenic niche over the course of development . The third NPC cluster , NPCc , appeared together with NPCb in pseudotime and was distinguished from the other NPCs by higher expression of genes involved in or regulated by retinoic acid signaling . The retinoic-acid binding protein CRABP2 has been identified as an NPC marker in other reports [10 , 57] . We speculate that NPCc are the result of spatially varying concentrations of retinoic acid , which is produced in the cortical interstitium [58] . Finally , NPCd appeared between NPCb–c and PTA in pseudotime and were clearly distinguished from the other NPC clusters by increased proliferation , at least as far as that can be inferred from gene expression data . In agreement with our analysis , NPCd were classified as “proliferating cells” by comparison with the Lindström data set [19] . NPCd cells also lowly expressed markers of induction towards differentiation ( such as LEF1 , LHX1 , WNT4 ) , which indicates a transitory state between induced and/or primed NPCs and PTA . The suggested developmental flow from NPCa via NPCb–c to NPCd was supported by a gradual decrease of OSR1 , which is a well-known marker of the early CM . By in situ detection of CITED1 , SIX2 , and other genes , we also explored the spatial localization of the different NPC clusters . NPCa seemed to reside closest to the tip of the UB , the induced and/or primed NPC b and c were situated closer to the stalk , and NPCd were closest to the PTA . This finding is consistent with the recent report of NPCs streaming from their niche at the UB tip towards the UB branch point to form the PTA and RV [19] . On this path , the cells gradually lose the NPC transcriptional program , and differentiation is induced . In the mouse , trajectories of NPCs also seem to have a large stochastic component: NPCs repeatedly detach from the UB and attach again and also shuttle back and forth between the “uninduced region” at the UB tip and the “committed region” around the stalk of the UB [59] . This observation could indicate that varying expression levels of genes such as CITED1 occur as a consequence of cell migration and are not necessarily functionally relevant . Indeed , Cited1 knockout has no adverse effects on kidney development in the mouse [60] . Taken together , our results support a model in which ( self-renewing ) NPCa reside at the tip of the UB , probably in close proximity or even in contact with the UB . Movement away from the UB tip , toward the stalk , is accompanied by decreased CITED1 expression and transformation to the ( induced and/or primed ) NPCb–c states . Arrival at the stalk of the UB is characterized by the NPCd expression state , increased proliferation , and eventually transformation to the PTA . It is conceivable that cells sometimes visit the different NPC states in reverse order , which would reconcile this model with the observed high , multidirectional motility of NPCs [59] . In the prevailing model of mammalian kidney development , self-renewing NPCs are not prepatterned to develop into a certain lineage . When the developing nephron first displays signs of proximal–distal patterning is an important , outstanding question . Mugford and colleagues have found evidence that the PTA , which succeeds the NPCs , is already polarized [26] . A recent study by Lindström and colleagues [19] proposed an intriguing mechanism that couples temporal and spatial cues: whereas NPCs that are recruited to the PTA early develop into distal cell types , NPCs that are integrated later contribute to the proximal compartment . In our study , we identified the PTA by known marker genes ( CCND1 , LHX1 , and WNT4 ) and high proliferation . We were unable to detect any substructure within the PTA , which might be due to the limited resolution of our scRNA-seq method . RV/CSB , the next developmental stage , however , was split in two clusters ( a and b ) . Pseudotime analysis suggested that RVCSBa was a heterogeneous cluster comprising early RV cells ( which appeared before RVCSBb ) and the distal segment of the RVCSB . This observation is consistent with the time-dependent recruitment model by Lindström and colleagues [19] in the sense that in that model , distal specification precedes proximal patterning . Single-cell transcriptomics studies of various organs have brought to light many new , intermediate cell states . This has provoked the question of whether we should consider gene expression in complex tissues as a continuum rather than a collection of distinct expression profiles . In developmental systems , it is certainly useful to think about gene expression change as a continuous process . Nevertheless , there are clearly distinct intermediate cell states even within linear developmental paths . In our study , we observed that RVCSBb gave rise to Pods via an intermediate state , the SSBpod , which directly preceded the Pods in pseudotime . Specific expression of OLFM3 made it likely that this cluster is identical to previously identified podocyte precursors , which were marked by this gene [19 , 46] . Menon and colleagues defined a cluster of “immature” or “early” podocytes , characterized by high OLFM3 and low MAFB expression . In that study , podocytes showed increased MAFB expression but loss of OLFM3 [46] . Lindström and colleagues located OLFM3 positive cells to the proximal part of the SSB [19] . In our study , we confirmed all of these observations: OLFM3 was localized to the visceral part of the proximal segment of the SSB , and OLFM3 negative Pods showed higher expression of mature podocyte markers compared to the OLFM3 positive SSBpod . Functional annotation analysis of genes that were differentially expressed between SSBpod and Pods revealed enrichment of a binding site for AP-1 . This transcription factor has been found to be important for the development of the skin [61] , neural precursor cells [62] , and the heart valve [63] in mice . A role of AP-1 in kidney development has not been described yet , and further research is needed to elucidate its potential function . The analysis of the kidneys from different gestational ages showed high similarity of cell types across different ages with the exception of Pods . These displayed a systematic change in expression pattern , which might indicate the continued maturation of Pods over time . This observation is in agreement with a study by Brunskill and colleagues [64] in the mouse , which compared embryonic ( E13 . 5 and E15 . 5 ) with adult podocytes ( defined as MafB positive cells ) . That study found hundreds of genes that were differentially expressed between embryonic and adult podocytes . Furthermore , targeted experiments are needed to demonstrate the possible maturation of podocytes in human kidney development . In summary , we have leveraged a combination of single-cell transcriptomics and in situ imaging to study the intricate structure of the developing human kidney . The transcriptomics data , accessible via a web application ( http://www . semraulab . com/kidney ) , will be a valuable starting point for discovering gene regulatory mechanisms or finding new disease mechanisms .
The collection and use of human material in this study was approved by the Medical Ethics Committee from the Leiden University Medical Center ( P08 . 087 ) . The gestational age was determined by ultrasonography , and the tissue was obtained by vacuum aspiration from women undergoing elective abortion . The material from six embryos ( w9 , male; w11 , male; w13 , female; w15 , female; w16 , male; and w18 , female ) was donated with written informed consent . Questions about the human material should be directed to S . M . Chuva de Sousa Lopes ( Lopes@lumc . nl ) . | Regenerative medicine offers an exciting avenue to potential cures of kidney disease . However , a detailed knowledge of the structure and embryonic development of the kidney is crucial , both for stimulating regeneration in the body and for growing healthy kidney tissue in a dish . Most of such knowledge has been obtained from mice , whose development differs in crucial ways from that of humans . We therefore studied the composition of human fetal kidney tissue from five developmental ages , using a technique that can measure gene expression in individual cells . Our measurements revealed 22 distinguishable cell types , some of which we localized in the tissue by fluorescence microscopy . We found several subpopulations of nephron progenitors—cells that give rise to the nephron , the functional unit of the kidney . Our study also focused on the development of podocytes , a cell type that is crucial for the filtration function of the kidney , and our results might inform attempts to recreate these cells in a dish . We hope that our data set , made conveniently accessible through a web application , will help scientists develop new regenerative medicine approaches to kidney disease . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"medicine",
"and",
"health",
"sciences",
"kidney",
"development",
"marker",
"genes",
"precursor",
"cells",
"developmental",
"biology",
"organism",
"development",
"stem",
"cells",
"nephrons",
"molecular",
"biology",
"techniques",
"kidneys",
"research",
"and",
"analysis",
"methods",
"specimen",
"preparation",
"and",
"treatment",
"staining",
"animal",
"cells",
"gene",
"expression",
"organogenesis",
"molecular",
"biology",
"methods",
"&",
"resources",
"immunostaining",
"anatomy",
"cell",
"biology",
"genetics",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"renal",
"system"
] | 2019 | Single-cell transcriptomics reveals gene expression dynamics of human fetal kidney development |
Muscle contracts due to ATP-dependent interactions of myosin motors with thin filaments composed of the proteins actin , troponin , and tropomyosin . Contraction is initiated when calcium binds to troponin , which changes conformation and displaces tropomyosin , a filamentous protein that wraps around the actin filament , thereby exposing myosin binding sites on actin . Myosin motors interact with each other indirectly via tropomyosin , since myosin binding to actin locally displaces tropomyosin and thereby facilitates binding of nearby myosin . Defining and modeling this local coupling between myosin motors is an open problem in muscle modeling and , more broadly , a requirement to understanding the connection between muscle contraction at the molecular and macro scale . It is challenging to directly observe this coupling , and such measurements have only recently been made . Analysis of these data suggests that two myosin heads are required to activate the thin filament . This result contrasts with a theoretical model , which reproduces several indirect measurements of coupling between myosin , that assumes a single myosin head can activate the thin filament . To understand this apparent discrepancy , we incorporated the model into stochastic simulations of the experiments , which generated simulated data that were then analyzed identically to the experimental measurements . By varying a single parameter , good agreement between simulation and experiment was established . The conclusion that two myosin molecules are required to activate the thin filament arises from an assumption , made during data analysis , that the intensity of the fluorescent tags attached to myosin varies depending on experimental condition . We provide an alternative explanation that reconciles theory and experiment without assuming that the intensity of the fluorescent tags varies .
Muscle contraction underlies most voluntary and involuntary motion of multicellular animals . The last 100 years have revolutionized our understanding of this process . Advances in light microscopy in the early 20th century culminated in the sliding filament theory , the discovery that muscle contracts via the relative sliding of two sets of protein filaments ( Fig 1A ) [1 , 2] . To achieve this motion , it was hypothesized that myosin molecules in thick filaments form transient bonds with specific binding sites on actin in thin filaments [3] . Subsequent experiments characterized the biochemical reactions between myosin and actin , culminating in the determination of how these reactions couple to the hydrolysis of ATP [4] . More recently , single molecule techniques have allowed direct observation of myosin binding to actin [5] and measurement of the effects of external force on this chemical interaction [6 , 7] . However , despite this wealth of knowledge , a clear understanding of the connection between muscle contraction at the molecular and cell or organ scale remains unclear . Besides integrating a century of research on muscle physiology , such an understanding might allow , for example , novel treatments of genetic heart disease [8] . Connecting the molecular and macro scales of muscle contraction is a complex problem . Forming this connection requires 1 ) understanding how an isolated myosin molecule interacts with a thin filament; 2 ) understanding how this interaction changes when multiple myosin molecules work together; 3 ) understanding how the collective behavior of multiple myosin motors changes upon the addition of accessory proteins like titin and myosin binding protein-C ( MyBP-C ) ; and 4 ) understanding how collections of myosin with accessory proteins work together when arranged as sarcomeres in series , myofibrils in parallel and so on up to whole muscle ( Fig 1A ) . While recent single molecule experiments have given a relatively clear picture of how isolated myosin molecules interact with actin filaments e . g . [5–7] , the remaining three ingredients of a molecular to macro scale understanding remain much less clear . Thus , one obstacle to connecting muscle at the molecular and macro scale is that myosin motors interact with each other via a common actin filament . This coupling occurs in two ways . The first , which we call mechanochemical coupling , occurs because myosin motors apply forces on each other when they bind to the same actin filament . These forces can accelerate chemical reaction rates , and lead to emergent ensemble behavior [9 , 10] . If the actin filament is relatively stiff , then this coupling affects every molecule equally—that is , it doesn’t matter where on the actin filament the bound myosin molecules are located . The second type of coupling , herein termed local coupling , occurs when a myosin molecule binds to the actin filament and affects the reaction rates of nearby myosin molecules , but has no effect on distant myosin . Local coupling happens when myosin binds to actin at low calcium in the presence of the regulatory proteins troponin and tropomyosin . Under these conditions , the protein tropomyosin , which wraps around the actin filament , sterically hinders myosin binding by covering the binding sites on actin ( Fig 1A ) [11] . If a myosin molecule then binds , it deforms tropomyosin . This deformation , being local , makes it easier for neighboring myosin molecules to bind to actin , but has no effect on distant myosin molecules ( Fig 1B–1D ) [12–17] . Direct observation of either type of coupling presents a major challenge , since traditional single molecule techniques cannot be used . With the lack of direct measurements , mathematical models can be used to interpret indirect observations . However , it is unclear whether mechanochemical and/or local coupling can be captured by current muscle models , since many muscle models incorporate the assumption that molecules act independently . Mathematical modeling has played an important role in our understanding of the connection between muscle contraction at the micro- and macro-scale . Shortly after the sliding filament and cross-bridge theories of muscle contraction were proposed [1 , 2] , they were implemented in a mathematical model [3] . Since then , as biochemical and biophysical experiments have uncovered more about the molecular scale , increasingly detailed and increasingly accurate mathematical muscle models have been developed [18–30] . The majority of these models build on the ideas and assumptions of Huxley’s original model [3] . One of these assumptions is that each myosin molecule acts independently of neighboring myosin molecules , the independent force generator assumption [31] . One might expect that any type of coupling between myosin molecules in an ensemble would violate the independent force generator assumption . However , mechanochemical coupling does not . This is because , assuming that the thin and thick filaments are stiff , mechanochemical coupling affects each actin-bound myosin molecule equally , regardless of its position on actin . Thus , the future behavior of each myosin molecule in an ensemble can be determined from the average properties of the ensemble , an alternate statement of the independent force generator assumption [10] . In contrast , local coupling does violate the independent force generator assumption [32] . This is because the future behavior of a myosin ensemble cannot be predicted simply by average ensemble properties , but also depends on the spatial distribution of those properties . For example , two ensembles with , say , 25% of the myosin molecules bound to actin will behave differently if all of the bound molecules are tightly clustered together or if the bound molecules are distributed evenly [13] . Thus , muscle models that incorporate Huxley’s independent force generator assumption cannot capture aspects of muscle activation where local coupling plays an important role . Since local coupling violates the independent force generator assumption , modeling locally coupled myosin ensembles is challenging . Some of the earliest studies [13] recognized that local coupling occurs in some physical models . In particular , the 1D Ising model considers the behavior of an ensemble of molecules , each of whom is coupled to its nearest neighbor . For this model , analytic expressions can be derived for the equilibrium distribution of the ensemble [13 , 33] . But this approach is also limited , in that it is unclear how to model long-distance coupling or non-equilibrium effects , both of which are of central interest in muscle modeling . An alternate approach is to explicitly model each molecule in a myosin ensemble [14 , 34–40] . To capture stochastic effects , Monte-Carlo methods are used [41] . One advantage of this approach is that , since each molecule is considered explicitly , local coupling can be easily included . Thus , effects like filament elasticity , which can contribute to local coupling , can be incorporated into models [35 , 37] . These models require significant computational expense , and so have only recently become tractable . But , even with modern computational power , optimizing fits of such models to data is difficult . As a result , most muscle models that include local coupling cannot describe a large suite of experiments . With the aim of developing a muscle model that describes the largest number of experimental results , a model has recently been proposed that includes both mechanochemical and local coupling in a realistic way and yet can be implemented with minimal computational expense [32 , 42 , 43] . This model describes a series of experiments that indirectly measure local coupling between myosin molecules . In particular , the local coupling model ( summarized in more detail in the Methods section ) , which has only two parameters , when incorporated into a model of mechanochemical coupling [10] , describes measurements of in vitro motility of thin filaments at low ATP [14 , 42] , at more physiological ATP and in the presence of MyBP-C [43–45] and at physiological ATP and variable calcium [14 , 32 , 46] . Importantly , the same model with consistent parameters describes these in vitro motility experiments and then reasonably predicts fiber level experiments [32 , 47–51] . It is therefore plausible that this model describes how myosin motors work together , and how that interaction changes with the addition of the accessory protein MyBP-C . Given that these are requirements to a multi-scale understanding of muscle contraction , the model represents progress in that direction . However , as yet , the connection of this local coupling model to the molecular scale is speculative , as the model has not been compared to direct molecular-scale measurements . Recently , the first direct observations of myosin binding to thin filaments were made across a range of activating conditions [52] . Analysis of these data , which clearly demonstrate local coupling , revealed a discrepancy between experiment and theory: the experimental data point towards two heads minimally being required to activate the thin filament whereas the model predicts only one head is required . To understand this discrepancy , we simulated the experimental data using a stochastic method and analyzed the outputs using the same algorithms used for the experimental data . Remarkably , by assuming that fluorescently-tagged myosin emits a constant amount of light per unit time , we found that the model is consistent with the experimental measurements . This assumption , for which we find indirect support , contrasts with an assumption made by Desai et al . [52] that emission depends on experimental condition . Thus , here we propose an alternative scenario where a single myosin head is sufficient to activate a thin filament , thereby reconciling experimental data and theoretical predictions .
To describe the local coupling between myosin molecules , we use a simplified mechanochemical model of the thin filament , based on the continuous flexible chain model [15–17] . In this model , troponin/tropomyosin is modeled as a continuous elastic beam , which contrasts with more structurally detailed models of local coupling that explicitly model each troponin and tropomyosin molecule , e . g . [36 , 37] . Although this simplifying assumption makes some effects more difficult to include ( e . g . potential variations in troponin density [37] ) , one advantage of this approach is that model parameters are related to physical properties of the system . The local coupling model used here starts with three assumptions based on the continuous flexible chain model [15–17] 1 . troponin/tropomyosin ( Tn-Tm ) is a slender , infinite , linear-elastic beam constrained to a plane; 2 . the interaction of Tn-Tm with the actin filament , described by potential energy density , W ( z ) , varies azimuthally along actin but does not vary ( or the variations can be averaged out ) longitudinally along actin; 3 . myosin binding induces a local displacement of the Tn-Tm beam . These assumptions are shown in Fig 1B . A few simplifications differentiate this model from the continuous flexible chain model [42] . The first simplification is based on the observation that the energy of the Tn-Tm beam ( E ) contains two quantities: an elastic component that is minimized when the beam is straight , and a potential energy component that depends on the details of W ( z ) . Importantly , the elastic component dominates over short distances , while the potential energy component dominates over long distances . Therefore , when two nearby myosin molecules bind to actin , the Tn-Tm beam between the two myosin remains relatively straight , minimizing elastic energy; conversely , when two distant myosin molecules bind to actin , they each induce independent deformations , minimizing potential energy ( Fig 1C ) . These ideas lead to specific predictions of the energy change of the actin-Tn-Tm system when myosin binds to actin ( ΔE ) . Consider , for example , the energy change that occurs when a myosin molecule binds to actin when another myosin molecule , a distance L away , has previously bound to actin . If L is small , the Tn-Tm beam should be displaced uniformly , minimizing elastic energy . And , since W ( z ) does not vary longitudinally along actin , the energy change should increase linearly with L . Conversely , if L is large , then myosin binding should induce an independent deformation in the Tn-Tm beam , minimizing potential energy , and the energy change should not depend on L ( see Fig 1D ) . It is convenient to assume that the transition between these two regimes is abrupt , and occurs at some critical separation L = LC . Then , the final curve is defined by two parameters , C and ε . The first is the critical length scaled by the separation between myosin molecules , C = L C / Δ L . The second is the reduction in attachment rate due to the regulatory proteins—i . e . if myosin binds to unregulated actin at a rate k0 , then myosin binds to regulated actin at a rate εk0 when L ≫ LC ( see Fig 1D ) . These results are independent of the details of Tn-Tm’s interaction with actin ( i . e . the exact shape of W ( z ) [42] ) . These ideas define any pair-wise interaction between myosin molecules , and therefore we can calculate the energy change that would occur with the binding of any myosin molecule . Assuming that attachment rate depends exponentially on this energy , we can then incorporate regulation into cross-bridge muscle models [10] . More specifically , given a cross-bridge muscle model , regulation is incorporated by modifying the attachment rate and leaving all other rate constants unchanged . Thus , if myosin binds to unregulated actin at a rate k0 in a particular model , and supposing that a given molecule’s ith neighbor to the left and jth neighbor to the right are bound to actin , then that given molecule has an attachment rate k from the following equation [42] k ( i , j , C , ε ) = { ε k 0 : i , j ≥ C ε i / C k 0 : i < C , j ≥ C ε j / C k 0 : i ≥ C , j < C ε ( i + j - C ) / C k 0 : i , j < C , i + j > C k 0 : i + j ≤ C ( 1 ) As this attachment rate depends on the state of nearby molecules , it introduces local coupling to the model . Differential equation methods can be used to simulate the behavior of myosin ensembles with both this kind of local coupling and also mechanochemical coupling [32 , 42] . These differential equation methods are much more computationally efficient than Monte-Carlo methods , allowing optimization of fits and parameter estimation . Thus , from fits to in vitro motility data , we can obtain values for the parameters C = 11 and ε = 0 . 003 at pCa 8 [42] , where pCa is the negative log of the calcium concentration . With these parameters , the model fits motility experiments in the presence of MyBP-C , suggesting that MyBP-C both activates the thin filament and specifically binds to actin , thereby competing with myosin and creating a viscous drag [43] . By assuming that calcium only affects ε , and that C remains constant , force-pCa experiments from muscle fibers and fiber twitch experiments can be reproduced and then motility-pCa curves and twitch summation experiments are successfully predicted [32] . The reasonable agreement between model and experiments provides support for its assumptions . If this model is correct , then at low calcium , the binding of one myosin molecule can locally activate the thin filament . Desai et al . [52] interpret their experiments to support the view that at least two myosin molecules are necessary to activate the thin filament , contradicting the model . To understand this difference we used stochastic methods to simulate the experimental observations . In the previously published experiments , the head domain of myosin ( S1 ) was directly observed binding to thin filaments ( actin with troponin and tropomyosin ) [52] . The experiment starts by using fluid flow to suspend a thin filament between two silica beads affixed to a glass surface ( Fig 2A ) . Then , a solution is added that includes variable concentrations of 1 ) GFP-labeled S1 domains of myosin ( GFP-S1 ) ; 2 ) ATP; and 3 ) calcium . The thin filaments are then imaged using Oblique Angle Fluorescence microscopy to detect fluorescent GFP-S1s bound to the thin filament . With appropriate concentrations of GFP-S1 , ATP and calcium , individual fluorescent spots are observed ( Fig 2A ) . These spots can vary in brightness , indicating the binding of multiple GFP-S1s , can appear/disappear or can randomly diffuse along the thin filament . Movies of these spots were visualized in two ways , as kymographs ( Fig 2B ) and as histograms ( Fig 2D ) . To generate both a kymograph and a histogram from a movie , Desai et al . [52] performed the following steps . Pixels lying along the thin filament were isolated ( Fig 2A , red box ) . For the kymograph , the intensity along this line of pixels was plotted for each frame of the movie , generating a 2D image with position in the ordinate and time on the abscissa ( Fig 2B ) . Any GFP-S1 interactions with the thin filament appear as streaks that start when the GFP-S1 binds and end when the GFP-S1 detaches . For a histogram , the intensity of the line of pixels was plotted as a function of pixel position ( Fig 2C ) . These data were then fit with custom code [52] that determines the best-fit Gaussians of fixed standard deviation , but variable amplitude . The amplitudes of these best-fit Gaussians were determined for every frame of the movie and then plotted as a histogram ( Fig 2D ) . As concentrations of GFP-S1 , ATP and calcium were varied , clear differences are observed in both the kymographs and the histograms [52] . In order to compare our model most directly to experimental measurements , we made the following assumptions: A single excited GFP emits a constant average amount of light per unit time , e . In the movies from Desai et al . [52] , the recorded fluorescence intensity of a single GFP ( I1 ) is inversely proportional to frame rate ( f ) . Thus , a single GFP , imaged at f = 2 Hz , will be half as bright as a single GFP , imaged at f = 1 Hz . There are two sources of signal noise . The first is background noise ( e . g . electrical noise , unbound GFP-S1s , etc . ) that is uniform across the image . This noise varies , depending on experimental condition . We assume this noise is Gaussian with mean zero and standard deviation σN ( see Fig 3A ) . The second source of signal noise is due to temporal variations in GFP intensity ( e . g . flexing of the thin filament causing the center of the GFPs to deviate from the nominal average position of the thin filament , Fig 2A , red box ) . This noise is constant across all conditions . We assume this noise is Gaussian with mean zero and standard deviation σF ( see Fig 3A ) . Since GFP-S1s are coming out of solution , they can bind to either side of the actin filament . We assume each side is regulated independently [53] . Thus , each thin filament contains two independent systems of the sort shown in Fig 1C . From fits to in vitro motility , we estimated C = 11 [42] . Since C = L C / Δ L , where ΔL is the spacing between myosin molecules , this gives a value of LC = 385nm , using ΔL = 35nm for myosin molecules attached to a glass surface [54] . Here , the GFP-S1s are not constrained by being bound to a surface so they can bind to each actin monomer , spaced 5 . 5 nm apart . Thus , ΔL = 5 . 5 nm , and so C = 385 / 5 . 5 = 70 . We assume that this value is independent of calcium concentration . The value of ε is a function of calcium , and should lie between 0 . 003 ( at very low calcium ) and 1 ( at high calcium , or in the absence of regulation ) [42] . GFP-S1 myosin obeys the kinetic model described in Walcott et al . [10] , but here it is coming out of solution and the rate limiting step is no longer the weak to strong transition , but rather the initial formation of the weak-binding complex ( Fig 3B ) . Thus , the attachment rate per μm of actin is κ b 0 [ Myo ] , where [Myo] is the concentration of GFP-S1 and κ b 0 = 0 . 2 nM - 1 μ m - 1 s - 1 [52] . Since binding sites are spaced 5 . 5 nm apart along both sides of an actin filament , this corresponds to a binding rate of k b 0 = ( 0 . 2 nM - 1 μ m - 1 s - 1 ) ( 1 μ m 1000 nm ) ( 5 . 5 nm 2 binding sites ) = 0 . 00055 nM - 1 s - 1 per binding site Since myosin is coming out of solution , the myosin molecules do not apply forces on each other and so there is no mechanochemical coupling . Assumptions 1 and 2 , which posit that a single excited GFP emits , on average , a constant amount of light per unit time ( e ) and that the measured signal intensity ( I ( t ) ) is inversely proportional to frame rate ( f ) , differ from the assumptions made by Desai et al . [52] that the recorded intensity of a GFP varies depending on experimental condition . We estimate e = 450 fluorescent units/s , giving a GFP intensity of I1 = 45 at f = 10 Hz and I1 = 118 at f = 3 . 8 Hz ( see S1 Supplementary Material ) . To eliminate differences in measured signal intensity due to frame rate , we introduce a dimensionless quantity I F = I ( t ) f / e , the scaled intensity . With this definition , a cluster of n GFP-S1s generates a scaled intensity of I F = n . To model the experiments , we estimated the background noise , defined by standard deviation σN , and the variability in GFP intensity , defined by standard deviation σF . For all simulations , we used σF = 0 . 22 ( scaled intensity ) . The background noise varied depending on experimental condition , so we estimated σN for each condition . The values vary from σN = 0 . 21–0 . 36 ( scaled intensity ) , and the exact values are given in Table 1 . We can estimate overall signal noise by the root mean squared error between the measurement and the Gaussian fits . A comparison between the fitting error from a simulated experiment and from a measurement , consisting of 500 frames , suggests that our simulations have reasonably captured experimental noise ( Fig 3C ) . With these assumptions , and given a value for the local coupling parameter ε , we can perform simulations of the experiments performed by Desai et al . [52] . To do so , we used the following algorithm: For each experimental condition , we performed five simulations . To perform the simulations , we must specify the local coupling parameter ε . From assumption 7 , we have 0 . 003 < ε < 1; we can further refine this estimate by comparison to measurements in the absence of regulation ( ε = 1 ) and at pCa 4 and pCa 8 [52] . Even at high calcium ( pCa 4 ) , Desai et al . [52] observed a significant decrease in binding events in the presence of regulatory proteins compared to in their absence . We can reasonably replicate their data if we assume that 0 . 01 < ε < 0 . 5 in the presence of regulatory proteins , with the maximum and minimum values being obtained at pCa 4 and pCa 8 , respectively ( Fig 3D ) . This is consistent with the observation that , even at saturating calcium , myosin strong binding to actin induces a shift in the position of tropomyosin [55] . For each calcium concentration ( pCa 5 , 6 and 7 ) , we estimated ε by trial-and-error , with the goal being to replicate the experimental measurements ( see S1 Supplementary Material ) . Our model therefore has a single fitting parameter at each calcium concentration . The values used in our simulations are ε = 0 . 4 at pCa 5 , ε = 0 . 06 at pCa 6 and ε = 0 . 02 at pCa 7 . All model parameters are summarized in Table 2 . In the model , local coupling is determined by the parameters ε and C . These affect the rate at which a myosin molecule binds to actin , kb , according to Eq 2 . Importantly , the value of kb depends on where neighboring myosin molecules are bound , thereby defining the local coupling in the system . Besides defining this reaction rate , the parameters ε and C also relate to mechanochemical properties of the thin filament . This connection is made explicitly , and mathematical expressions are provided , in Walcott 2013 ( where the parameter ε is called δ ) [42]; here we provide a more intuitive interpretation . The parameter ε defines how the Tn-Tm beam decreases the attachment rate of an isolated myosin molecule . That is , if ε = 0 . 1 , then the regulatory proteins decrease the binding rate of an isolated myosin 10-fold . Since calcium affects how easily myosin can bind to actin , ε is a function of calcium concentration . The value of ε can be related to physical properties of the thin filament , given the assumption that attachment rate depends exponentially on the energy required to displace the Tn-Tm beam , ΔE . In particular , since ε is defined as the reduction in an isolated myosin’s attachment rate due to the regulatory proteins , the exponential dependence of attachment rate implies that limL→∞ exp ( −ΔE ) = ε ( where ΔE is measured in kB T , Boltzmann’s constant time temperature ) . Thus , ΔE asymptotes to −ln ( ε ) as L becomes large ( Fig 1D ) . The parameter C is defined as the ratio of two lengths: C ≡ L C / Δ L . The length ΔL is the spacing between neighboring myosin molecules interacting with a given actin filament , and depends only on how the myosin molecules are arranged in a given experiment . The length LC is the critical separation between myosin molecules . If two myosin molecules are bound to actin and the separation between them is greater than LC , then they induce independent deformations of the Tn-Tm beam . That is , they are uncoupled . Conversely , if two myosin molecules are bound to actin and the separation between them is less than LC , then they completely activate the intervening length of thin filament . The value of LC depends on details of how the Tn-Tm beam interacts with myosin ( W ( z ) ) , upon its mechanical properties ( i . e . its flexural rigidity [42] ) , and upon the distance that the Tn-Tm beam is deformed upon myosin binding . These likely do not change upon changes in , say , ATP concentration or the geometry of the assay . Thus , we have assumed that LC is constant between the experiments considered here and the motility assay [32 , 42 , 43] . In contrast , the interaction between the Tn-Tm beam and actin , W ( z ) , might depend on calcium , and so it is possible that C varies with calcium . However , the local coupling model reasonably reproduces experimental measurements if C is assumed to be calcium independent , and only ε varies with calcium [32] . We therefore make that assumption here , fixing C ( see Table 2 ) and allowing ε to vary in order to give the best fit between model and measurement .
We simulated nine different experimental conditions to follow Desai et al . [52] , varying the amount of GFP-S1 ( [Myo] ) , the amount of calcium , and the amount of ATP ( pCa 6 , ATP = 0 . 1μM , [Myo] = 1 , 5 , 10 , 15nM , 1000 frames collected at 10Hz; pCa 5 , 6 , 7 , ATP = 0 . 5μM , [Myo] = 15nM , 500 frames collected at 3 . 8 Hz; pCa 6 , ATP = 0 . 1 , 0 . 5 , 1μM , [Myo] = 15nM , 500 frames collected at 3 . 8Hz ) . Representative kymographs from measurements and simulation are shown in Fig 4 . In all cases , agreement between simulation and experiments is reasonable . We can make a more quantitative comparison between simulation and experiment by calculating the mean fluorescence per pixel in the kymographs . The agreement between simulation and experiments is again good ( Fig 4D shows variable [Myo] , the remaining plots are in the S1 Supplementary Material ) . Importantly , both the simulation and measurements show a non-linear dependence of mean fluorescence on [Myo] . Mean fluorescence ( in units of scaled intensity ) corresponds to binding probability and , since increasing [Myo] increases binding rate linearly , in the absence of local coupling mean fluorescence should increase , at most , linearly in [Myo] . The observation , in both experiment and simulation , that this curve has a faster-than-linear increase therefore implies the presence of local coupling in the experimental system and that the model accurately describes this local coupling . The data contain more information than just mean fluorescence; each kymograph shows local clusters of fluorescence of varying brightness . The brightness of each spot can be quantified and plotted as a histogram , using custom code developed by Desai et al . [52] . We analyzed our simulated data with this code , as shown in Fig 2D , and the results are shown in Fig 5 . With the exception of one measurement at pCa 6 , [Myo] = 15 nM , and [ATP] = 0 . 1μM ( rightmost panel in Fig 5C ) , the agreement between simulation and measurement was good . Indeed , this agreement is remarkable since , at each calcium concentration , we had only one parameter ( the binding probability: ε ) with which to fit the data . The reason that the simulations do not fit all of the experimental measurements becomes apparent upon closer inspection . There were two measurements performed at pCa 6 , [Myo] = 15 nM , and [ATP] = 0 . 1μM , one that was imaged at 10Hz ( Fig 5A , rightmost panel ) and one that was imaged at 3 . 8 Hz ( Fig 5C , rightmost panel ) . While the simulations agree with the former , they do not agree with the latter . There are several possible reasons for this inconsistency , including variability in GFP fluorescence between the two experiments , but we believe the most likely explanation is that under these conditions where GFP-S1 binding is frequent and local coupling is strong , the system is highly sensitive to parameters . For example , with only a 50 nM increase in ATP concentration , the simulations are in reasonable agreement with the measurements ( Fig 5D ) .
To support the assumption of constant GFP-S1 fluorescence intensity , at e = 450 fluorescent units/s , we reanalyzed the data from Desai et al [52] . If GFP-S1 fluorescence varied significantly , then the fluorescent spots from single GFP-S1s between conditions could not be rationally rescaled . Using I F = I f / e we rescaled the histograms of fluorescent intensity observed by Desai et al . [52] under three conditions where isolated single GFP-S1s are expected to dominate binding . These conditions are: high ATP ( pCa 6 , ATP = 1μM , [Myo] = 15nM , f = 3 . 8Hz ) , low myosin ( pCa 6 , ATP = 0 . 1μM , [Myo] = 1nM , f = 10Hz ) , and low calcium ( pCa 7 , ATP = 0 . 5μM , [Myo] = 15nM , f = 3 . 8Hz ) . The resulting histograms are similar , supporting the assumption of equal intensity ( Fig 6 ) . From this study it is clear that the local coupling model agrees well with data recently collected using single molecule fluorescence . The wider agreement of this model with bulk and physiological data suggests this model has the potential to describe and predict systems behavior at multiple levels . Further refinement of this model provided by single molecule data requires challenging one outcome from the experiments: that two heads are required to activate the thin filament . To experimentally validate or refute this outcome , the transition from one to two bound molecules must be observed . However , Desai et al . [52] used a steady-state approach to analyze their data , leaving the time-domain unexplored . Determining if two heads are required for activation can be most effectively done using fast imaging techniques with high spatial resolution . Such experiments can directly determine if the binding of a second head is required to generate activation . This would manifest as a change in the apparent second order binding rate constant once two heads are bound , with no ( or little ) change if a single head is bound . As an additional benefit , these measurements could also define model parameters . For example , determining the spatial separation between the first two heads to bind and the degree of collaboration provides a direct measure of LC , the critical length over which two myosin molecules communicate . Desai et al’s [52] conclusion that few isolated GFP-S1s are bound to an activated thin filament not only requires that binding be coordinated , but also requires that unbinding be coordinated . In support of this view , they do not clearly observe stepwise dissociation [52] . The model described in this study predicts that unbinding should occur stepwise and that single GFP-S1s should exist in activated conditions . Coordinated detachment requires that the regulatory proteins act to remove low numbers of myosin molecules , which would challenge the view that regulation affects only myosin’s attachment rate [11] . Again , high spatial and temporal imaging in these challenging conditions would provide a measurement of the process of thin filament relaxation—a process of high relevance to disease [60] . Recent experimental observations [52] of thin filament activation by single myosins were well predicted by a local coupling model described in this study . The importance of this result is that a relatively simple model of local coupling , which can be incorporated into efficient differential equation models , now fits in vitro motility data at high and low calcium [42] , in the presence of MyBP-C [43] , fiber data [32] and , as we have just shown , direct molecular-scale measurements . This validation of the model across size scales suggests that it captures the essential phenomenology of the local coupling between myosin molecules that occurs upon muscle activation . As it captures this coupling across scales , the model has the potential to provide insight into precisely how molecular level parameters affect macroscopic function—a particularly important problem , since mutations that affect local coupling have been implicated in some genetic cardiomyopathies [61] . | Despite decades of study , there is no clear connection between muscle contraction at the molecular and the macroscopic scale . For example , we cannot yet predict how a genetic defect in a muscle protein will result in a physiological change in the heart . This multi-scale understanding is difficult , in part , because molecules cooperate during muscle contraction; that is , one molecule’s behavior is influenced by the behavior of its neighbors . It is difficult to make direct measurements from such coupled molecular systems and also difficult to describe them quantitatively . Despite these obstacles , we recently published experimental measurements and theoretical models of this coupling , but there were apparent discrepancies between the two . Here , we use detailed computer simulations of these experiments to show that , in fact , the measurements agree with the model to a remarkable extent . This agreement suggests that the model captures the essential molecular events that underlie the coupling between muscle molecules . This removes a major obstacle to a multi-scale understanding of muscle contraction and , while more work is necessary , suggests that a connection between the molecular and macroscopic scale is within reach . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | Direct Measurements of Local Coupling between Myosin Molecules Are Consistent with a Model of Muscle Activation |
Rubella virus ( RuV ) infection of pregnant women can cause fetal death , miscarriage , or severe fetal malformations , and remains a significant health problem in much of the underdeveloped world . RuV is a small enveloped RNA virus that infects target cells by receptor-mediated endocytosis and low pH-dependent membrane fusion . The structure of the RuV E1 fusion protein was recently solved in its postfusion conformation . RuV E1 is a member of the class II fusion proteins and is structurally related to the alphavirus and flavivirus fusion proteins . Unlike the other known class II fusion proteins , however , RuV E1 contains two fusion loops , with a metal ion complexed between them by the polar residues N88 and D136 . Here we demonstrated that RuV infection specifically requires Ca2+ during virus entry . Other tested cations did not substitute . Ca2+ was not required for virus binding to cell surface receptors , endocytic uptake , or formation of the low pH-dependent E1 homotrimer . However , Ca2+ was required for low pH-triggered E1 liposome insertion , virus fusion and infection . Alanine substitution of N88 or D136 was lethal . While the mutant viruses were efficiently assembled and endocytosed by host cells , E1-membrane insertion and fusion were specifically blocked . Together our data indicate that RuV E1 is the first example of a Ca2+-dependent viral fusion protein and has a unique membrane interaction mechanism .
Rubella virus ( RuV ) is the causative agent of “German measles” , an airborne , relatively mild childhood disease [for review] , [ see 1 , 2] . However , RuV infection of pregnant women can have very serious consequences due to transplacental infection of the fetus , causing first trimester miscarriages and severe fetal malformations known collectively as rubella congenital syndrome [2] , [3] . Humans are the only known host for this virus . An efficient vaccine has all but eliminated the incidence of RuV disease in the Americas , although a reduction in vaccination recently led to resurgent cases [4] . RuV remains endemic worldwide and is a serious public health problem in countries without an effective vaccination program , which included 59% of the world birth cohort in 2012 [5] . RuV is the sole member of the Rubivirus genus , which together with the Alphavirus genus makes up the Togaviridae family . As for all togaviruses , RuV is an enveloped virus with a positive strand RNA genome [for overview , see 2] . RuV is roughly spherical but relatively pleomorphic , with diameters ranging from ∼60–80 nm [6]–[8] . The RuV genome is organized similarly to that of alphaviruses , and encodes the non-structural proteins p150 and p90 , which mediate viral RNA replication and transcription , and the structural proteins capsid , E2 , and E1 . The genome associates with capsid proteins to form the internal nucleocapsid core , which is enveloped by a lipid bilayer containing heterodimers of the E2 and E1 transmembrane glycoproteins . While this general organization is similar to that of the alphaviruses , cryo-electron tomography studies show that the RuV surface is organized with parallel rows of E2-E1 heterodimers rather than the T = 4 icosahedral symmetry of the alphaviruses [7] , [8] . During RuV biogenesis the structural polyprotein p110 is translated in association with the endoplasmic reticulum ( ER ) , and signal peptidase cleavage produces the capsid , E2 and E1 proteins . E2 and E1 form stable heterodimers within the ER , with E2 acting as a chaperone to promote the folding and sorting of E1 [9]–[12] . Unlike alphavirus E2 , which is processed by cellular furin to prime virus infectivity [13] , RuV E2 and E1 do not undergo a proteolytic maturation step [2] . RuV particles bud into the Golgi apparatus , traffic through the secretory pathway , and are released into the extracellular milieu [14]–[16] . Most of the RuV antigenic determinants and all of its neutralizing epitopes are located on the E1 protein [17] , [18] , which is therefore postulated to mask the E2 protein in the virus particle . During RuV entry E1 binds to cell surface receptors [19] that mediate virus uptake by clathrin-dependent [20] and/or macropinocytic [21] pathways . Similar to the alphaviruses [22] , RuV fuses in endosomes via low pH-triggered conformational changes in the E1 fusion protein [23]–[26] . However , in spite of this general outline of RuV entry , mechanistic information is relatively limited and is heavily based on analogy with the alphavirus entry and fusion pathways . The alphavirus and flavivirus membrane fusion proteins have similar structures and have been grouped together as “class II” fusion proteins [reviewed in 27] . The postfusion structure of the RuV E1 ectodomain was recently determined [28] , revealing it to be a class II fusion protein despite the lack of amino acid sequence conservation with the alpha- or flavivirus fusion proteins . Similar to other class II proteins , RuV E1 is composed of three β-sheet-rich domains: a central domain I ( DI ) connecting on one side to the elongated DII , and on the other to the Ig-like DIII followed by the stem and transmembrane ( TM ) regions [28] . As for the postfusion forms of other class II proteins , postfusion RuV E1 is a stable homotrimer with a central trimer core composed of DI and DII , against which the DIII and stem regions fold back to form an overall hairpin-like configuration . Extensive studies of class II proteins demonstrate that fusion is driven by the initial insertion of the fusion loop ( FL ) at the tip of DII into the target membrane , followed by refolding to the hairpin thus bringing the viral and target membranes together [reviewed in 27] , [29] . In addition to these similarities , the structure of RuV E1 revealed a puzzling difference from other class II fusion proteins , and indeed from all other known viral fusion protein structures [28] . Rather than the single FL of other class II fusion proteins , RuV E1 contains two FLs at the tip of DII [30] ( FL1 , residues 88–93; FL2 , residues 131–137 ) , making a more extensive surface for target membrane interaction . Other fusion proteins such as those of the baculoviruses , herpesviruses , and rhabdoviruses have also been shown to contain two FLs [30]–[32] . However , in the case of RuV , a metal binding site is located between the two FLs . The E1 structure shows that this site coordinates either Na+ or Ca2+ ions via interactions with residues N88 in FL1 and D136 in FL2 [28] . Ca2+ binds with apparent higher affinity and displaces Na+ . Ca2+ binding affects the conformation of FL2 around D136 with respect to the Na+ bound form , but does not cause other changes in E1 . No viral fusion protein has ever been observed to require a metal ion for its function . Here we addressed the role of Ca2+ in RuV infection . We found that Ca2+ was critical for productive viral entry . The absence of Ca2+ caused a block in E1-target membrane insertion and thus inhibited virus-membrane fusion . Similarly , alanine substitution of the Ca2+-coordinating residues N88 and D136 blocked virus membrane interaction , fusion and infection . Our observations thus document the first described example of a Ca2+-dependent viral fusion protein .
The importance of calcium during RuV entry was tested by pre-binding virus to target Vero cells on ice and then incubating at 37°C in medium containing various concentrations of CaCl2 . To avoid possible effects of calcium deprivation on endosomal acidification [33] , the internalization period was limited to 20 min . Cells were then incubated in growth medium containing NH4Cl to neutralize endosomal pH , and infected cells scored at 48 h post-infection by immunofluorescence ( Fig . 1A ) . RuV infection was strongly dependent on the concentration of Ca2+ , with infection efficiency decreased by more than 90% in the presence of 30 µM CaCl2 vs . 2 mM CaCl2 . Infection by the low pH-dependent alphavirus Semliki Forest Virus ( SFV ) was independent of Ca2+ under these conditions , excluding aberrant endosomal acidification . Addition of CaCl2 after the 20 min internalization period but prior to endosomal neutralization by NH4Cl did not cause significant rescue of RuV infection ( p>0 . 05 ) ( Fig . 1B ) . This result suggests that the internalized RuV particles were inactivated by endosomal acidity during the 20 min uptake period . While the lack of rescue could be due to several factors such as rapid acid-inactivation and/or an inability of added calcium to access the virus-containing endosomes , it demonstrates that Ca2+ deprivation did not simply delay RuV endocytosis . Together our data support a critical and specific role of Ca2+ during productive RuV entry . Given the location of bound Ca2+ between the two E1 FLs [28] , we hypothesized that it plays a role in RuV fusion with target membranes , a complex process not yet fully understood . Although RuV is known to require acidification to initiate fusion [23] , [24] , its pH threshold has not been determined . A fusion-infection assay previously developed for alphaviruses [34] was used to measure this precisely . RuV was prebound to Vero cells on ice , pulsed for 3 min at 37°C with medium of defined pH , and the resultant infected cells scored by immunofluorescence ( Fig . 2A and B ) . No infection was detected upon treatment at pH 7 . 0 , indicating that virus fusion in endosomes does not occur within this time frame . Infection was observed after treatment at pH 6 . 2 , and gradually decreased at lower pH values , presumably reflecting virus inactivation by acidic pH . The kinetics of RuV fusion were then determined by pulsing bound RuV at pH 7 . 0 or 6 . 0 for various times ( Fig . 2C ) . Infection increased with time of treatment at pH 6 . 0 , reaching a maximum at 4 min . Infection from pH 7 . 0 incubation was first detected after 5 min and increased thereafter , reflecting the time required for virus endocytic uptake and endosomal fusion . Together our data indicate that the RuV fusion reaction is optimal at 4 min of treatment at ∼pH 6 . 0–6 . 2 at 37°C . We then used this assay to determine the role of Ca2+ in RuV fusion ( Fig . 3A ) . Treatment for 4 min at neutral pH did not allow RuV fusion whether Ca2+ was present or not . In contrast , fusion at pH 6 . 0 or 5 . 8 was strongly dependent on Ca2+ , with increases of up to 46-fold with 2 mM CaCl2 compared to calcium-free treatment . Addition of the calcium chelator EDTA completely abrogated fusion at both pH 5 . 8 and 6 . 0 . Treatment at very low pH ( 5 . 0 ) did not rescue fusion in the absence of Ca2+ ( Fig . S1 ) , indicating that the fusion defect is not due to an acid-shift in the RuV pH threshold . A similar calcium dependence for RuV fusion was observed using HeLa cells or primary human umbilical vein endothelial cells as targets ( Fig . S2 ) . As predicted [34] , [35] , SFV fusion was strongly pH-dependent but calcium-independent ( Fig . 3B ) . To control for a possible effect of calcium on RuV binding to target cells , we performed fusion-infection assays in which the binding and/or fusion steps were carried out in the presence of EDTA or CaCl2 ( Fig . 3C ) . As observed also in Fig . 3A , fusion was efficient when Ca2+ was present during both steps , and totally impaired when EDTA was included in the fusion media . However , even when RuV was prebound in presence of EDTA , it fused efficiently when subsequently treated at low pH in the presence of Ca2+ . Together our data support a specific role for Ca2+ during virus fusion , and the lack of a Ca2+ requirement for virus-cell binding . Any effects of Ca2+ chelation on the virus were reversible by adding back Ca2+ during the fusion step . To more rigorously determine the Ca2+ concentration requirement for RuV fusion , we performed fusion-infection assays in which Ca2+ was buffered with EGTA , a more selective Ca2+ chelator . RuV fusion in this assay was maximal at 2 mM CaCl2 and showed a gradual reduction at decreasing Ca2+ concentrations ( Fig . 4A ) . Mg2+ , Mn2+ and Zn2+ did not substitute for Ca2+ even at concentrations of 2 mM ( Fig . 4A , Fig . S3 ) . When fusion was triggered using a sub-optimal concentration of Ca2+ ( 0 . 5 mM ) , the addition of Mg2+ at concentrations up to 20 mM neither substituted nor competed with Ca2+ ( Fig . 4B ) . The structure of the RuV E1 homotrimer shows that the bound Ca2+ is coordinated in part by an acetate ion [28] . However , when tested in the fusion-infection assay CaCl2 and Ca ( C2H3O2 ) 2 promoted fusion with a similar concentration dependence ( Fig . 4A ) , implying that acetate is not critical for the role of Ca2+ during fusion . Low pH triggers the irreversible rearrangement of the prefusion RuV E2-E1 heterodimer to the E1 hairpin-like homotrimer , driving fusion of the virus with a target membrane [24] , [28] or virus inactivation in the absence of target membranes [24] . To determine if the role of Ca2+ in fusion is via effects on this conformational change , we first examined the effect of Ca2+ on low pH inactivation [36] . RuV particles were immobilized on poly-D-lysine-coated wells , treated with neutral or low pH buffers in the presence or absence of CaCl2 or EDTA , overlaid with Vero cells and incubated for 48 h in growth media to quantitate virus infectivity ( Fig . 5A and B ) . Fluorescence microscopy confirmed that buffer treatments did not cause elution of the adsorbed virus from the wells ( Fig . 5A , Fig . S4 ) . Virus infectivity was maintained after treatment at neutral pH whether or not Ca2+ or EDTA was present , but low pH caused virus inactivation under either condition . To directly test the role of Ca2+ in the RuV E1 conformational change , we used a previously described assay [24] to compare the trypsin-resistance of prefusion vs . postfusion E1 . Radiolabeled RuV was treated at neutral vs . low pH , solubilized in non-ionic detergent , digested with trypsin , and precipitated with a monoclonal antibody ( mAb ) to E1 ( Fig . 5C and D ) . Non-digested samples treated at either pH showed efficient co-precipitation of the E2-E1 dimer . While trypsin completely digested the samples treated at pH 8 . 0 , three distinct species were immunoprecipitated from the trypsinized low pH-treated samples: a tight doublet migrating slightly faster than uncleaved E1 ( ∼55 kDa ) and a band of lower molecular weight ( ∼41 kDa ) . These species were immunoprecipitated by mAb to the E1 protein but not the E2 protein ( Fig . S5 ) , confirming that they were derived from digestion of the E1 protein . Low pH-triggered production of trypsin-resistant E1 was unaffected by the presence of either CaCl2 or EDTA . Thus , Ca2+ is not required for the low pH-triggered conformational change that leads to RuV inactivation and E1 trypsin-resistance . An early step in viral membrane fusion pathways is the insertion of the fusion peptide or fusion loops into the target membrane . This step confers virus-target membrane association and can be monitored by testing the cofloatation of virus with target liposomes [37] . To test the role of Ca2+ in E1-membrane insertion , radiolabeled RuV was mixed with liposomes and treated at neutral or low pH in the presence of EDTA or CaCl2 . Liposomes plus associated virus were separated on sucrose floatation gradients ( Fig . 6A ) . Little virus was associated with the floated liposomes when treated at either pH in the absence of Ca2+ . However , efficient virus-liposome interaction ( ∼74% of input virus ) was observed when low pH treatment was carried out in the presence of 2 mM CaCl2 . Even at pH 8 . 0 , 2 mM CaCl2 caused a small increase ( ∼10% ) in virus-liposome association . Once the virus was associated with liposomes , stable interaction was maintained even though the gradients did not contain Ca2+ . Ca2+ did not promote non-specific electrostatic interactions of virus with membranes , since SFV-liposome floatation was strictly low pH-dependent and unaltered by the presence or absence of CaCl2 ( Fig . 6B ) . Our data indicate that complete RuV fusion is Ca2+-dependent . To test whether initial lipid mixing between RuV and cell membranes can occur in the absence of calcium , we labeled RuV with a self-quenching concentration of the lipophillic dye DiD . Virus was prebound to Vero cells at 4°C , and then samples were pulsed for 2 min at 37°C with medium at pH 6 . 0 or 7 . 0 and containing 2 mM CaCl2 or EDTA as indicated . Lipid mixing was assessed by measuring the increase in fluorescence that resulted from the dequenching of the DiD probe ( Fig . 6C and 6D ) . Upon pulsing at pH 7 . 0 , only background fluorescence was observed . In contrast , a strong homogeneous signal was detected at the surface of the cells after treatment at pH 6 . 0 in presence of CaCl2 . This low pH-triggered fusion was blocked by omission of CaCl2 or the addition of EDTA . We then tested whether calcium was also required for fusion and lipid mixing in endosomes . DiD-labeled RuV was prebound to Vero cells and virus uptake allowed to occur for 20 min at 37°C in medium containing 2 mM CaCl2 or 20 mM NH4Cl . Cells were immediately fixed after the internalization period and analyzed by confocal microscopy . Virus fusion in endosomes resulted in numerous bright fluorescent foci when uptake occurred in presence of CaCl2 ( Fig . 6E–D ) . DiD dequenching was severely impaired by preventing endosomal acidification with NH4Cl , by preventing endocytic uptake by incubation at 4°C , or by removing calcium from the media during virus internalization . Together , our data support a critical role of calcium for RuV E1-membrane interaction and for the initial lipid-mixing stage of fusion . The RuV E1 trimer structure shows that Ca2+ is coordinated by residues N88 and D136 in FL 1 and 2 , respectively [28] ( Fig . 7A ) . Such polar/charged residues are commonly involved in coordination of Ca2+ , but are usually absent from fusion peptides [38] , having the potential to hinder their insertion into the target membrane . We tested whether we could overcome the RuV calcium requirement by substituting these residues with alanine , an uncharged residue that is common in fusion loops [38] . Mutant viral RNAs were transcribed from the RuV infectious clone and electroporated into BHK-21 cells . Cells electroporated with WT RuV RNA produced infectious particles as soon as 24 h , and reached a maximum titer of 1×107 IC/ml at 72 h post-electroporation ( Fig . 7B ) . In contrast , neither the single N88A or D136A mutants nor the double mutant N88A , D136A produced detectable infectious virus under the same conditions . Equivalent expression of E1 , E2 and capsid was detected in WT vs . mutant-infected cells , ruling out a problem in mutant viral protein expression ( Fig . 7C ) . The amount of virus particles pelleted from the medium at 48 h post-electroporation was comparable between WT and mutants ( Fig . 7C ) . The efficiency of particle release ( normalized to expression levels in the cell lysates ) was equivalent between WT and mutants ( Fig . 7D ) . Thus , mutation of the Ca2+-coordinating residues in RuV E1 does not overcome the requirement for calcium . Instead , these mutations are lethal to RuV infection without affecting virus synthesis and assembly . Fusion-infection assays of the mutants revealed that their infectivity was not rescued even by low pH-treatment in the presence of 20 mM CaCl2 ( Fig . S6 ) . We therefore tested specific steps of entry and fusion for the N88A , D136A double mutant . Binding studies showed that the efficiency of radiolabeled WT and mutant virus binding to Vero cells on ice was comparable ( Fig . 8A ) . Co-immunoprecipitation analysis of solubilised virus showed that both the WT and mutant E1 proteins were efficiently recognized by a commercial mAb to RuV E1 and that similar levels of E2 were retrieved , indicating comparable stability of the E2-E1 dimer interaction ( Fig . 8B ) . Together with the efficient assembly of the mutant ( Fig . 7D ) , these results thus suggest that the folding of N88A , D136A E1 is not aberrant . To test the possibility that these mutations compromise the low pH-triggered structural reorganization of the E1 protein , we evaluated the conversion of E1 N88A , D136A to trypsin-resistance ( Fig . 8C ) . Both WT and mutant E1 were totally digested by trypsin after treatment at neutral pH , consistent with a similar heterodimeric structure . After low pH treatment similar proportions of the mutant and E1 WT were trypsin-resistant ( Fig . 8C ) . We then compared the membrane interaction of WT virus and the N88A , D136A mutant ( Fig . 8D ) . While low pH treatment in the presence of Ca2+ produced efficient liposome association of the WT virus , the mutant virus showed only background levels of liposome floatation at either pH . Thus , our data indicate that the N88A , D136A mutant virus was comparable to the WT in all properties save for the ability of mutant E1 to interact with the target bilayer , a critical step in using the energy of E1 homotrimer formation to drive the fusion of the virus and target membranes . The properties of the E1 N88A , D136A mutant completely recapitulated the defect in the fusion of WT RuV in the absence of calcium , thus confirming the importance of Ca2+ coordination during the RuV fusion process . Together our results suggest that the primary role of RuV E1 calcium binding is not to compensate for the negatively charged D136 residue , but rather a more specific role in promoting E1-membrane interaction .
Here we demonstrated that Ca2+ is strictly required for WT RuV liposome association , membrane fusion , and virus infection . A similar specific block resulted when the key Ca2+ binding residues E1 N88 and/or D136 were substituted with alanine . In contrast , RuV binding to the cell surface , endocytic uptake , and E1 conformational changes did not require Ca2+ coordination . An early study showed an effect of calcium in cell-cell fusion mediated by HIV-1 , but the mechanism of this effect is undefined [39] . Cellular calcium levels are known to be critical for infection by non-enveloped viruses such as polyomavirus and rotavirus [40] , [41] . However , in these examples Ca2+ stabilizes the virus core and the low Ca2+ concentration of the cytosol promotes uncoating during virus entry . In contrast , WT RuV particles were stable in the absence of Ca2+ and the N88 and D136 mutants showed normal virus assembly and release . Our results thus define a specific requirement for Ca2+ during RuV fusion . To our knowledge , RuV E1 represents the first example of a Ca2+-dependent virus fusion protein . The low pH-triggered interaction of RuV with liposomes was inhibited in the absence of Ca2+ , reflecting the Ca2+-dependence of E1 insertion into the target membrane . In contrast to our results with virus particles , association of the purified RuV E1 homotrimer with target membranes occurs at either neutral or low pH in the presence of Ca2+ , and at low pH in the absence of Ca2+ [28] . Although Dubois et al . hypothesized that calcium could be involved in direct attachment of the virus to the target cell lipid bilayer at neutral pH [28] , our data demonstrated that Ca2+ is not required for the binding of virus to cell surface receptors , while stable E1-membrane interaction requires both low pH and Ca2+ . The differences between our results and those of DuBois et al . probably reflect several differences between experimental systems , including the use of the postfusion E1 ectodomain homotrimer , overnight pH treatment , and target membranes containing negatively-charged lipids . Nevertheless , despite these differences , both studies are consistent with an important role of calcium in E1-membrane interaction . Our data demonstrate that RuV encounters low pH and a calcium concentration compatible with fusion during virus transit through the endocytic pathway ( Fig . 6E ) . This is intriguing because both pH and Ca2+ concentration are reported to fluctuate in this pathway . While the Ca2+ concentration is high in the extracellular milieu , one report indicates that the level of calcium drops to concentrations ≤10 µM in early endosomes [33] due to the presence of pH-dependent Ca2+ channels that compensate for the influx of protons into the endosomal lumen [42] . The Ca2+ concentration then appears to increase to 500 µM in late endosomes/lysosomes [43] . How might RuV fuse in the environment of the endocytic pathway ? Our understanding of the Ca2+ concentration in the endocytic pathway is currently incomplete , and endosomes themselves are heterogeneous organelles with distinct characteristics [44] . The simplest explanation would be that virus-containing vesicles reach a proper pH for RuV fusion before the Ca2+ concentration becomes suboptimal . Alternatively , Ca2+ could remain bound to RuV E1 for sufficient time to promote fusion . Our data indicate that removal of calcium immediately prior to low pH treatment impaired fusion at the plasma membrane , in keeping with the relatively weak nature of Ca2+-E1 binding [28] . However , calcium binding might be stabilized by coordination by a lipid headgroup , replacing the acetate ion observed in the crystal structure . Alternatively , RuV might first encounter low pH in the early endosome but fuse in a later endosomal or lysosomal compartment that has a higher Ca2+ concentration [43] , [45] . However , this model does not agree with the rapid inactivation we observed when RuV was exposed to low pH in the absence of Ca2+ . The Ca2+ concentration required for virus fusion in the endosome might differ from those we observed in vitro , perhaps due to differences in lipid composition , luminal ionic environment , or membrane-associated Ca2+ . Our limited knowledge of endosomal calcium concentration relies on imprecise probes that detect luminal calcium , while the critical concentration of calcium is presumably that at the endosomal membrane . It is even possible that the virus somehow increases the Ca2+ concentration of the endosome compartment , as there are examples of viruses that alter cellular calcium levels during entry [e . g . , 46] . While more work will be required to define the intracellular site and Ca2+ conditions for RuV fusion in vivo , our data clearly demonstrate that both RuV fusion and infection require calcium . The calcium-dependent RuV-lipid interaction is reminiscent of other calcium-dependent lipid binding proteins such as the TIM family [28] , [47] and the synaptotagmin proteins [48] , [49] . Bound Ca2+ can enable these proteins to interact with negatively-charged lipids such as phosphatidylserine [50]–[52] . However , to date our results suggest that rather than inducing binding of negatively charged lipids , Ca2+ plays a different role in the function of RuV E1 . The liposomes used in our membrane interaction studies did not contain negatively-charged lipids , and efficient RuV-plasma membrane fusion occurred in spite of the scarcity of negatively-charged lipids in the outer leaflet of the PM [53] . E1's Ca2+ binding does not act simply to compensate for N88 polarity and D136 negative charge , since their substitution with apolar alanine residues did not rescue RuV membrane interaction or fusion . The two FLs are presumably optimized to form a structurally specific Ca2+ binding site . Indeed , while Na+ binds to the two FLs , Na+ did not permit RuV-membrane interaction or fusion . The Na+ bound conformation of FL2 differs significantly from that of Ca2+ bound FL2 , including an important rotation of D136 [28] . These differences in conformation upon Ca2+ binding could align the two FLs in the correct orientation for optimal membrane insertion . It is also important to note that the effect of Ca2+ on the conformation of the FLs could differ in target membrane-inserted RuV E1 vs . the crystal structures , which were determined in solution . The calcium requirement is not simply a consequence of the presence of two FLs , since viral fusion proteins such as vesicular stomatitis virus G protein and herpes simplex virus gB also have two FLs , but function efficiently without having a metal binding site [31] , [32] . Thus the mechanistic role of Ca2+ in RuV fusion , while clearly critical , is as yet unclear , and it will be important to address the effect of Ca2+ on the membrane-inserted conformation of the FLs . To date , RuV E1 is the only viral fusion protein that is known to require metal binding for its function . Within the Togaviviridae and Flaviviridae , RuV is unique in having two FLs , a Ca2+ requirement , and humans as the only natural host . The selective advantage of the unique RuV fusion mechanism and the nature of the evolutionary pressure that produced it are unclear . Presumably these relate to the key step of fusion protein-membrane insertion , a process that is not well understood for any viral fusion protein . Definition of the role of Ca2+ in this process thus will help to advance our understanding of the overall function of viral fusion proteins and may prove useful for their therapeutic control .
Vero and HeLa cells from ATCC were a kind gift of Dr . Kartik Chandran , and were maintained at 37°C in growth medium: Dulbecco's modified Eagle's medium ( DMEM ) containing 5% fetal bovine serum , 1% L-glutamine , 100 U penicillin/ml and 100 µg streptomycin/ml . BHK-21 cells were cultured in growth medium with the addition of 10% tryptose phosphate broth . Primary human umbilical vein endothelial cells were obtained from a commercial source ( Lonza , NJ ) and maintained according to manufacturer's instructions in Endothelial Cell Growth Medium ( Lonza , NJ ) . The RuV M33 infectious clone ( pBRM33 ) [54] was a kind gift from Dr . Tom Hobman ( University of Alberta , Edmonton , Canada ) . In vitro mutagenesis was performed using intermediate plasmids derived from pGEM-5zf ( - ) plasmid ( Promega ) . pGEM33-SphI/BamHI and pGEM33-XbaI/HindIII ( in which the BamHI in the multiple cloning site was removed by site-directed mutagenesis ) contained the corresponding fragments from the M33 clone . pGEM33-SphI/BamHI was used as a template for circular site-directed mutagenesis [55] , cut with SphI/BamHI and ligated into pGEM33-XbaI/HindIII . The mutated Xba/HindIII fragments were then subcloned into pBRM33 to generate the final RuV E1 N88A , D136A , and N88A , D136A infectious clones . Sequences were confirmed by automated DNA sequencing . The SFV strain used was a well-characterized plaque-purified isolate propagated in BHK-21 cells [56] . Purified radiolabeled SFV was derived from the pSP6-SFV4 infectious clone [57] . RuV viral RNAs were in vitro transcribed from pBRM33 and electroporated into BHK-21 cells to generate virus stocks [57] . Culture supernatants were harvested 48–72 h post-electroporation , supplemented with 10 mM Hepes pH 7 , clarified by centrifugation at 1000× g for 5 min and frozen at −80°C in aliquots . RuV titers were determined by a modified infectious center assay [58] . Briefly , serial dilutions of viruses were incubated with Vero cells for 4 h at 37°C , a time point that permits maximal RuV entry , and then cultured for 48 h in the presence of 20 mM NH4Cl to block secondary infection . This time point was chosen because of the relatively slow replication kinetics of RuV . Infected cells were quantitated by immunofluorescence and titers calculated as infectious centers ( IC ) /ml . To produce radiolabeled RuV , BHK-21 or Vero cells were electroporated with in vitro transcribed viral RNA [57] . At 24 h post-electroporation , infected cells were radiolabeled with 150 µCi/ml [S35]-methionine/cysteine for 24 h at 37°C . Supernatants were then harvested , clarified and pelleted at 210 , 000× g for 3 h at 4°C . The virus pellet was resuspended in 50 mM Tris pH 7 . 4 , 100 mM NaCl and purified on a continuous Pfefferkorn sucrose gradient [59] and stored in aliquots at −80°C . Purified WT RuV was still infectious , reaching titers of up to 1×108 IC/ml . As described previously [34] , cells were washed in PBS , fixed in 4% paraformaldehyde for 20 min , then permeabilized 5 min in PBS containing 0 . 2% Triton X-100 . Permeabilized cells were washed once , then incubated 2 h at 37°C in PBS containing 5% BSA and an anti-RuV ( LifeSpan BioSciences , Inc ) or an anti-SFV envelope protein serum [60] . Labeled cells were washed twice , and then incubated at room temperature for 30 min in PBS containing secondary antibodies and 1 µg/ml Hoechst ( Invitrogen ) to counterstain nuclei . Stained cells ( ≥5 fields/condition ) were imaged by conventional epifluorescence microscopy using a Zeiss Cell Observer system ( Zeiss ) equipped with an Axiovert 200M microscope and 10× objective . The proportion of infected cells was evaluated by automated counting of infected cells versus nuclei using the CellProfiler cell image analysis software ( www . cellprofiler . org/ ) . Multiplicity was adjusted to be in the linear range . RuV ( MOI = 1 ) or SFV ( MOI = 10 ) was bound on ice to pre-chilled Vero cells for 1 . 5 h in binding medium ( RPMI without NaBicarbonate , plus 0 . 2% BSA and 10 mM Hepes , pH 7 . 0 ) . Unbound viruses were removed by washing , and cells were shifted to 37°C to permit entry for 20 min in calcium-free media ( calcium-free MEM , Sigma-Aldrich ) supplemented with 10 mM Hepes pH 7 , 0 . 2% BSA and the indicated concentrations of CaCl2 . The cells were then incubated in growth media ( with calcium ) containing 20 mM NH4Cl for 48 h at 37°C , and infected cells were scored by immunofluorescence . Alternatively , after binding on ice the cells were shifted to 37°C for 20 min+/−2 mM CaCl2 , followed by an additional 1 h incubation at 37°C+/− CaCl2 to test if infection in calcium-free media can be rescued by the subsequent addition of Ca2+ . Cells were then cultured in growth media containing 20 mM NH4Cl and scored by immunofluorescence 48 h post-infection . Similar to a published assay [34] , RuV ( MOI 1–5 ) or SFV ( MOI 2 . 5–5 ) was prebound for 1 . 5 h on ice to Vero cells in binding medium . Cells were washed and incubated at 37°C for the indicated time in fusion medium of the indicated pH ( either RPMI or calcium-free MEM , without bicarbonate , plus 0 . 2% BSA and 20 mM MES ) . For experiments testing pH values lower than 5 . 8 , fusion media was buffered with a combination of 10 mM MES and 10 mM Na-succinate . After the pH pulse cells were washed and incubated in growth medium containing 20 mM NH4Cl for 48 h at 37°C , and scored by immunofluorescence . To test defined concentrations of free Ca2+ or Mg2+ , fusion buffer was composed of 20 mM MES , 130 mM NaCl , 5 mM glucose , 0 . 2% BSA and 1 mM EGTA plus added CaCl2 or MgCl2 , and the concentration of free Ca2+ or Mg2+ in presence of EGTA was calculated using Webmaxc web tool ( http://www . stanford . edu/~cpatton/webmaxcE . htm ) . Similar to a previously described protocol [36] , RuV was spun at 4°C for 1 h at 1500× g onto poly-D-lysine treated coverslips . Coverslips with immobilized virus were washed in PBS , and treated with RPMI-based fusion medium at the indicated pH for 15 min at 37°C . A single cell suspension of Vero cells in growth medium was overlaid onto the washed coverslips and centrifuged at 200× g for 5 min . Cells were transferred to 37°C for 1 h to permit virus entry , incubated for 48 h in growth medium plus 20 mM NH4Cl , and scored by immunofluorescence . Note that the initial amount of virus added to the coverslips corresponded to an MOI of 1 . The low pH-triggered conformational change in RuV E1 was quantitated by trypsin-resistance essentially as previously described [24] . In brief , purified radiolabeled RuV was treated at pH 8 or 6 in the presence of 2 mM CaCl2 or 1 . 5 mM EDTA for 5 min at 37°C . Samples were then adjusted to pH 8 , solubilized with 1% NP40 , and digested with 125 µg trypsin/ml for 30 min at 37°C . Digestion was quenched by addition of a final concentration of 125 µg/ml soybean trypsin inhibitor , 1 mM PMSF and 0 . 1 TIU aprotinin/ml . Trypsin-resistant E1 was recovered by immunoprecipitation [60] using a mAb to RuV E1 ( Meridian LifeScience ) . Samples were boiled in LDS sample buffer under reducing conditions , resolved by SDS-PAGE and analyzed by autoradiography . The E1-E2 dimer was also assessed by immunoprecipitation of radiolabeled virus with E1 or E-specific mAbs ( Meridian LifeScience ) . Liposomes composed of a 1∶1∶1∶1 . 5 molar ratio of phosphatidylcholine ( egg yolk ) , phosphatidylethanolamine ( egg yolk ) , sphingomyelin ( bovine brain ) and cholesterol ( Avanti Polar Lipids , Alabaster , Ala ) were prepared in MES saline ( 20 mM MES pH 7; 130 mM NaCl ) by freeze-thaw and extrusion through 400 nm polycarbonate filters [61] . Low pH-triggered virus-liposome association was determined based on a published method [59] . Briefly , radiolabeled RuV was mixed with 0 . 2 mM liposomes plus 1 . 5 mM EDTA or 2 mM CaCl2 where indicated and treated at pH 8 . 0 or 6 . 0 for 1 min at 37°C . Samples were adjusted to pH 8 . 0 and 40% sucrose , layered onto a 60% sucrose cushion and overlaid with 25% and 5% sucrose ( all percentages are wt/vol ) in 50 mM Tris pH 8 and 100 mM NaCl . Gradients were centrifuged at 210 , 000× g for 2 h at 4°C in the TLS-55 rotor , and fractionated into seven equal fractions . Virus-liposome co-floatation was calculated as the percentage of the total virus radioactivity recovered in the top two liposome-containing fractions . Total recovery of virus radioactivity was ∼95% . Purified RuV ( 1 . 2 µg ) was labeled with a self-quenching concentration of the DiD lipid probe . Virus was diluted in 48 µl of buffer ( 5 mM HEPES , 150 mM NaCl , 0 . 1 mM EDTA pH 7 ) , 2 µl 0 . 25 mM DiD ( Invitrogen ) in DMSO was injected while mildly vortexing and the mixture was incubated at room temperature for 10 min . Labeling did not cause a significant decrease in virus infectivity . Labeled particles were diluted in binding medium , and 150 ng virus was incubated for 1 h on ice with Vero cells plated in 8 well Lab-Tek chambers ( Thermo Scientific , Waltham , MA ) . To trigger fusion at the plasma membrane , cells were then directly treated for 2 min at 37°C in fusion media at the indicated pH , supplemented with 2 mM CaCl2 or 2 mM EDTA . To allow fusion in endosomes , cells were washed and then incubated for 20 min at 37°C or 4°C in MEM fusion medium containing 2 mM CaCl2 as indicated . In parallel , 1 set of cells was pre-incubated with 20 mM NH4Cl for 2 h at 37°C and maintained with NH4Cl throughout the experiment . Following treatments , cells were washed , fixed in 4% paraformaldehyde , and nuclei were counterstained with Hoechst . Samples were analyzed using a SP5 Leica confocal microscope equipped with an Acousto-Optical Beam Splitter ( Einstein Analytical Imaging Center ) . Several 0 . 25 µm Z-stacks from several fields were taken using exactly the same settings . A maximal Z-stack projection was generated from which the DiD signal from ≥200 cells was quantified using ImageJ ( NIH , http://imageJ . nih . gov/ij ) . The data are presented as the average amount of DiD signal calculated per nucleus enumerated . BHK-21 cells were electroporated with RuV RNAs , and then seeded in four 35 cm wells and cultured in growth media for 72 h . Culture medium was harvested at the indicated times and infectious particles titered by infectious center assay , all as described above . The efficiency of particle assembly was determined at 48 h post-electroporation by pelleting half of the culture medium at 210 , 000× g for 3 h at 4°C and lysing the cells . Proteins from both samples were resolved by SDS-PAGE and analyzed by Western-blot using an anti-RuV serum ( Meridian Life Science ) . Virus particle release was evaluated by comparing the ratio of the virus and cell lysate E1 , E2 and capsid bands using the Odyssey Infrared system ( Li-Cor ) . Binding of radiolabeled RuV to Vero cells was determined as previously described [59] , using a 1 . 5 h incubation in binding medium on ice . | Rubella virus ( RuV ) is a small enveloped RNA virus causing mild disease in children . However , infection of pregnant women can produce fetal death or congenital rubella syndrome , a constellation of severe birth defects including cataracts , hearing loss , heart disease and developmental delays . While vaccination has greatly reduced disease in the developed world , rubella remains prevalent in developing countries and other undervaccinated populations . RuV infects cells by endocytic uptake and a low pH-triggered membrane fusion reaction mediated by the viral E1 protein . The postfusion structure of E1 revealed a metal ion complexed at the membrane-interacting tip of the protein . Here we demonstrated that RuV infection and fusion are completely dependent on calcium , which could not be replaced functionally by any other metal that was tested . In the absence of calcium , RuV entry and low pH-conformational changes were unchanged , but E1's interaction with the target membrane was specifically blocked . Mutations of the calcium-binding residues in E1 caused a similar inhibition of E1 membrane interaction , fusion and infection . Thus , RuV E1 is the first known example of a calcium-dependent virus fusion protein . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"infectious",
"diseases",
"medicine",
"and",
"health",
"sciences",
"viral",
"transmission",
"and",
"infection",
"virology",
"biology",
"and",
"life",
"sciences",
"microbiology",
"viral",
"diseases",
"viral",
"structure",
"rubella"
] | 2014 | Rubella Virus: First Calcium-Requiring Viral Fusion Protein |
The diagnosis of Human African Trypanosomiasis relies mainly on the Card Agglutination Test for Trypanosomiasis ( CATT ) . While this test is successful , it is acknowledged that there may be room for improvement . Our aim was to develop a prototype lateral flow test based on the detection of antibodies to trypanosome antigens . We took a non-biased approach to identify potential immunodiagnostic parasite protein antigens . The IgG fractions from the sera from Trypanosoma brucei gambiense infected and control patients were isolated using protein-G affinity chromatography and then immobilized on Sepharose beads . The IgG-beads were incubated with detergent lysates of trypanosomes and those proteins that bound were identified by mass spectrometry-based proteomic methods . This approach provided a list of twenty-four trypanosome proteins that selectively bound to the infection IgG fraction and that might , therefore , be considered as immunodiagnostic antigens . We selected four antigens from this list ( ISG64 , ISG65 , ISG75 and GRESAG4 ) and performed protein expression trials in E . coli with twelve constructs . Seven soluble recombinant protein products ( three for ISG64 , two for ISG65 and one each for ISG75 and GRESAG4 ) were obtained and assessed for their immunodiagnostic potential by ELISA using individual and/or pooled patient sera . The ISG65 and ISG64 construct ELISAs performed well with respect to detecting T . b . gambiense infections , though less well for detecting T . b . rhodesiense infections , and the best performing ISG65 construct was used to develop a prototype lateral flow diagnostic device . Using a panel of eighty randomized T . b . gambiense infection and control sera , the prototype showed reasonable sensitivity ( 88% ) and specificity ( 93% ) using visual readout in detecting T . b . gambiense infections . These results provide encouragement to further develop and optimize the lateral flow device for clinical use .
Human African Trypanosomiasis ( HAT ) , also known as Sleeping Sickness , is a disease caused by Trypanosoma brucei gambiense and T . b . rhodesiense [1] , [2] , [3] . The parasites are transmitted in sub-Saharan Africa by the bite from an infected tsetse fly . HAT is of great public health significance , with epidemic outbreaks recorded several times over the past century with , at times , estimates of 300 , 000 or more infected individuals [4] . Today , the recorded number of new cases has dropped below 10 , 000 per year , yet HAT still continues to place a large burden on individuals and communities in terms of disability-adjusted life years [5] , [6] . The identification of infected individuals is crucial for therapeutic and public health intervention . New tools could aid eradication of this disease when used in coordination with other efforts [5] , [7] , [8] . Infection with T . b . gambiense or T . b . rhodesiense progresses through two defined stages . The first stage is when trypanosomes are limited to the blood and lymphatic systems . The second stage occurs when the parasites invade the central nervous system [2] . The latter leads to neurological damage , sleep cycle disruption , coma and death if the patient does not receive treatment [9] , [10] , [11] . The two stages are treated with different drugs , and those used for the second stage have severe toxic side effects [12] , [13] . Staging of the infection , to select the appropriate therapeutics and follow up , is currently done by sampling the cerebral spinal fluid to search for the presence of parasites and/or increased numbers of lymphocytes [14] . The view that human trypanosome infections are invariably fatal if not treated has been challenged recently [15] , [16] but , nevertheless , early diagnosis is extremely important both for individual patient outcomes and for controlling epidemic spread [17] , [18] . The identification of infected individuals relies on dedicated screening teams that visit at-risk communities or patients seeking medical examination [19] . HAT diagnosis in the field faces many difficulties; not least the logistical challenges for the screening teams to attend communities in rural locations . In endemic areas , civil disturbance usually increases the incidence of HAT and decreases the frequency of screening [20] , [21] , [22] . Once the screening teams are with the communities , they face further challenges to recruit the entire local population into the HAT screening programme , which can lead to under-reporting and under-estimations of infection rates [23] , [24] , [25] , [26] . The current HAT screening regimen uses the Card Agglutination Test for Trypanosomiasis ( CATT ) , a serological test that detects whether antibodies from an individual are able to aggregate a suspension of fixed and stained T . b . gambiense trypanosomes [27] , detecting primarily antibodies to the variant surface glycoproteins ( VSGs ) on the fixed cells . If patients have a positive CATT result , microscopic examination of their blood is carried out to detect trypanosomes . If this is positive , a lumbar puncture is performed to stage the infection . Over the years , the CATT test has been optimised to improve sensitivity , specificity and stability . Such modifications include dilution of the blood samples , the use of multiple trypanosome clones expressing different VSG variants and improvements in thermostability [28] , [29] , [30] , [31] . Despite the usefulness and wide deployment of the CATT test , it has several widely accepted limitations [32] , [33] , [34] , [35] . These include varying degrees of sensitivity and specificity , its inability to detect T . b . rhodesiense infections , the requirement for trained screening personnel to use it and the specialised manufacture which precludes production on a scale necessary to saturate the market [26] , [36] . There have also been other post-CATT test diagnostic enhancements . For example , the concentration of trypanosomes from infected blood to improve microscopic detection [37] , [38] , [39] . Further , the detection of trypanosome DNA in blood by loop-mediated isothermal amplification of DNA ( LAMP ) [40] methods are under investigation and are summarised in a recent review [41] . However , these diagnostic methods require relatively sophisticated laboratory equipment . In summary , there is well accepted case for developing an extremely simple , low-cost diagnostic device with greater sensitivity and specificity than current field tests [4] . Lateral flow devices are simple tests that can rapidly detect nanogram amounts of antibodies or antigens in finger-prick blood samples without the need for any ancillary equipment [42] . T . b . gambiense infections are characterised by very low parasitemias , often <1000/ml , the equivalent of <5 ng total trypanosome protein/ml blood . Thus , using currently available technology , it is not feasible to directly detect a trypanosome protein and a lateral flow test that detects host antibodies is perhaps more likely to have the required sensitivity and specificity . The manufacture of large numbers of lateral flow devices requires milligram to gram amounts of diagnostic antigen , therefore potential diagnostic antigens for such devices should preferably derive from recombinant or synthetic sources . Recently the Foundation for Innovative New Diagnostics ( FIND ) has invested in developing new diagnostic tests for human African Trypanosomiasis [43] . With a similar aim in mind , we also set out to identify novel diagnostic antigens , and to create a prototype lateral flow test device , but using a non-biased ( proteomics ) approach to select potential biomarker antigens . The results of this antigen selection and the performance of a prototype lateral flow device are reported here .
All human serum samples were collected with the informed consent of the patients that they could be used anonymously for diagnostic development . Rodents were used to propagate sufficient T . brucei parasites to make the detergent lysates for immunoaffinity chromatography and proteomics . The animal procedures were carried out according the United Kingdom Animals ( Scientific Procedures ) Act 1986 and according to specific protocols approved by The University of Dundee Ethics Committee and as defined and approved in the UK Home Office Project License PPL 60/3836 held by MAJF . Two sets of human serum samples were used , the first was kindly provided by Philippe Büscher ( Institute of Tropical Medicine , Antwerp ) and consisted of nine sera from T . b . gambiense infected patients and nine from matched non-infected patients . These samples underwent virus inactivation using a procedure that retains antibody reactivity [44] . Briefly , 1% Tri ( n-butyl ) phosphate ( TnBP ) and 1% Triton X-45 were added to thawed serum samples and incubated at 31°C for 4 h . Sterile castor oil was added , mixed and the samples were centrifuged ( 3800× g , 30 min ) . The oil-extraction was repeated three times and the virus-inactivated sera ( lower phases ) were aliquoted and stored at −80°C . The second set of 145 patient sera ( 200 µl aliquots ) was obtained from the WHO Human African Trypanosomiasis specimen bank [45] . Serum samples were aliquoted and stored at either −80°C for long-term storage or in 50% glycerol at −20°C when prepared for ELISA analysis . Freeze-thawing was kept to a minimum; samples from P . Büscher and WHO were freeze-thawed three times and twice , respectively , prior to use in ELISA tests . Following virus inactivation , 125 µl of sera from four infected and four uninfected ( control ) patients were pooled . Each pool was applied to a 1 ml protein G column ( GE Healthcare ) equilibrated in phosphate buffered saline ( PBS ) . The columns were washed with 10 ml of PBS and the bound IgG antibodies were eluted with 50 mM sodium citrate pH 2 . 8 , and collected in 1 ml fractions into tubes containing 200 µl of 1 M Tris-HCl , buffer pH 8 . 5 . Peak fractions containing IgG were combined and dialysed for 16 h against coupling buffer ( 0 . 1 M NaHCO3 , 0 . 5 M NaCl , pH 8 . 3 ) . CNBr-activated Sepharose ( GE Healthcare ) was hydrated in 1 mM HCl and then equilibrated in coupling buffer . Aliquots ( 0 . 75 ml packed volume ) were mixed with 7 . 2 mg of purified infection IgG or purified control IgG in a final volume of 3 ml coupling buffer for 16 h at 4°C . The coupling of IgG was confirmed by measuring the absorbance of the supernatant at 280 nm before and after coupling . The Sepharose-IgG conjugates were centrifuged at 500× g ( 10 mins , 4°C ) and the beads were resuspended in 15 ml 1 M ethanoamine , pH 9 , to block remaining amine-reactive sites for 2 h at room temperature . Following this , the IgG-Sepharose beads were washed with three cycles of 0 . 1 M Tris-HCl , pH 8 . 0 , 0 . 5 M NaCl followed by 0 . 1 M sodium acetate buffer , pH 6 . 0 , 0 . 5 M NaCl and finally washed and stored in PBS containing 0 . 05% NaN3 . Six BalbC mice were injected with T . b . brucei Lister 427 variant MITat 1 . 4 cells . After three days , infected mouse blood was harvested with citrate anticoagulant , adjusted to 107 parasites per ml with PBS and aliquots of 0 . 5 ml were injected into the peritoneal cavity of 12 Wistar rats . The rat blood was harvested after 3 days with citrate anticoagulant and centrifuged at 1000× g for 10 min at 4°C . Plasma was removed and the buffy layer was resuspended in separation buffer plus glucose ( SB + glucose; 57 mM Na2HPO4 , 3 mM KH2PO4 , 44 mM NaCl , 10 g/l glucose ) and applied to a DE52 DEAE-cellulose ( Whatman ) column that had been pre-equilibrated with SB + glucose . The trypanosomes were washed through the column with SB + glucose , counted , centrifuged ( 900 g , 15 min , 4°C ) , resuspended in 1 ml PBS and then adjusted to 1×109 parasites/ml in ice-cold lysis buffer ( 50 mM Na2PO4 , pH 7 . 2 , 2% n-octyl β-D-glucopyranoside ( nOG ) detergent , 1 mM PMSF , 1 mM TLCK , 1 µg/ml aprotinin , 1 µg/ml leupeptin and 1× Roche protease cocktail minus EDTA ) . The lysate was incubated for 30 min on ice and then centrifuged at 100 , 000 g for 1 h at 4°C . Aliquots of T . b . brucei lysate ( 1010 cell equivalents ) were incubated with 0 . 75 ml packed volume of each of the Sepharose-IgG ( infection and non-infection/control ) gels , rotating for 3 h at 4°C . The gels were then packed into disposable 10 ml columns and washed with 10 ml of 10 mM Na2PO4 , pH 7 . 2 , 200 mM NaCl , 1% nOG , followed by 10 ml of 5 mM Na2PO4 pH 7 . 2 , 1% nOG . The trypanosome proteins were eluted 3 times with 750 µl of 250 mM sodium citrate , pH 2 . 8 , 1% nOG and the eluates were pooled and neutralised with 1 . 5 M Tris-HCl , pH 9 and further concentrated to 140 µl using a centrifugal concentrator ( Millipore , 0 . 5 ml capacity with 3 kDa MW cut off membrane ) . To remove eluted IgG , this fraction was mixed with 30 µl PBS-equilibrated Protein G agarose beads ( Pierce ) and incubated for 10 min and removed by centrifugation . The supernatant , containing the trypanosome proteins , were then transferred to low binding Eppendorf tubes and the proteins were precipitated by adding 1 ml ice-cold ethanol and incubation for 34 h at −20°C . Following ethanol precipitation , the proteins eluted from the infection IgG and control IgG gels were dissolved in SDS sample buffer , reduced with DTT and run on a precast 4–12% BisTris gradient SDS-PAGE gel ( Invitrogen ) using the MES running system . The gel was stained with colloidal Coomassie blue and equivalent regions of the infection and control lanes were cut out , reduced and alkylated with iodoacetamide and digested in-gel with trypsin . The tryptic peptides were analysed by LC-MS/MS on a Thermo Orbotrap XL system and MASCOT software was used to match peptides to the predicted trypanosome protein databases ( combined GeneDB and UniProt predicted protein sequences ) . Trypanosome proteins identified uniquely in the infection IgG immunopurified fractions were considered for recombinant expression . Within these , proteins with high MASCOT scores , likely to be the most abundant , were prioritised for recombinant expression and purification trials . These proteins included Gene Related to Expression Site Associated Gene ( GRESAG ) 4 , Invariant Surface Glycoprotein ( ISG ) 75 , ISG65 and ISG64 . The identified protein sequences were used to BLAST search the T . b . brucei predicted protein database , revealing several related protein sequences in each family . CLUSTALW2 alignments were carried out in order to better understand sequence sub-groups within those protein families . Representative gene segments from each protein sub-group that contained the peptide sequences identified by mass spectrometry were amplified from EATRO1125 genomic DNA ( for ISG65-1 , ISG65-2 , ISG64-2 , ISG64-3 and ISG75-1 ) or from stain 427 genomic DNA ( for ISG64-1 and GRESAG4 ) by PCR using the primers described in the Supporting Information ( Table S1 ) . In each case , the products of three separate PCR reactions were cloned into a TOPO-TA vector ( pCR2 . 1 ) for sequencing ( DNA Sequencing Service , College of Life Sciences , University of Dundee ) . The amplified ISG gene segments were cloned into various pET bacterial expression plasmids that provide a His-tag fused either to the N-terminus or C-terminus of the protein , in some cases via a TEV protease cleavage site , as indicated in ( Figure 1 ) . Multiple constructs were designed for GRESAG4 encoding the predicted full-length extracellular domain and several small globular domains based on predictions from GLOBplot software [46] ( Figure 1 ) . These constructs were amplified from genomic DNA using the primers described in the Supporting Information ( Table S1 ) , cloned into TOPO-TA vector pCR2 . 1 and verified by DNA sequencing . The constructs were either cloned into the pET15bTEV vector , such that the proteins they encode are fused at the N-terminus to a His tag , or into a pGEX-TEV vector such that the protein is fused at its N-terminus to a glutathione S-transferase ( GST ) sequence via a TEV cleavage site ( Figure 1 ) . The details of protein expression in E . coli and subsequent purification are described in the Supporting Information ( Text S1 ) . White ( Costar ) un-treated 96 well plates were coated at 50 µl/well for 16 h at 4°C with 2 µg/ml recombinant protein diluted in plating buffer ( 0 . 05 M NaHCO3 , pH 9 . 6 ) . Plating solution was removed and wells were blocked with PBS containing 5% BSA , 200 µl/well for 3 h at 22°C or 16 h at 4°C . Plates were stored at 4°C and used within 24 h . Aliquots ( 50 µl ) of serial serum dilutions ( see below ) were transferred in triplicate by a liquid handling device ( Bio-Tek , Precision ) to the ELISA plates and incubated for 1 h at room temperature , aspirated and 150 µl ELISA wash buffer was added to each well by the liquid handling device , left for 10 min and aspirated . This wash cycle was performed three times . Biotinylated goat anti-human-IgG ( Jackson Immunoresearch ) was diluted to 1∶5000 and 50 µl aliquots were applied to each well . After 1 h incubation at room temperature the secondary antibody solution was removed and wells were washed three times , as described above . Horseradish peroxidase ( HRP ) conjugated to NeutrAvidin ( Sigma ) was diluted to 1∶4000 and applied to the wells ( 50 µl/well ) for 1 h at room temperature . Wells were washed as before . Finally , chemiluminescent Femto substrate ( Pierce ) diluted 1∶5 ( i . e . , 0 . 5 ml solution A , 0 . 5 ml solution B with 4 ml PBS ) was applied to the wells at 50 µl/well and plates were read using an Envision plate reader after 2 . 5 min incubation at 22°C . ELISA measurements were made with both pooled and individual serum samples . Serum pools were made by combining patients sera from; stage 1 T . b . gambiense patients ( n = 10 ) , stage 2 T . b . gambiense patients ( n = 40 ) and matched uninfected patients ( n = 50 ) ; and from stage 1 T . b . rhodesiense patients ( n = 5 ) , stage 2 T . b . rhodesiense patients ( n = 20 ) and matched uninfected patients ( n = 25 ) . The pooled sera were diluted to 1∶60 in 50% glycerol , PBS and 1% BSA and stored at −20°C . For ELISA assays , the 1∶60 diluted pooled sera were further diluted to 1∶1000 in PBS , 0 . 1% BSA and then serially diluted ( doubling dilutions ) to 1∶32 , 000 . For the individual sera , the 1∶60 diluted samples were further diluted to 1∶1000 immediately before use . Sera were randomised by a member of the University of Dundee Tissue Bank . Forty T . b . gambiense infected patients sera and forty T . b . gambiense uninfected patients sera were randomly selected from the fifty T . b . gambiense infected and fifty uninfected WHO patient sera . These eighty serum samples were then randomised and coded . Serum aliquots ( 5 µl ) were diluted with 15 µl PBS and applied to the sample pad . Chase buffer ( 80 µl of PBS , 0 . 05% Tween 20 ) was added to the sample pad and the test was allowed to develop for 30 min . The test line was visually scored and the device was opened and the sample pads ( at top and bottom of nitrocellulose membrane ) were removed to prevent backflow . The lateral flow tests were photographed and scanned using a densitometer ( CAMAG TLC scanner 3 , CAMAG ) . Bar graphs and scatter plots ( x by y ) were generated by Microsoft Excel . Box plots , Receiver Operator Characteristic ( ROC ) curves , antigen scatter plots ( y axis only ) were generated by SigmaPlot 12 . Statistical analysis included Mann-Whitney ( Rank Sum Test ) and Dunn's post-hoc ( Analysis of Variance ( ANOVA ) on rank ) in SigmaPlot 12 . Data were tested for normality by Kolmogorov-Smirnov test and were further processed by Mann-Whitney or Dunn's post-hoc tests . The P values were recorded for Mann-Whitney with <0 . 05 set as the cut off for statistical significance .
We took a non-biased proteomics approach to identify proteins that adsorb selectively to pooled infection IgG , and not to pooled control IgG ( Figure 2A ) . Each serum pool contained four individual sera of patients clinically defined as having an infection with T . b . gambiense or as being uninfected . IgG fractions were purified from the pooled sera by affinity chromatography on protein G and then immobilised to cyanogen bromide-activated Sepharose beads . Equal amounts of infection and control IgG-Sepharose were incubated with equal amounts of T . b . brucei detergent cell lysate . Proteins that bound to the IgG columns were eluted by low pH , precipitated with cold ethanol , dissolved in SDS-sample buffer , reduced , separated by SDS-PAGE and stained with colloidal Coomassie blue ( Figure 2B ) . More protein was seen in the eluate from the infection IgG column , consistent with infection-specific anti-trypanosome immune responses . Equivalent sections were cut out from the infection and control lanes , as indicated ( Figure 2B ) . The excised gel pieces underwent in-gel S-alkylation and tryptic digestion , and the tryptic peptides were analysed by LC-MS/MS . Mascot software matched the peptide spectra to proteins in the T . b . brucei predicted protein database and scored the quality of the identifications . Lists of the proteins retained by infection IgG-Sepharose and control IgG-Sepharose were compared in each gel section . Twenty-four proteins with a MASCOT protein score above 50 were found uniquely in the infection IgG eluate and these are described in ( Table 1 ) . Several of the infection-specific proteins were defined as ‘hypothetical’ , but other hits included known cell surface proteins , such as: Invariant Surface Glycoprotein ( ISG ) 75 , ISG65 , ISG64 , Gene Related to Expression Site Associated Gene ( GRESAG ) 4 , and the transferrin receptor subunits ESAG 6 and 7 . As a starting point , the proteins with high MASCOT scores were prioritised . The rationale for this selection was that , by using an excess of trypanosome lysate in the affinity purification step , the amount of an eluted antigen should reflect , to a first approximation , the relative amount of antigen-specific immobilised IgG . The latter should , in turn , correspond to the immune response to that antigen in infected patients . Using this criterion , the protein antigens selected for study were ISG75 , ESAG7 , GRESAG4 , ISG65 , ISG64 and ESAG6 ( Table 1 ) . Next , we looked into the likely ease of protein expression of these antigens in E . coli . At this stage , we de-selected ESAG6 and ESAG7 because they form a heterodimer ( adding the complication of dual expression ) and because successful ( but low level ) protein expression has only been reported in a eukaryotic baculovirus expression system [47] . On the other hand , E . coli recombinant expression of domains of ISG75 , ISG65 and ISG64 had either been reported in the literature [48] or were known to the authors ( Mark Carrington , unpublished data ) . Consequently , we selected all three ISGs for protein expression trials . Finally , we performed expression trials on the predicted extracellular domain of GRESAG4 , for which there was no literature precedent . The selected purified recombinant trypanosome proteins , see Supporting Information ( Figure S1 ) , were used to prepare ELISA plates , as described in Experimental Procedures , and these were screened against various pooled human sera . These pools were derived from the 145 individual serum samples provided by the WHO Human African Trypanosomiasis specimen bank . The pooled sera were for stage 1 T . b . gambiense patients ( n = 10 ) , stage 2 T . b . gambiense patients ( n = 40 ) and matched uninfected patients ( n = 50 ) ; and from stage 1 T . b . rhodesiense patients ( n = 5 ) , stage 2 T . b . rhodesiense patients ( n = 20 ) and matched uninfected patients ( n = 20 ) . The results indicated that both stage 1 and stage 2 T . b . gambiense infection sera have significant antibody titres against all of the rISG64 and rISG65 proteins , compared to pooled non-infection sera ( Figure 3A ) , whereas infection sera titres against rISG75 and GRESAG4a were much closer to those for the control sera . The best performing recombinant protein was ISG65-1 , which had the highest infection to control signal . For the T . b . rhodesiense pooled sera , the signals were generally significantly lower , with the stage 2 pooled sera giving a significantly higher signal than the stage 1 pooled sera . There was one exception to this; the T . b . rhodesiense stage 1 pool had the highest antibody titre against rISG75 ( Figure 3B ) . However , as will be described later , the rISG75 result was due to a very high antibody titre in a single individual . From these results , all the rISG proteins were taken forward and screened against the individual sera but GRESAG4a ( rG4a ) was abandoned at this stage because it had poor infection versus non-infection discrimination . Recombinant protein ELISA plates that performed well in the pooled sera ELISAs were further screened against all of the individual sera . These antigens included three rISG64 proteins , two rISG65 proteins and one rISG75 protein . In this case , a total of 163 individual serum samples ( 145 from the WHO HAT specimen bank and 18 from the Institue of Tropical Medicine , Antwerp ) were diluted and applied in triplicate to wells coated with single recombinant proteins . T . b . gambiense and T . b . rhodesiense patient sera ELISA results were analysed separately ( Figure 4 ) and ( Figure 5 ) , respectively . The data are shown as box plots for each different recombinant antigen ELISA plate ( Figures 4A and 5A ) to provide a visualisation the range of antibody titres and the heat maps provide a different view of the same data ( Figures 4B and 5B ) . Both views suggest that rISG65 proteins provide the highest detection sensitivity whereas the rISG64-1 may provide slightly greater specificity . The rISG75 protein did not perform as well as the rISG65 or rISG64 proteins by both criteria and , indeed , only the stage 2 sera had statistically significant levels of IgG to rISG75-1 compared to controls ( Q = 4 . 616 , P = <0 . 05 ) . Dunn's post-hoc tests ( not shown ) demonstrated that , whereas there are significantly higher levels of anti-rISG64 and anti-rISG65 IgG antibodies in both stage 1 and stage 2 sera compared to uninfected controls , there is no statistically significant difference between the stage 1 and stage 2 groups . In other words , relative immunoreactivity to rISG64 or rISG65 antigens cannot be used to stage of the disease . Formal sensitivity ( i . e . , the proportion of correct positive results ) and specificity ( i . e . , the proportion of correct negative results ) parameters for each test were calculated by ROC curve analysis ( Figure 4C and 5C ) and are collated in ( Table 2 ) . The recombinant antigens that best discriminated between T . b . gambiense infected and control patients by ELISA were rISG65-1 and rISG64-1 , which had areas under the ROC curve of 0 . 99 and 0 . 98 respectively ( Figure 4C ) . The rISG65-1 ELISA antigen had sensitivity of 96 . 6% ( with a 95% Confidence Interval ( CI ) of 88 . 3 to 99 . 6% ) and specificity of 93 . 2% ( 95% CI of 83 . 5 to 98 . 1% ) , whereas sensitivity and specificity of rISG64-1 antigen was 93 . 2% ( 95% CI of 83 . 5 to 98 . 1% ) and 94 . 9% ( 95% CI of 85 . 9 to 98 . 9% ) , respectively ( Table 2 ) . It was more difficult to find a recombinant protein antigen that reliably discriminated T . b . rhodesiense infected patient sera from non-infected sera . The box plots , heat maps ( Figure 5A and B ) and Dunn's post hoc analyses ( not shown ) all indicate that , whereas the stage 2 sera show statistically significant immunoreactivity to all the antigens compared to controls , the immunoreactivities of the stage 1 sera are not statistically significant . rISG65-2 was the most sensitive at identifying T . b . rhodesiense infection sera 92% ( 95% , CI of 74 to 99% ) , but at a cost to specificity 85% ( 95% , CI 62 . 1 to 97% ) ( Figure 5C and Table 2 ) . As mentioned above , the pooled sera ELISA experiments had indicted that stage 1 T . b . rhodesiense infection sera might have high antibody titres towards rISG75 . However this proved not to be the case and was due to a single serum sample with a very high anti-rISG75 titre . Based on the ROC curve analyses of the performances of the ELISA plates , we selected rISG65-1 ( ROC curve area 0 . 99 for T . b . gambiense sera ) for development of a lateral flow prototype . Purified rISG65-1 was supplied to BBInternational ( Dundee , www . bbigold . com ) a company that specialises in lateral flow technology . The lateral flow approach that was utilised is illustrated in ( Figure 6 ) . Thus , rISG65-1 was both immobilised in a band on a nitrocellulose membrane and coupled to colloidal gold that was then localised in the conjugate pad . When the sera and chase buffer are applied to the sample pad , the rISG65-colloidal gold conjugate is resuspended . The absorbent pad at the top of the lateral flow device draws the liquid across the nitrocellulose membrane . During this time , any anti-rISG65 antibody in the serum binds to the rISG65-gold conjugate and when the antibodies reach the rISG65 test band , one Fab arm of the IgG binds to the immobilised rISG65 while the other Fab domain bridges to the rISG65-gold-conjugate . Accumulation of this specific antibody sandwich generates a visible test line . The control line is an internal positive control for the lateral flow test and does not relate to the infection status of the patient but indicates successful test flow . The final reading of this test should be as follows; the appearance of only a control line ( upper band ) indicates non-infected sera , whereas , the appearance of two lines , a control and test line ( upper & lower bands ) indicates infected sera , examples are shown in ( Figure 6 ) . Absence of a control line ( upper band ) indicates an invalid test , irrespective of the appearance of the test line and the test should be repeated . Eighty randomised and coded WHO ‘test’ T . b . gambiense sera , comprising forty infected and forty non-infected sera were applied to the lateral flow prototypes . Each serum sample ( 5 µl ) was diluted with 15 µl of PBS and applied to a lateral flow device sample pad . Within about 30 s , 80 µl of chase buffer was added and the test was left for 30 min , at which point a visual score was recorded . The sample pads were removed to prevent back flow and the visual scores were decoded ( Figure 7 ) . Sensitivity and specificity were calculated by ROC curve analysis , and for visual scores a cut off of 2 . 5 gave 100% sensitivity ( 95% CI of 91 . 1 to 100 ) and 87 . 5% specificity ( 95% CI of 73 . 2 to 95 . 8% ) . An analysis of the test lines was also carried out using a densitometer , where an arbitrary cut off at 265 . 6 RU gave 100% sensitivity ( 95% CI of 91 . 2 to 100% ) and 92 . 5% specificity ( 95% of CI 79 . 6 to 98 . 4% ) ( Table 3 ) indicating there is potential for separation between infection and non-infected individual scores . Principally the end user will interpret the results visually therefore further optimisation of the test line will be necessary to reduce false positive results due to non-specific binding . A checklist , Supporting Information ( Table S2 ) , and flow diagram , Supporting Information ( Figure S2 ) , are provided according to the STAndards for the Reporting of Diagnostic accuracy studies ( STARD ) guidelines .
The overall goal of this project was to develop an immunodiagnostic lateral flow prototype for human African trypanosomiasis that might be developed into a field-based device to replace the CATT screening tool . To do this , we needed to identify potential diagnostic antigen candidates , investigate whether they could be adequately expressed and purified and assess their diagnostic potential with patient sera . We took an unbiased proteomics approach to identify more than twenty potential diagnostic protein antigens , several of which were known cell-surface glycoproteins . This list was filtered pragmatically; first , the proteins with high proteomic MASCOT scores , generally synonymous with their abundance , were selected because , by using an excess of trypanosome lysate in the affinity purification step , the amount of an eluted antigen should reflect the relative amount of antigen-specific IgG in infection sera . The latter should , in turn , correspond to the immune response to that antigen in infected patients . From this list we eliminated ESAG6 , since it is known to form a heterodimer with ESAG7 and is , therefore , relatively complicated to express [46] . Our attempts to express parts of the extracellular domain of GRESG4 in E . coli were not very successful , although we were able to isolate the A-domain fused to GST . However , this G4a construct and the ISG75 protein construct did not perform well in the ELISA studies and were removed from this study . Nevertheless , these antigens should not be ignored for diagnostic development as they may simply have been miss-folded in the absence of endoplasmic reticulum folding and quality control components [49] . Indeed , recombinant ISG75 has been shown to have diagnostic potential for T . b . brucei animal infections [47] . Future expression attempts might include bacterial expression systems that target recombinant proteins into the periplasmic space [49] and/or eukaryotic expression systems such as insect cells and Pichia pastoris . Our data on the diagnostic potential of ISGs 64 , 65 and 75 for detecting T . b . rhodesiense infections were somewhat hampered by the small number of sera available for testing . Nevertheless , it is clear from the ELISA data that IgG antibody responses to these antigens are lower than in T . b . gambiense infections . Further , the IgG responses are particularly low in stage 1 T . b . rhodesiense patient sera . This may be due to the differing nature and speed of progression of the infections; T . b . rhodesiense infections are usually acute and progress faster whereas T . b . gambiense infections are chronic and progress over months or years [50] , which could in turn lead to a greater amount and diversity of antibodies present in these sera . A previous study also struggled to identify diagnostic antigens for T . b . rhodesiense infections [51] . A good approach may be to repeat the procedures described here using immobilised IgG from T . b . rhodesiense patient sera . Further research is also required to measure the half-life of antibodies in patients after they have been treated for HAT , as persistent antibodies may lead to false positives . It has been described that antibodies can persist up to 3 years post cure , however it is not known which class of antibodies persist or which antigens they recognise [52] . Ideally , a longitudinal study could be carried out to gain a greater insight into this and how it could affect the diagnostic potential of any future lateral flow test relying on antibodies [32] . Lateral flow tests , whilst having limitations , could potentially be more suitable for use in the field because of their stability and the fact that they can be used by non-specialists [53] . In summary , we report here the selection of ISG65-1 as a potential diagnostic antigen for T . b . gambiense infections and its performance in both conventional ELISA and prototype lateral flow device assays looks promising . The performance of the prototype ISG65 lateral flow device encourages us to further develop and optimize it , perhaps adding an additional antigen or antigens to improve sensitivity and specificity , while aiming for a production cost of <US$1 per unit . | Human African Trypanosomiasis is caused by infection with Trypanosoma brucei gambiense or T . b . rhodesiense . Preliminary diagnosis of T . b . gambiense infection relies mainly on a Card Agglutination Test for Trypanosomiasis ( CATT ) , which has acknowledged limitations . New approaches are needed , first to identify new diagnostic antigens and , second , to find a more suitable platform for field-based immunodiagnostic tests . We took an unbiased approach to identify candidate diagnostic antigens by asking which parasite proteins bind to the antibodies of infected patients and not to the antibodies of uninfected patients . From this list of twenty-four candidate antigens , we selected four and from these we selected the one that worked the best in conventional immunodiagnostic tests . This antigen , ISG65 , was used to make lateral flow devices , where a small sample of patient serum is added to a pad and thirty minutes later infection can be inferred by simple optical read out . This simple prototype device works as well as the CATT test and may be developed and optimized for clinical use in the field . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"biochemistry",
"infectious",
"diseases",
"diagnostic",
"medicine",
"immunology",
"biology",
"microbiology",
"proteomics"
] | 2013 | Proteomic Selection of Immunodiagnostic Antigens for Human African Trypanosomiasis and Generation of a Prototype Lateral Flow Immunodiagnostic Device |
Weight-loss interventions generally improve lipid profiles and reduce cardiovascular disease risk , but effects are variable and may depend on genetic factors . We performed a genetic association analysis of data from 2 , 993 participants in the Diabetes Prevention Program to test the hypotheses that a genetic risk score ( GRS ) based on deleterious alleles at 32 lipid-associated single-nucleotide polymorphisms modifies the effects of lifestyle and/or metformin interventions on lipid levels and nuclear magnetic resonance ( NMR ) lipoprotein subfraction size and number . Twenty-three loci previously associated with fasting LDL-C , HDL-C , or triglycerides replicated ( P = 0 . 04–1×10−17 ) . Except for total HDL particles ( r = −0 . 03 , P = 0 . 26 ) , all components of the lipid profile correlated with the GRS ( partial |r| = 0 . 07–0 . 17 , P = 5×10−5–1×10−19 ) . The GRS was associated with higher baseline-adjusted 1-year LDL cholesterol levels ( β = +0 . 87 , SEE±0 . 22 mg/dl/allele , P = 8×10−5 , Pinteraction = 0 . 02 ) in the lifestyle intervention group , but not in the placebo ( β = +0 . 20 , SEE±0 . 22 mg/dl/allele , P = 0 . 35 ) or metformin ( β = −0 . 03 , SEE±0 . 22 mg/dl/allele , P = 0 . 90; Pinteraction = 0 . 64 ) groups . Similarly , a higher GRS predicted a greater number of baseline-adjusted small LDL particles at 1 year in the lifestyle intervention arm ( β = +0 . 30 , SEE±0 . 012 ln nmol/L/allele , P = 0 . 01 , Pinteraction = 0 . 01 ) but not in the placebo ( β = −0 . 002 , SEE±0 . 008 ln nmol/L/allele , P = 0 . 74 ) or metformin ( β = +0 . 013 , SEE±0 . 008 nmol/L/allele , P = 0 . 12; Pinteraction = 0 . 24 ) groups . Our findings suggest that a high genetic burden confers an adverse lipid profile and predicts attenuated response in LDL-C levels and small LDL particle number to dietary and physical activity interventions aimed at weight loss .
Dyslipidemia is a strong risk factor for atherosclerotic heart disease [1]–[3] , has a well-defined genetic basis [4] , and is modifiable through therapeutic lifestyle changes and weight-loss interventions [5] , [6] . Individuals at risk for diabetes are also at high risk of cardiovascular disease [7] , and individualized lifestyle intervention programs , like the one incorporated into the Diabetes Prevention Program ( DPP ) , have a salutary effect on dyslipidemia and cardiovascular disease risk in this population . However , the cost of widespread implementation of such interventions has been highlighted as a major limitation [8] and not all benefit equally from such interventions . Identifying persons most likely to benefit from intensive lifestyle modification could provide justification for targeting this subpopulation first , making the clinical translation of findings from studies such as the DPP more feasible . Selection of persons whose dyslipidemia is likely to respond well to lifestyle interventions or pharmacotherapy could help target resources and optimize prevention strategies . To do so requires knowledge of the underlying risk factors for the trait and knowledge of how personal characteristics interact with exercise , diet , and weight loss . Although the heritability of polygenic dyslipidemia [9]–[11] and its sequelae [12] have been elucidated , little is known of how lifestyle interventions modify the effects of these loci , singly or in combination , on lipid profiles . Thus , learning how a person's genetic background modulates his or her response to therapeutic lifestyle changes and weight-loss interventions might help optimize the targeting of interventions designed to mitigate cardiovascular and metabolic disease risk . The purpose of this study was to examine whether loci reliably associated with polygenic dyslipidemia modified the response to cardio-protective interventions in the DPP , a randomized clinical trial of intensive lifestyle modification , metformin treatment , or placebo with standard care . We hypothesized i ) that the baseline lipid profiles of DPP participants would be associated with gene variants known to associate with polygenic dyslipidemia and ii ) that improvement in lipidemia following treatment would depend on these same genetic variants . We also used NMR spectroscopy to characterize the associations of these previously reported loci with lipoprotein subfractions .
Thirty-two SNPs previously associated with triglycerides ( TG ) , low-density lipoprotein-cholesterol ( LDL-C ) and/or high-density lipoprotein-cholesterol ( HDL-C ) levels were considered [10] . Thirty-one of these were successfully genotyped in the DPP , and two SNPs in CETP , serving as HapMap proxies ( r2≥0 . 90 ) for rs173539 , including rs247616 , were subsequently successfully genotyped , with rs247616 retained as the replacement for rs173539 . Twenty-three of these 32 non-redundant SNPs replicated with their respective traits in a directionally consistent manner ( P≤0 . 05 ) , including 8/11 for TG , 9/14 for HDL-C and 8/11 for LDL-C . Two of the SNPs , rs12678919 and rs964184 , replicated for both HDL-C and TG ( Table S1 ) . Additionally , we evaluated the associations of the 32 lipid loci with baseline lipids and nuclear magnetic resonance ( NMR ) -derived lipoprotein traits ( Large HDL particles , Small HDL particles , Total HDL particles , HDL size , LDL size , Total LDL particles , Small LDL particles , Total VLDL particles , Large VLDL particles , VLDL size ) . Of all analyses of baseline traits , roughly one third of the tests were nominally significant associations , and 35 associations were significant after correcting for all 448 hypothesis tests; these involved 12 SNPs and 13 traits ( Table S2 ) . Interestingly , SNP rs10401969 did not replicate for LDL-C ( C vs . T: β±SEM = −0 . 1±1 . 6 mg/dl , additive P = 0 . 94 ) , but was associated with decreased large VLDL ( mean 5 . 43 , 4 . 26 , 4 . 07 nmol/L for TT , TC , CC genotypes respectively , additive P = 4×10−5 ) and smaller VLDL size ( 53 . 18 , 50 . 94 , 49 . 55 nm , P = 2×10−6 ) . SNP rs7679 did not quite reach nominal significance for decreased HDL-C ( C vs . T: β ±SEM = −0 . 016±0 . 009 ln mg/dl , additive P = 0 . 07 ) , but was very strongly associated with increased small HDL particle number ( 17 . 93 , 20 . 14 and 21 . 89 µmol/L for TT , CT and CC genotypes respectively; additive P = 4×10−18 ) and consequently total HDL particle number ( 34 . 07 , 35 . 10 and 36 . 78 µmol/L , P = 2×10−5 ) . A lipid GRS was calculated for each individual by first replacing missing genotypes with ethnicity-specific imputed means and then adding up the number of risk alleles possessed for each of the 32 independent SNPs . Of the 32 SNPs evaluated , 11 were originally associated in the meta-analysis with LDL cholesterol , 10 with HDL cholesterol only , seven with triglycerides only , and four with both HDL cholesterol and triglycerides . A risk allele was defined as one associated with increased TG or LDL-C or decreased HDL in the original meta-analysis [10] . After adjustment for age , sex , ethnicity , and BMI , the GRS was significantly associated with all baseline traits evaluated except total HDL particles ( P = 0 . 26 , Table 2 ) . The following are P-values for the effects of the GRS , as a quantitative covariate , and geometric means for the upper and lower ethnicity-specific GRS quartiles for each trait . A higher GRS was associated with elevated baseline levels of: total cholesterol ( P = ×410−11 , 206 vs . 195 mg/dl ) , LDL-C ( P = ×910−8; , 129 vs . 121 mg/dl arithmetic means ) , TG ( P = ×410−19 , 160 vs . 127 mg/dl ) , total VLDL particles ( P = ×610−14 , 67 vs . 53 nmol/L ) , large VLDL particles ( P = ×110−14 , 6 . 57 vs . 4 . 21 nmol/L ) , total LDL particles ( P = ×210−10 , 1412 vs . 1262 nmol/L ) , small LDL particles ( P = ×210−11 , 743 vs . 543 nmol/L ) , small HDL particles ( P = µ0 . 0005 , 19 . 17 vs . 18 . 10 mol/L ) , and VLDL particle size ( P = ×110−5 , 53 . 86 vs . 51 . 84 nm ) . A higher GRS was also associated with lower baseline levels of: HDL-C ( P = ×110−15 , 43 vs . 47 mg/dl ) , LDL particle size ( P = ×110−19 , 0 . 256 vs . 0 . 269 nm ) , large HDL particles ( P = ×210−8µ , 2 . 98 vs . 3 . 68 mol/L ) , and HDL particle size ( P = 0 . 0003 , 8 . 82 vs . 8 . 90 nm ) . All of these results are consistent with a greater number of risk alleles increasing the atherogenicity of the lipoprotein profile . Two traits showed evidence of GRS×lifestyle interaction: LDL-C ( P = 0 . 02 ) and small LDL particles ( P = 0 . 01 , Table 3; Figure 1; Figure S1a–S1f ) . For these two traits , there was a residual detrimental impact of GRS in the lifestyle ( i . e . , the GRS was associated with higher levels at one year even after adjusting for baseline levels ) but not the metformin or placebo group , suggesting that the lifestyle intervention was less effective at lipid-lowering in those with a higher genetic burden . A unit ( allele ) GRS increase was associated with higher residual LDL-C levels in the lifestyle group ( β+±0 . 087 , SEE0 . 022 mg/dl , P = ×810−5 ) but not in the metformin ( β−±0 . 03 , SEE0 . 22 mg/dl , P = 0 . 90 ) or placebo ( β+±0 . 20 , SEE0 . 22 mg/dl , P = 0 . 35 ) groups ( Figure 1 ) . Similarly , the GRS was associated with higher residual ln-small LDL particles in the lifestyle group ( β+±0 . 030 , SEE0 . 0 . 012 ln nmol/L , P = 0 . 01 ) , but not in the metformin ( β−±0 . 013 , SEE0 . 008 ln nm/L , P = 0 . 12 ) or placebo ( β−±0 . 002 , SEE0 . 008 , P = 0 . 74 ) groups ( Figure 1× ) . There were no metforminGRS interactions significant at the P = ×0 . 05 level . In addition to the three traits discussed , several traits showed residual detrimental effects of the GRS in one or more strata ( total cholesterol , TG , LDL size , total VLDL particles , large HDL particles , and HDL size ) without any statistical evidence of treatmentGRS interaction ( Table 3 ) .
In the present study , the majority of previously associated SNPs replicated for baseline lipid traits , and there was a statistically significant relationship between the GRS and the vast majority of standard lipid traits and NMR lipoprotein subfractions . Importantly , in several cases the evidence for association with NMR subfractions was much stronger than for the original standard lipid trait , which may be owing to the relative proximity of the subfractions to the genetic loci . For example , SNP rs7679 , originally associated with total HDLC levels in previous GWASs , in the DPP was not significantly associated with HDL-C ( P = 0 . 07 ) but was strongly associated with small HDL particle levels ( P = ×410−18 ) . This SNP is near PLTP , encoding phospholipid transfer protein , a molecule directly influencing HDL particle size [14] . Such findings extend our understanding of lipid biology and suggest that , compared to standard lipid levels , measurements of lipoprotein subfractions may provide a more effective way of capturing genetically influenced risk . This is particularly important , as recent studies have shown that HDL-C is a heterogeneous trait , which in the context of clinical use may benefit from sub-stratification by genotype [15] . We also observed that within the lifestyle group but not the placebo or metformin groups , the GRS was associated with higher LDL-C and small LDL particle levels after one year of intervention . These findings suggest that the genetic burden on these traits cannot be completely overcome by lifestyle modification . However , even those with the greatest genetic burden benefit to a limited extent from lifestyle intervention in terms of LDL-C reduction ( Figure 1A ) , although the effect on LDL particle size reduction is almost completely ablated ( Figure 1B' ) . Even a true residual effect of lifestyle on LDL-C in people with the highest GRS does not negate the clinical relevance of our findings in terms of potential to facilitate tailored treatment decisions . Seeing less of an effect of lifestyle in a particular patient subgroup indicates that these persons may benefit from more frequent surveillance , more intense lifestyle interventions , or aggressive pharmaceutical interventions to supplement lifestyle interventions . Conversely , knowing that lifestyle intervention is likely to be adequate in persons with the lowest genetic burden may maximize the patients diet adherence and potentially reduce the costs and side effects associated with prescribing lipid-lowering medications unnecessarily . The availability of information on genetic background may also facilitate patient-provider dialogue , owing to improved diagnostic accuracy . This is similar to the strategy used to control cholesterol levels in patients with a monogenic disorder such as familial hypercholesterolemia due to a severe loss of function mutation in LDLR , where lifestyle intervention combined with pharmacotherapy is needed to bring LDL-C levels within an acceptable range ( Third Report of the NCEP-ATP III on the Detection , Evaluation , and Treatment of High Blood Cholesterol in Adults ) . ×It is also important to bear in mind that the interaction effects may be underestimated in our paper . This is because the majority of SNPs included in the GRS are likely to be imperfect proxies for unobserved functional variants , resulting in some degree of genotype misclassification . Moreover , all 32 SNPs were included within the GRS , even though not all SNPs convey statistically significant effects in the DPP and do not individually modify the effects of the interventions . A parsimonious GRS including only those SNPs that are statistically significant in the DPP would likely be overfitted to our data , resulting in biased conclusions about the strength and magnitude of genetreatment interactions . Dyslipidemia is a long-established risk factor for CVD [1]–[3] . Thus , the primary and secondary prevention of atherosclerotic CVD often involves intervening on lipid levels [16] . Lifestyle interventions [17] and metformin treatment [18]; that result in weight loss have the potential to improve lipid profiles nevertheless , as long recognized [19] , changes in lipid profiles following interventions vary greatly from one person to the next . Some of the variability in response to interventions may be because genotypes modulate the effects of preventive interventions on lipid homeostasis and CVD risk [20] . Of the many known dyslipidemia-predisposing loci discovered so far [10]× , only a handful have been the focus of studies testing hypotheses of genetreatment interactions [21]–[28] , and most of these studies are small ( N<150 ) , non-randomized trials of dietary intervention . Although some of these studies have focused on genomic regions that are confirmed to harbor dyslipidemia-predisposing loci , such as APOB , CETP , LIPC and LPL [21] , [22] , [24]–[28] , no exhaustive studies testing whether GWAS-discovered loci [10] , [29] , [30] modify response to treatments have been previously reported . The GRS used in this study attenuated the impact of the DPP lifestyle intervention on LDL levels and small LDL particle number . This suggests that a genetic predisposition to high LDL levels and more small LDL particles is difficult to overcome through lifestyle intervention alone . These data also unmask the effects of an underlying genetic defect of LDL levels and small LDL particles found in individuals with a high genetic burden , which becomes visible when adiposity and blood TG content are reduced through lifestyle intervention . This information may justify the combination of lifestyle and lipid lowering drug treatment from the outset in these individuals , rather than the usual approach of stepping from lifestyle to drug therapy when the former fails . ;The DPP lifestyle intervention prioritized weight loss , daily fat gram intake and physical activity goals over intake of saturated fat , cholesterol , viscous fiber and plant stanols/sterols this may have influenced the nature of the changes in the lipid profile . When compared to the metformin and placebo groups , the lifestyle intervention group reported improved physical activity levels and reductions in calorie intake , resulting in significantly greater weight losses [31] , each of which has major influences on TG levels . The lifestyle intervention group reported significantly greater reductions in percent calories from total fat and saturated fat than the metformin and placebo groups [31] . However , they did not , on average , achieve the National Cholesterol Education Program target for saturated fat intake and did not focus on the other therapeutic lifestyle changes , such as the additional dietary changes mentioned above , that often have the largest effects on LDL concentrations . The ethnic diversity of the DPP cohort facilitates the generalizability of results , but may also lead to confounding by population stratification in genetic analyses . However , Sensitivity analyses in the European White sub-cohort of the DPP yielded comparable effect estimates to the results obtained in the entire DPP genetics cohort ( results for baseline traits shown in Table S3 ) , supporting the conclusion that confounding by population stratification is unlikely to explain our findings . ×Interestingly , no significant interaction was observed between the GRS and other biochemical components of the lipid profile in the present study . It is important to bear in mind , however , that despite being the largest clinical trial of its kind , the DPP is only moderately powered to detect genetreatment interactions [32];× it is likely , therefore , that genetreatment interactions that are small in magnitude will have been overlooked here . Moreover , during the course of writing this paper , many smaller impact lipid loci have been discovered [9] , [11] . Thus , it is possible that with a larger sample size and the inclusion of some or all of these additional loci , we may have discovered interaction effects on other lipid traits . In summary , we have shown that common genetic loci that influence polygenic dyslipidemia also modify the effects of clinical interventions designed to mitigate cardiovascular and metabolic risk . This report is the first comprehensive effort to examine validated lipid loci within the context of a large randomized clinical trial . The findings of this study may facilitate the implementation of complex trait genetics into the clinical setting .
The DPP was a multi-center randomized controlled trial that examined the effects of metformin or intensive lifestyle modification on the incidence of type 2 diabetes [33] , [34]∼∼%%∼– = . Briefly , overweight persons with elevated but non-diabetic fasting and post-challenge glucose levels were randomized to receive placebo , metformin ( 850 mg twice daily ) or a program of intensive lifestyle modification . The lifestyle intervention was designed to achieve 150 min/wk of physical activity and 7 weight loss via focus on daily fat gram goals . Fat gram goals were based on initial weight and 25 of calories from fat using a calorie level estimate to produce a weight loss of 0 . 51 kg/wk . The principal endpoint was the development of diabetes by confirmed semi-annual fasting plasma glucose or annual oral glucose tolerance testing ( OGTT ) . Other phenotypes , such as changes in weight , waist circumference , lipids , insulin and glucose , were also ascertained . Written , informed consent was obtained from each participant , and each of the 27 DPP centers obtained institutional review board approval prior to initiation of the study protocol . A total of 2 , 993 participants in the placebo , lifestyle and metformin groups had DNA available and provided consent for genetic analysis . Individuals taking lipid lowering medications at baseline ( n145 ) were excluded from all analyses . ≥[]All participants fasted for 12 hrs the night before blood was drawn from an antecubital vein . Standard blood lipid measurements ( triglyceride TG , total cholesterol , HDL-C , calculated LDL-C ) were performed at the DPP central biochemistry laboratory . TG and total cholesterol levels were measured using enzymatic methods standardized to the Centers for Disease Control and Prevention reference methods [35] . HDL fractions for cholesterol analysis were obtained by the treatment of whole plasma with dextran sulfate Mg+2 [36] . LDL cholesterol was calculated by the Friedewald equation [37]>β . In participants with TGs4 . 5 mmol/l , the lipoprotein fractions were separated using preparative ultracentrifugation of plasma by quantification [38] . Lipoprotein subclass particle concentrations and average VLDL , LDL , and HDL particle diameters were measured by NMR spectroscopy at LipoScience , Inc ( Raleigh , NC ) with modification of existing methods [39] . Thirty-two SNPs previously associated with lipid concentrations in GWAS meta-analyses [10] were selected . DNA was extracted from peripheral blood leukocytes using standard methods . Genotyping was performed by allele-specific primer extension of multiplex amplified products and detection using matrix-assisted laser desorption ionization time-of-flight mass spectrometry on a Sequenom iPLEX platform [40]%% . The mean genotyping success rate was 96 . 7 . The minimum call rate was 94 . 0 . All SNPs were in Hardy-Weinberg equilibrium within each self-reported ethnic group . The SAS software v9 . 2 ( SAS , Carey , NC ) was used for analyses . Baseline total cholesterol , HDL-C , TG and all lipoprotein sub-fraction levels were natural log transformed for non-normality , and LDL-C was evaluated directly . For replicating the previously reported associations of SNPs with baseline traits and evaluating the association of the individual SNPs with NMR lipoprotein particle sizes and numbers , measurements were compared across genotypic groups by ANCOVA ( general model , 2 df F test for three possible genotypes ) , and evidence for an additive effect of genotype was also evaluated using the measured genotype approach , in which each genotype was assigned a value of 0 , 1 or 2 according to the number of minor alleles . Analyses of baseline traits were adjusted for age , sex , self-reported ethnicity ( to minimize confounding due to potential differences in both allele frequency and lipid traits across ethnicities ) and BMI . For the individual SNP analyses , the Bonferroni-corrected P-value for significance was set at P<× = ; = 0 . 0001 to account for multiple comparisons ( 32 SNPs14 traits448 tests 0 . 05/4480 . 0001 ) . A genetic risk score ( GRS ) was calculated from the 32 SNPs using the direction of association from the initial association seen in the published meta-analysis [10]; for each SNP , an allele was designated as a risk allele if it was associated with higher TG or LDL-C and/or lower HDL-C . In order to be able to incorporate all individuals in the analysis , including those missing genotypes at one or more loci , a simple imputation procedure within each self-reported ethnic group was implemented ( in order to account for allele frequency differences across ethnicities ) prior to score calculation . First , after coding the genotype as the number of minor alleles ( 0 , 1 or 2 ) , an ethnicity-specific mean genotype was calculated and rounded to the nearest whole number . Missing genotypes were replaced by the appropriate rounded mean genotype [41]%×× . We calculated a GRS for each individual by adding up the number of risk alleles for each of the 32 tested SNPs , where a risk allele was defined as one associated with increased TG or LDL-C or decreased HDL . The GRS was then included as a quantitative independent variable in a multiple regression model for each baseline lipid/lipoprotein trait to test for association , adjusted for age , sex , self-reported ethnicity , and BMI . GRS quartiles were constructed separately within each self-reported ethnicity prior to calculating quartile-specific arithmetic means or geometric means and 95 confidence intervals . To test for interaction of the risk score with treatment , a multiple regression model was constructed with the 1 year value as the outcome variable and including GRS , lifestyle and metformin treatment and GRSlifestyle and GRSmetformin terms , along with adjustments for the corresponding baseline trait , baseline age , sex and self-reported ethnicity . A priori power calculations are an important study-planning tool , providing relevant information on likely effect sizes and variances is accessible . It is possible to obtain a broad understanding of the power constraints of our study by extrapolating results from other experimental settings ( as described in detail in [42] ) , but specific a priori× power calculations could not be performed for the current study because reliable effect estimates and variances for tests of genetreatment interactions for the index genotypes and phenotypes were unavailable in the published literature at the time this study was planned . Post-hoc' power calculations were not performed , as these are well known to cause bias when interpreting a studys results [43]–[46] . However , confidence intervals are included in the figures , which give insight into the precision of the GRS effect estimates and hence the power to detect those estimates in the DPP cohort . | The study included 2 , 993 participants from the Diabetes Prevention Program , a randomized clinical trial of intensive lifestyle intervention , metformin treatment , and placebo control . We examined associations between 32 gene variants that have been reproducibly associated with dyslipidemia and concentrations of lipids and NMR lipoprotein particle sizes and numbers . We also examined whether genetic background influences a person's response to cardioprotective interventions on lipid levels . Our analysis , which focused on determining whether common genetic variants impact the effects of cardioprotective interventions on lipid and lipoprotein particle size , shows that in persons with a high genetic risk score the benefit of intensive lifestyle intervention on LDL and small LDL particle levels is substantially diminished; this information may be informative for the targeted prevention of dyslipidemia , as it suggests that genetics might help identify persons in whom lifestyle intervention is likely to be an effective treatment for elevated lipids and lipoproteins . The NMR subfraction analyses provide novel insight into the biology of dyslipidemia by illustrating how numerous genetic variants that have previously been associated with lipid levels also modulate NMR lipoprotein particle sizes and number . This information may be informative for the targeted prevention of cardiovascular disease . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"medicine",
"diabetes",
"mellitus",
"type",
"2",
"drugs",
"and",
"devices",
"endocrinology",
"pharmacogenetics",
"diabetic",
"endocrinology",
"pharmacology",
"diabetes",
"and",
"endocrinology"
] | 2012 | Genetic Modulation of Lipid Profiles following Lifestyle Modification or Metformin Treatment: The Diabetes Prevention Program |
Previous studies of repetitive elements ( REs ) have implicated a mechanistic role in generating new chimerical genes . Such examples are consistent with the classic model for exon shuffling , which relies on non-homologous recombination . However , recent data for chromosomal aberrations in model organisms suggest that ectopic homology-dependent recombination may also be important . Lack of a dataset comprising experimentally verified young duplicates has hampered an effective examination of these models as well as an investigation of sequence features that mediate the rearrangements . Here we use ∼7 , 000 cDNA probes ( ∼112 , 000 primary images ) to screen eight species within the Drosophila melanogaster subgroup and identify 17 duplicates that were generated through ectopic recombination within the last 12 mys . Most of these are functional and have evolved divergent expression patterns and novel chimeric structures . Examination of their flanking sequences revealed an excess of repetitive sequences , with the majority belonging to the transposable element DNAREP1 family , associated with the new genes . Our dataset strongly suggests an important role for REs in the generation of chimeric genes within these species .
Gene duplication followed by the acquisition of novel molecular function is a fundamental process underlying biological diversity . It has been theoretically and empirically demonstrated that functionally distinct duplicates are capable of evolving through a neofunctionalization process in which there is an accumulation of mutations in a redundant copy of a preexisting gene [1–3] . In addition , there is mounting evidence for the rapid generation of new genes through the recombination of preexisting exons and functional domains . This latter process does not exclude , and in fact often relies on , the duplication of the loci involved [4 , 5] . Excluding chimeric genes formed through retroposition [6–8] , more than three hundred gene families are believed to have originated through exon shuffling [9] . Most of these gene families have introns , suggesting that DNA level recombination was involved ( DLR; DLR as opposed to a retroposition event involving an RNA intermediate ) . Since its initial proposal [10] , the genetic mechanisms involved in the formation of chimeric genes through exon shuffling have largely remained a mystery . The classic model states that nonhomologous recombination ( NHR ) brings together exons or domains from ectopic positions [10] . Experimental evidence for the role of NHR has been gained through transfection experiments [11 , 12] and through surveys of rearrangement hotspots which are often disease-associated [13–15] . Breakpoint analyses on these datasets revealed little or no sequence identity between the loci recombined , supporting a NHR model . While these experiments show such a model is possible for exon shuffling , it remains an open question how frequently such processes in non-artificial systems , and over evolutionary time , will contribute to the formation of fixed chimeric genes . Another potential NHR mechanism that can mediate nonhomologous recombination is through the activity of transposable elements ( TEs ) . If a TE is capable of mobilizing adjacent sequence , novel junctions that share no sequence identity could be generated [16] . The capacity for such events has been documented with the imprecise excision of well studied TEs such as P elements [17] as well as in plant pack-MULE and Helitron TEs [18–20] . These investigations implicate a role for TEs in the generation of chimeric genes . Whether these shuffled products are under functional constraint remains an interesting question . Alternatively , non-allelic homologous recombination ( NAHR ) between ectopic sequences can lead to the formation of chimeric genes . Recently , a surge of evidence has begun to demonstrate the importance of NAHR to genomic architecture , especially in primates [21–26] . Intriguingly , several studies have reported on a limited number of chimeric gene structures , some of which appear functional and nondeleterious , but most remain putative [24 , 27] . Focus has primarily been placed on NAHR's role in human disease [26] . However , given that NAHR appears to be a common mutational mechanism , a new hypothesis for exon shuffling has been motivated: Despite the frequently deleterious effects , NAHR is capable of making a contribution to the origin of new chimeric genes as an exon shuffling mechanism [24 , 25 , 28 , 29] . A difficulty in investigating the relative contributions of these mechanisms to the formation of chimeric genes is that most of the available examples are evolutionarily ancient [9] . These genes provide few clues for understanding the recombination mechanisms that generated their initial structures because the sequence features , especially those non-constrained sequence traits , that may have fostered their formations have likely been lost ( the half life is 120 mys for mammals and 10 mys in Drosophila [30] ) . While sequence analyses of ancient chimeric genes provide little mechanistic insight , a sample of young chimeric genes that potentially retain these sequence features may . A second difficulty arises from the limited number of young chimeric genes that are thought to have arisen by DLR . While several case studies exist , evolutionary analyses demonstrating that the new chimeras are functional are largely lacking [24 , 27 , 31] . Here we report on a large-scale experimental genomic screen for young chimeric genes generated by DLR within the D . melanogaster subgroup . We utilized an integrated approach based on fluorescent in situ hybridizations ( FISH ) , Southern hybridizations , expression and transcript experiments , BLAST queries , and evolutionary analyses . This approach allowed us to focus on dispersed duplication events , ignoring tandem duplications . Consequently , the total number of chimeric formations are likely larger than the total we report on here . Nonetheless , our results show that , rather than providing redundant copies , dispersed duplication events via DLR have generated new chimeric structures at a high frequency . Interestingly , none of these chimeric structures involved two or more genic sequences; all chimeric regions were formed from the fusion of the duplicated loci and intergenic sequences . Furthermore , we provide strong evidence that REs , in particular the TE family DNAREP1 , are a major mediator of these events . Finally , using multiple well-established methods [6 , 7 , 32–34] , we demonstrate that most of these new chimeric genes are functional .
Two cDNA unigene libraries from D . melanogaster comprised of ∼7 , 000 cDNA probes were used for cFISH experiments over all tested species . Each hybridization generated at least two images for each species . In total , our experiment produced ∼112 , 000 primary images . Including those probes that gave weak or paradoxical signals , the Drosophila Gene Collection ( DGC ) library version 1 . 0 set resulted in 266 candidates . The unigene library included 1 , 000 cDNA probes , most of which were included in the DGC 1 . 0 library . From this set , 5 new genes , jingwei [33] , Hun [32] , sphinx [34] , monkey-king [35] , and Dntf-2r [36] have previously been described . To exclude false positives from the 266 candidates , we carried out Southern hybridizations and conducted BLAST searches against the available genome sequences of D . simulans ( droSim1 ) , D . yakuba ( droYak1 ) , D . sechellia ( droSec1 ) , D . melanogaster ( dm2 ) and D . erecta ( droEre1 ) ( http://genome . ucsc . edu ) ( Figure 1 ) . The Southern and BLAST analyses confirmed 17 young duplicates generated through DLR ( Table 1; Figure 2 ) . The genomic sequences of all 17 dispersed duplicates contain the intron ( s ) and/or non-coding flanking sequences that exist in their parental copies , suggesting that the new genes originated through DLR . In addition , we also identified ten new copies of retrogenes and 53 young copies of REs including retroelements and other repetitive sequences . In this report , we have focused on the 17 dispersed duplicates and investigate possible DLR mechanisms that generate dispersed duplications . Interestingly , the kep1 gene family has six new duplicates that have been dispersed to different chromosomal locations , while the other 11 gene families have only a single new duplicate ( Table 1 ) . Thirteen of these duplications are intrachromosomal , and 4 are interchromosomal ( Table 1 ) . Two putative pseudogenes exist in this list: CR33318 and CR9337 . CR33318 is found only in D . melanogaster , however CR9337 has a disrupted reading frame in D . melanogaster but is intact in D . sechellia and D . simulans . Mapping these results onto the species tree reveal an age <8 mys for almost all these origination events except the 12-my-old CG5372 ( Figure 2 ) . Excluding the two putative pseudogenes ( CR33318 and CR9337 ) paralog-specific reverse transcriptase ( RT-PCR ) experiments detected transcripts for all paralogs . Twelve out of these 15 duplicates display differential expression patterns from their parental copies in development and/or sex ( Table S1 ) . These observations indicate that most of the new genes have evolved divergent expression patterns , and that generally the patterns are more restricted . To examine whether the new duplicates have evolved chimeric gene structures , we utilized previously reported cDNA sequences , RACE , or RT-PCR based on computationally predicted structures ( Materials and Methods ) . Among the 17 new genes , 13 were found to have evolved chimeric gene sequences through the recruitment of flanking sequence near the insertion site or as the result of extensive deletions ( CG5372 , CG9902 , CG4021 , CG3875 , CG3927 , CR9337 , CG7635-r , CG3101-r , CG3071-r , d-r , Dox-A3-r , Hun , and klg-r; Figure 3 ) . Among these chimeric genes , 11 can encode chimeric proteins ( CG5372 , CG9902 , CG4021 , CG3875 , CG3927 , CR9337 , CG7635-r , CG3101-r , CG3071-r , Hun and d-r ) . For example , d-r and CG9902 have both recruited novel coding regions following their duplications , and possibly in conjunction with their deletions events that followed ( Figure 3 ) . These observations reveal that the majority of young duplicated genes have evolved chimeric gene structures . In addition , it is notable that the chimeric genes that we have detected involve only the duplicated loci and integenic sequences . This suggests that for dispersed duplication events , the formation of chimeric genes by recombining two or more genic sequences may be relatively rare . To test for functional constraint , we conducted substitution analyses by estimating the Ka/Ks ratio for both paralogous and orthologous comparisons . For the paralogous comparisons , our conservative null hypothesis was that the parental genes are under strong functional constraint with the new copy subject to no constraint ( a pseudogene ) . These estimates suggest that most of the genes are under functional constraint: Ka/Ks values are lower than 0 . 5 for 8 genes , lower than 1 but higher than 0 . 5 for 5 genes , and ∼1 for 2 genes ( Table S2 ) . Furthermore , analyses of the functional domains for these genes ( Materials and Methods ) , revealed that almost all genes have Ka/Ks ratios lower than or close to 0 . 5 ( Table S2 ) . For orthologous comparisons , the null hypothesis was that the new copies are pseudogenes ( Ka/Ks = 1 ) . The results were similar , showing that Ka/Ks ratios are significantly less than 1 for most genes except CG3071-r ( Ka/Ks = 2 . 3091 ) and CG8490-r ( Ka/Ks = 1 . 2230 ) , indicating the possibility that positive selection may be acting on these two ( Table S3 ) . The statistical tests of the null hypothesis of neutrality [1] in the paralogous and orthologous comparisons reveal that most of these new genes are under significant functional constraint over the tested coding sequences . These complementary analyses of expression , gene structure , and nucleotide substitution suggest that all 15 new genes are functional and that many of these have undergone neofunctionalization by evolving new gene structures with new expression patterns . The classical models of gene duplication assume a completely redundant ( in sequence and function ) duplicate copy [1 , 2] . In these models the most likely outcome is that one copy will become non-functionalized , with a low probability that one or the other becomes neofunctionalized or subfunctionalized through subsequent mutations [37 , 38] . However , our results show that the majority of new duplicates generated through DLR in Drosophila are not structurally , and are thus unlikely to be functionally , identical to their parental copies . It is also a general result that DLR is an important mechanism for the generation of dispersed genes with novel functions , adding to other potential mechanisms [39] . Interestingly , Katju and Lynch [40] have recently found that many new duplicates in C . elegans have unique exons in one or both members of a duplicate pair . Consistent with our observations , these latter cases are also likely DLR-derived duplicates that have recruited new gene fragments and have evolved stable chimeric structures . Having established that 15 of these new duplicates are likely functional , with many having chimeric structures , we then investigated the mutational mechanisms that generated them . Data , largely originating from detailed sequence analyses of human disease-related loci , have shown correlations between structural variation and REs , most notably Alu elements in primate genomes [13 , 22 , 23 , 25 , 28 , 41] . Though a causal relationship between the repetitive elements and segmental duplications is difficult to establish , several studies have argued for their causative role in genomic rearrangements through NAHR . Based in part on these findings , we were interested in whether there was evidence for repetitive sequence surrounding these duplicated regions . We identified both 5′ and 3′ breakpoints for each young duplicate by comparing genomic sequences of each of these new gene duplicates with its parental copy ( Table 2 ) . Interestingly , we observed REs at or near the breakpoints for 10 out of the 17 duplicates ( including the 2 duplicates that are likely pseudogenes ) ( Table 2; Figure S1 ) . These REs consist of 7 TEs , 2 satellite sequences , and 1 simple repeat . They are associated with the new genes that are in different genomic locations , suggesting independent events . Furthermore , all TEs belong to the DNAREP1 family , the largest TE family in Drosophila which has very diverged members [42 , 43] . Among these 10 pairs associated with REs , 5 have shared repeats at or near the breakpoints of both the parental and the new duplicate copies ( Table 2 ) . For these 5 paralog pairs , 4 ( CG3875-CG3927 , mkgr-mkgr2 , CG3101-CG3101-r and CR9337-CR9337-r ) maintain very high sequence identity over the flanking elements; the remaining CR9337-CR33318 pair , though both harboring DNAREP1 sequence at their 5′ ends , provides a weak alignment . The other five paralog pairs contain a repetitive element at the breakpoint of one copy ( Table 2; 2 examples with highly similar TEs shown in Figure 4 ) . In addition , klg-r , CG7635-r and CG8490-r ( not included in the ten above ) were found next to sequencing gaps in the genomic databases ( Table 2 ) , and resequencing these regions resulted in sequence profiles characteristic of repetitive sequences ( data not shown ) . If these are included , the majority of new duplicates ( 13/17 , 76 . 5% ) are associated with repetitive elements . Four lines of evidence indicate that this association has not been observed by chance . The first is based on orthology assignments available from current genome databases , indicating that all ten in our set are euchromatic and not on the 4th chromosome . High-resolution analyses of D . melanogaster TEs have verified that the paracentromeric regions of the major chromosome arms and chromosome 4 harbor the highest densities of TEs [44] . Second , simulations show that the probability that the number of genes flanked by TEs ≥7 given the sample size of seven genes ( with 14 breakpoints ) is low ( p < 0 . 05 ) given a TE-free region ( TFR ) of ∼15 kb or larger ( Figure S2; Materials and Methods ) . Despite TE differences between species , 15 kb is less than half the mean TFR found in D . melanogaster [44] . Given that the TEs in our dataset are comprised primarily of DNAREP1 family members , the distance is even greater . Furthermore , the probability that both paralogs contain the same TE sequence in their flanking regions , as three ( and possibly four ) do in our dataset , is much lower ( Table 2; Figure S1 ) . Finally , our data reveal a gradation of degeneration in the TEs and other REs with the ages of the gene duplicates that the repeats flank ( Figure 5 ) . This gradation is consistent with observed degeneration rate of functionless elements in Drosophila [30] , as well as any potential internal deletions that could be part of a self-regulation system as seen in D . melanogaster TEs [45] . The striking association with REs provides evidence for the relationship between RE sequences and genomic rearrangements leading to novel functions . This relationship differs from previous reports of TE themselves becoming part of a novel transcript in D . melanogaster [46 , 47] . Instead , our dataset supports a model whereby REs are mediating the recombination of flanking sequences to form chimeric products that do not include RE sequence . The precise mechanism defining “RE-mediation” would likely be NAHR or the mobilization of flanking sequence through the activity of the DNAREP1 transposons . Recent studies of DNAREP1 elements suggest a burst of activity occurred just prior to or during the formation of the D . melanogaster subgroup , followed by nearly complete inactivation ∼5–10 mya [42] . Interestingly , there is evidence of a very recent revival of activity in the D . yakuba lineage [43] . If these estimates on inactivity are correct , NAHR would be the most likely mechanism generating the rearrangement in our dataset . This possibility is also supported by the identified non-mobile repeat sequences that are associated with the new chimeric genes ( Table 2 ) . However , if DNAREP1 has been active in the D . melanogaster subgroup for a longer period than reported , as implicated by the observation in D . yakuba [43] , and if this class of TEs does in fact mobilize flanking DNA , a combination of mechanisms is possible . Alternatively , the REs flanking the new duplicates could be the result of larger duplications that included the REs ( segmental duplication ) , rather than the REs mobilizing the region . However , we would expect that under this hypothesis we would see longer stretches of identity outside REs . Inspecting the flanking regions of our dataset indicate that identity is lost in close proximity with the repetitive sequences . A second alternative hypothesis is that the repetitive sequence presents a preferential site for strand breakage . Similar suggestions have been made for Alu , satellite repeats , and other sequence demonstrating fragility [23 , 31 , 48] . If imperfect repair were to follow strand breakage , this too would be akin to a nonhomologous end-joining event and would support the classical view of exon-shuffling . Further experimental work is needed to address this possibility . Our observation that there is an excess of repetitive elements around dispersed functional duplicates is of general importance in light of advancements in identifying copy number variation in other model organisms , and the increased recognition for the role of repetitive sequences in shaping chromosomal architecture [14 , 22–26 , 31 , 49 , 50] . Despite these advancements , little is known about the potential non-deleterious outcomes that such rearrangements may present . Our work helps fill this void by providing an extensive chimeric gene dataset that is supported by experiments that test for functionality . Evidence from previous case studies has indicated that once a duplicate has been generated the recruitment of exons and/or flanking gDNA is a heterogeneous process [32 , 51] . Within our dataset , we also observe this . The first instance is the direct recruitment of genomic DNA flanking the insertion site of the new copy . Eight new genes , representing eight gene families ( CG9902 , CG5372 , Hun , CG7635-r , CG3701-r , d-r , Dox-A3-r and klg-r ) , were created this way . The second involves dramatic mutations within the new duplicates . In the kep1 family , numerous deletions in the duplicated 3′ regions have resulted in varying peptide sequences in the C terminal ( Figure 3 ) . In CG8490-r , both the start and the stop codons have been shifted , resulting in different peptides at both N and C termini . Finally , CG3101-r has recruited part of its previous intron 3 , which becomes the 5′ UTR and a short stretch of protein-coding sequence in the new duplicate gene . We have used ∼7 , 000 cDNA probes to screen new gene duplicate copies . The estimated number of genes in the genome is ∼14 , 000 . The total number of new gene duplicates can be estimated as 17/7 , 000 × 14 , 000 = 34 , over an evolutionary time equal to ∼20 mys ( the sum of the branch lengths of the D . melanogaster subgroup ) . Thus , on average , the origination rate is 34/20 = 1 . 7 per mys per genome , or 0 . 121 × 10−9 per year per gene . We note that , because our method ignores tandem duplicates , and because our FISH probes were all based on D . melanogaster sequence , this is an underestimate . However , this rate is an order of magnitude higher than the gene duplication rate estimated in yeast [52] but still 30 times lower than a previous estimate that were based on the assumption of a molecular clock [53] . Our estimate may not be inconsistent with previous estimates [53] because our focus was much narrower , investigating DLR events only . Only two new duplicates ( d-r and Dox-A3-r ) in the yakuba-santomea-teissieri lineage ( yakuba lineage ) were observed , while 5 new duplicates were detected between the common ancestor of melanogaster and yakuba and the common ancestor of the melanogaster complex ( Figure 2 ) . In addition , we did not detect any new duplicate in D . erecta . This may be a technical result attributable to the difficulty of hybridization with D . erecta polytene chromosomes , or sequence divergence relative to our probes . Alternatively , the putative inconsistent duplication rate may be associated with episodic activities of transposons or repetitive sequences . For example , the transposon DNAREP1 members were associated with 5 new duplicates in the kep1 gene family and CG9902 . As noted above , it has been suggested that there was an active episode of DNAREP1 before the D . melanogaster lineage separated from the D . yakuba lineage and then again within the D . yakuba lineage [42 , 43] . Previous investigations have revealed several important roles for REs in the generation of evolutionary novelties including the donation of their own sequences into protein coding regions [46 , 47 , 54 , 55] , retrotransposing and recruiting novel gene sequence [5] , increasing genic diversity in the maize genome by the helitron-like transposons [56] , potentially providing greater overall genome plasticity [16 , 57] , and elevating expression of a nearby insecticide resistant gene [58 , 59] . The observation reported here further demonstrates a mechanistic role for REs in mediating the origins of new genes by facilitating gene recombination . The precise mechanism for this recombination is unclear , but likely include NAHR , as implicated by both TEs and non-TE repetitive sequences being detected , and NHR as a consequence of transposon enzymatic activities [43] . However , the conventional NAHR model is much more likely between the homologous repeats that are located on the same chromosome [22] . Four of the 17 new genes identified are on different chromosomes from their parental genes . These four new genes may have been generated by a different homology-dependent recombination model that assumes a replication-dependent mechanism involving no crossover [22] , the explicit model depicted in Figure 8 .
In order to screen for young chimeric genes systemically , we designed an experimental genomics approach using the D . melanogaster species subgroup as a comparative model system . This subgroup includes D . melanogaster ( hereafter abbreviated as mel in presented tables and figures ) , D . simulans ( sim ) , D . mauritiana ( mau ) , D . sechellia ( sec ) , D . yakuba ( yak ) , D . teissieri ( tei ) , D . santomea ( san ) , and D . erecta ( ere ) . D . orena was excluded from analyses because of its unclear placement in the phylogeny . The phylogeny of this subgroup is well resolved [60 , 61] and the divergence times among these species provide a considerable range over which to detect the presence of young genes . The polytene chromosomes of the salivary gland of Drosophila allow detection of gene copy number using a fluorescent in situ hybridization ( FISH ) approach . Therefore , we can use cDNA probes to visualize FISH signals that are about 100 kb away from each other in the species of D . melanogaster subgroup , and count the signal number in each species [34] . We carried out dual-color FISH on the polytene chromosome preparations of the aforementioned 8 species . Our probe sets comprised 5 , 928 full-length D . melanogaster cDNA clones from the Berkeley Drosophila Gene Collection ( DGC ) version 1 . 0 ( http://www . fruitfly . org/DGC/index . html ) and about 1 , 000 cDNA clones from an early Drosophila Unigene Library ( Research Genetics ) . Probes were labeled with digoxiginin ( DIG ) or biotin using PCR [34 , 62] . Polytene chromosomes from four species were simultaneously squashed on a slide and then hybridized with a pair of DIG and biotin labeled probes [34] . For a given probe , FISH is capable of resolving two signals across two adjacent polytene bands , which is equivalent to ∼100 kb in linear DNA sequence . As a result , all duplicates we report in this study have been involved in translocations; they are not tandem duplications . The probes that revealed extra signals in a particular lineage were subject to further confirmation using southern hybridization . Genomic DNAs of the eight species were extracted using the Puregene DNA isolation kit ( Gentra Systems ) . DNAs digested with restriction enzymes were separated on agarose gels and transferred to nylon membranes ( Roche Molecular Biochemicals ) by Southern blotting . The DIG-labeled probes were hybridized to the membrane to further confirm the copy number in different species . In addition , homology searches were carried out for those new genes that fell within sequenced genomes ( http://genome . ucsc . edu ) . To identify breakpoints and examine the type of sequence surrounding them , the genomic sequences of each pair of duplicate and parental copy , along with 5′ and 3′ flanking sequences , were aligned using the bl2seq software with default settings [63] . The length of the 5′ and 3′ flanking sequences for each pair was chosen to ensure that it extends 1kb beyond the point where sequence identity disappears . Breakpoints of duplicates were determined as the last nucleotide showing sequence identity between parental and new copy . For a multiple-copy gene family , the parental copy was defined as the copy that has the highest similarity to the new copy . RepeatMasker ( http://www . repeatmasker . org/ ) was used to identify whether there is repetitive sequence within a 100 bp window centered at each breakpoint . To examine the evolutionary forces operating on the new duplicates , we calculated synonymous ( Ks ) and non-synonymous ( Ka ) divergence between all paralogs except for the pseudogene CR33318 ( we included the putative pseudogene CR9337 and CR9337-r because they are still intact in the D . simulans complex ) . In addition , we also conducted substitution analyses between orthologous copies in different species . For 11 young duplicates we retrieved their orthologs from a second species's genome , and therefore also calculated Ka and Ks between the orthologous pairs . Estimates were obtained using MEGA 3 . 1 [64] . A Z-test implemented in MEGA 3 . 1 was used to test if Ka/Ks ratios deviate from the neutral expectation ( Ka/Ks = 1 ) . We tested functional constraint in the whole gene coding region and the functional domain separately . To define the functional domains , the coding sequences of genes were translated into the protein sequences . Then we performed rps-BLAST to detect whether the newly translated protein sequences have functional domains using a cutoff line E < 0 . 01 on NCBI website http://www . ncbi . nlm . nih . gov/Structure/cdd/wrpsb . cgi . Our approach is capable of observing three kinds of new duplicates , ( 1 ) direct duplicates that still keep intron ( s ) or flanking non-coding sequences , ( 2 ) retroposed copies that have lost ancestral introns , and ( 3 ) copies that have no obvious sequence features identifying them as either created by retroposition or direct duplication . Tandem duplication can be resulted from either replication slippage or DLR , but the assumption is that those dispersed duplicates across long chromosome distance , or between chromosomes , have originated through DLR . In this study , we only considered direct dispersed duplicates that were derived through DLR . For each of these duplicate genes , we designed copy-specific RT-PCR primers . RT-PCR experiments were carried out using cDNA from 5 developmental stages: embryo , instar larva 2 ( L2 ) , instar larva 3 ( L3 ) , pupa and adult . Total RNA was extracted from these samples using RNAeasy Mini RNA extraction kit ( Qiagen ) . To avoid contamination of genomic DNA , total RNA was treated with Dnase I ( amplification grade , Invitrogen ) prior to first strand synthesis . First strand cDNA was synthesized using Oligo-dT and SuperScript II Rnase H- reverse Transcriptase ( Invitrogen ) . All RT-PCR products were sequenced for verification . To establish the gene structures of the new genes , four types of data were used: ( 1 ) the draft genomes of D . simulans ( droSim1 ) , D . yakuba ( droYak1 ) , D . sechellia ( droSec1 ) , and D . erecta ( droEre1 ) ( http://genome . ucsc . edu ) were queried and provided addition verification and gDNA for primer design; ( 2 ) For those duplicates whose full length cDNAs are available in public databases ( http://www . ncbi . nlm . nih . gov/Database/ ) , we mapped the cDNA to their genomic positions if draft sequence was available; ( 3 ) For those duplicates without cDNA , and whose sequences have diverged enough to allow copy-specific primers , we carried out rapid amplification of cDNA ends ( RACE ) ; ( 4 ) For those duplicate pairs that are too similar to allow copy-specific primers , and for those that resulted in no RACE product ( possibly due to low expression levels or long ends ) , we used the Softberry software [65] to obtain a tentative chimeric gene structure prediction . We then tested these predictions using RT-PCR . To establish an approximate chromosomal position ( interstitial or not ) for each of these genes , we used the D . melanogaster genome as a reference . We carried out BLAST queries of the D . melanogaster genome using sequence flanking each of the genes . These flanking regions were then used to query available genome draft sequence ( http://www . ncbi . nlm . nih . gov/Database/ ) in order to determine orthologous chromosomal regions . The cytological positions were then extracted using NCBI's MapView ( www . ncbi . nlm . nih . gov/mapview/ ) . Two new copies ( CG7635 and klg ) , fell between sequence gaps . For these two we determined their approximate position based on our FISH images . To assess the significance of our observed association between TE sequences and the flanking regions of the paralogs , we carried out simulations based on the known frequencies of TEs in D . melanogaster [44] . The mean TE-free region ( TFR ) is 23 , 878 , with a median of 1 , 992 . The difference between the mean and the median results from the clustering of TEs within the pericentric regions and the fourth chromosome . However , the identified new genes are non-pericentromeric regions in which the density of TEs is much lower and there are few cases of non-random insertions to one particular locus . Therefore , we carried out simulations over a range of normally distributed TFRs in a conservative assumption of the 15 kb average . The length of each TE was normally distributed with a mean of 4 kb . The total length of simulated chromosomes was kept at ∼ 20 Mb . 14 breakpoints were introduced randomly into the sequence ( seven paralog pairs where only one copy is associated with TE sequence ) and an association was considered if the breakpoint was within 300 bp . This distance was also chosen to be conservative , given the distances observed in our data . 10 , 000 iterations were run and the upper 5% tail was calculated from the resulting distribution . | In numerous organisms , many new genes have been found to originate through dispersed gene duplication and exon/domain shuffling . What recombination mechanisms were involved in the duplication and the shuffling processes ? Lack of the intermediate products of recombination that share adequate sequence identity between homologous sequences , or the parental sequences from which the new genes were derived , often makes answering these questions difficult . We identified a number of young genes that originated in recently diverged branches in the evolutionary tree of the eight Drosophila melanogaster subgroup species , by using fluorescence in situ hybridization with polytene chromosomes . We analyzed the genomic regions surrounding 17 new dispersed duplicate genes and observed that most of these genes are flanked by repetitive elements ( REs ) , including a large and diverged transposable element family , DNAREP1 . Several copies of these REs are kept in both new and parental gene regions , and their degeneration is correlated with the increasing ages of the identified new genes . These data suggest that REs mediate the recombination responsible for the new gene origination . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
] | [
"evolutionary",
"biology",
"drosophila",
"1"
] | 2008 | Repetitive Element-Mediated Recombination as a Mechanism for New Gene Origination in Drosophila |
In Magnaporthe oryzae , the causal ascomycete of the devastating rice blast disease , the conidial germ tube tip must sense and respond to a wide array of requisite cues from the host in order to switch from polarized to isotropic growth , ultimately forming the dome-shaped infection cell known as the appressorium . Although the role for G-protein mediated Cyclic AMP signaling in appressorium formation was first identified almost two decades ago , little is known about the spatio-temporal dynamics of the cascade and how the signal is transmitted through the intracellular network during cell growth and morphogenesis . In this study , we demonstrate that the late endosomal compartments , comprising of a PI3P-rich ( Phosphatidylinositol 3-phosphate ) highly dynamic tubulo-vesicular network , scaffold active MagA/GαS , Rgs1 ( a GAP for MagA ) , Adenylate cyclase and Pth11 ( a non-canonical GPCR ) in the likely absence of AKAP-like anchors during early pathogenic development in M . oryzae . Loss of HOPS component Vps39 and consequently the late endosomal function caused a disruption of adenylate cyclase localization , cAMP signaling and appressorium formation . Remarkably , exogenous cAMP rescued the appressorium formation defects associated with VPS39 deletion in M . oryzae . We propose that sequestration of key G-protein signaling components on dynamic late endosomes and/or endolysosomes , provides an effective molecular means to compartmentalize and control the spatio-temporal activation and rapid downregulation ( likely via vacuolar degradation ) of cAMP signaling amidst changing cellular geometry during pathogenic development in M . oryzae .
Eukaryotes , ranging from yeasts to multicellular metazoans , interact with their environment , constantly sampling it for physico-chemical signals or cues for proper growth and development . Extracellular ligands or stimuli detected by membrane bound GPCR ( G- protein coupled receptors ) are transmitted to the cell interior by heterotrimeric ( αβγ ) guanine nucleotide binding proteins ( G-proteins ) , which function as intrinsic molecular switches . Ligand activated receptors promote the exchange of GDP to GTP on cognate GαS subunit , triggering its dissociation from the βγ , thereby rendering it active to signal downstream [1] . Both GαS·GTP and Gβγ moieties subsequently propagate the signal through a host of downstream effectors , which include ion channels , adenylate cyclases , phosphodiesterases and phospholipases [2] , [3] . The foremost of these is adenylate cyclase that synthesizes the second messenger Cyclic AMP ( cAMP ) from ATP . Active signaling by the GαS·GTP persists until the bound GTP is hydrolyzed to GDP , by the intrinsic GTPase activity of GαS , permitting GαS to re-associate with Gβγ to form an inactive complex , and thereby commencing a fresh cycle of signaling [1] , [4] . The duration of active signaling by GαS is dependent on the nucleotide state of the GαS subunit . In addition to the intrinsic GTPase activity of GαS , the hydrolysis of GTP to GDP is regulated and fine-tuned by RGS proteins , which ultimately influence the duration of active G-protein signaling [5] , [6] . RGS proteins are highly conserved negative regulators of G-protein signaling and function by stabilizing the switch regions on the GαS subunit which undergo a conformational change upon GTP hydrolysis , thereby accelerating the rate of hydrolysis of bound GTP to GDP on the active GαS subunit [7] , [8] , [9] , [10] , [11] . In conjunction with upstream GPCRs , heterotrimeric G-proteins regulate a wide-range of cellular and developmental pathways including aspects of nutrient sensing , mating and pathogenesis in fungi and yeasts [12] . For example , they are involved in sensing glucose in S . cerevisiae and S . pombe; amino acids in C . albicans and C . neoformans [13] , [14] , [15] , [16] and carbon sources in N . crassa , A . nidulans and B . cinerea [17] , [18] , [19] . Additionally , G-proteins are involved in the pheromone response pathway in S . cerevisiae [20] and regulate mating and fruiting body formation in ascomycete fungi such as A . nidulans , N . crassa and M . oryzae , as well as basidiomycetes such as U . maydis and C . neoformans [12] , [21] , [22] , [23] , [24] , [25] , [26] , [27] , [28] , [29] . G-protein signaling also plays an important role in regulating disease causing ability and progression ( infection structure formation , pathogenicity factors or toxin production ) in many plant and animal pathogens . A wide range of filamentous phyto-pathogens such as C . parasitica , M . oryzae , U . maydis , B . cinerea , F . oxysporum , C . trifolii , S . nodorum and A . alternata disrupted for trimeric G-proteins are defective in pathogenicity [12] , [23] , [24] , [25] , [30] , [31] , [32] , [33] , [34] . Similarly , functional G-protein cascades play a crucial role in establishing pathogenicity in human pathogenic fungi like C . neoformans A . fumigatus and C . albicans [14] , [16] , [27] , [28] , [35] , [36] . In general , when and where G-protein signaling modules are activated in the context of cell shape and cellular geometry has a profound implication on downstream responses and behavior . As a result , cells have evolved complex molecular mechanisms to compartmentalize and thereby control the spatial and temporal dynamics of signaling pathways . In mammalian cells , signal compartmentalization is achieved by sequestration and/or anchoring of key regulators and effectors on different subcellular organelles , supramolecular complexes or scaffold proteins such as AKAPs ( A-kinase anchoring proteins ) [37] , [38] . Anchoring of signaling proteins endows the cell with a host of advantages , which include the ability to specifically target and activate signaling pathways or components within a defined region of the cell , thereby preventing non-specific or off-target effects . Furthermore , signal anchoring per se can help the cell integrate and achieve modularity of signaling responses and ultimately regulate the amplitude and temporal duration of signaling in a precise manner [39] . The endosomal system represents a highly dynamic and biochemically specialized continuum of membranous tubulo-vesicular compartments involved primarily in the sorting , delivery and degradation of a diverse array of cargoes ( macromolecules , biological signals ) and maintaining cellular homeostasis . Cargo can enter the endolysosomal pathway either through endocytosis or internalization that takes place at the cell surface or through transport from internal biosynthetic pathways . The endosomal system is broadly categorized into early , late/MVBs ( multi vesicular bodies ) and recycling compartments . Each compartment is inherently programmed to carry out specific cellular functions . Early endosomes represent the first sites within which internalized components are deposited . Upon internalization , the components may be recycled back to the plasma membrane or may enter a late endosome/MVB for transport into the vacuole/lysosome . Early endosomal compartments usually mature over time , become increasingly acidic , and undergo homotypic fusions to form either late endosomes or accumulate intraluminal vesicles to develop MVBs . As they mature , the early endosomes first lose the Rab5 GTPase and eventually acquire Rab7 GTPase thereby becoming competent to undergo fusion with the degradative vacuole [40] , [41] . Endosomes are known to serve functions beyond cargo sorting . Recent findings suggest that endosomal compartments additionally function as signaling anchors/platforms , integrating membrane trafficking and intracellular signal transduction [42] , [43] , [44] , [45] , [46] , [47] . Upon hydration , the asexual spores produced by M . oryzae germinate to form slender germ tubes , which respond to inductive surface cues such as hydrophobicity and hardness , to undergo morphogenic transitions to ultimately form an infection cell known as the appressorium . One of the most critical steps during early pathogenic differentiation in M . oryzae is the “hooking” stage wherein the germ tube tip switches from polarized to isotropic growth ( 3–4 hpi ) and subsequently forms an appressorium ( 8–12 hpi ) . It is thought that hooking constitutes a “recognition phase” in which surface characteristics are sensed and assessed prior to commitment to appressorium formation [48] , [49] . A host of genetic studies carried out over the last decade have established that components of the G-protein/cAMP signaling cascade are highly conserved and play an important role in regulating various aspects of asexual development , appressorium formation , disease establishment and progression in M . oryzae [23] , [50] , [51] , [52] , [53] , [54] , [55] . However , little is known about the subcellular dynamics and/or the spatio-temporal organization of the critical cAMP-signaling module . In this study , we determine the intracellular localization of the G-protein cascade in M . oryzae with particular emphasis on the dynamics of active signaling components during the crucial stages of host surface sensing and recognition . We further demonstrate that the late endosomes represent important sites of active G-protein signaling and that the disruption of late endosomal integrity results in the disruption of host surface based signaling and subsequent disease causing ability . Our results unravel an elegant adaptive mechanism in M . oryzae for anchoring and trafficking of G-protein signaling components for cAMP synthesis during initiation of infection-related growth and development .
RGS proteins ( Regulator of G-protein Signaling ) play a crucial role in controlling the intensity and duration of early G-protein signaling [1] , [6] . We utilized live cell imaging to gain insight into the spatial and temporal dynamics of Rgs1 in M . oryzae , particularly during the early events of surface recognition and appressorium morphogenesis . Imaging was performed on germinating conidia at 2 and 4 hpi on an inductive surface . Ungerminated conidia ( 0 hpi ) were used as a control . Rgs1-mCherry ( mC ) was present on less dynamic punctate vesicles in ungerminated conidia; however it localized to highly dynamic and distinct tubulo-vesicular structures during the stages of germination and hooking ( Figure 1A ) . We next asked if the formation of the tubulo-vesicular structures or the observed dynamics of the Rgs1 compartments was influenced by inductive surface characteristics . On a non-inductive surface ( 1% agar ) that does not support appressorium formation [54] , Rgs1-mC failed to localize to dynamic tubulo-vesicular structures , but was predominantly vacuolar or present as small puncta in developing germ tubes ( Figure S1A in Text S1 ) . Likewise , the pth11Δ mutant , which is defective in sensing and responding to inductive surface cues [55] , lacked such tubulo-vesicular structures and showed a predominantly vacuolar Rgs1-mC signal in the conidial cells ( Figure S1B in Text S1 ) . Rgs1-mC localized to tubulo-vesicular compartments at identical time points in the WT control strain . Moreover , the dynamic mobility of Rgs1-containing structures was not random but inherently bidirectional , with a subset of the population moving towards the incipient appressorium , while a few punctae trafficked towards the terminal cell of the conidium ( Figure 1B; Video S1 ) . Taken together , we conclude that in response to inductive surface cues , such as hardness and hydrophobicity , Rgs1-mC localizes to dynamic tubulo-vesicular compartments that traverse bi-directionally during appressorium initiation in M . oryzae . In the absence of such inductive cues or signals , as in the pth11Δ strain , Rgs1 is predominantly targeted to the vacuoles . We have previously demonstrated that Rgs1 physically interacts with and negatively regulates the GTP-bound MagA in vitro [54] . In order to understand the physiological relevance of Rgs1 localization to the dynamic compartments especially in the context of active G-protein signaling , a M . oryzae strain expressing both Rgs1-mC and a GFP-tagged constitutively active MagA/GαS subunit [54] ( MagAG187S ) was generated by insertion of the GFP coding sequence in the alpha B-alpha C loop of MagA [56] . The wild-type MagA-GFP was used as control . Remarkably , we observed a differential pattern of localization between the constitutively active MagAG187S variant and the wild-type MagA in the vegetative mycelium . The active MagAG187S-GFP predominantly localized to distinct vesicles , whereas the wild-type MagA-GFP was primarily plasma membrane bound ( Figure 2A ) . Next , we tested if the active MagAG187S co-localizes with Rgs1 during early pathogenesis and specifically during the hooking stage ( 4 hpi ) . Indeed , active MagAG187S-GFP vesicles co-localized with Rgs1-mC ( Figure 2B ) , as determined by line scan analysis ( Figure 2C ) and verified quantitatively by Pearson's coefficient analysis ( value = 0 . 56 ) ( Figure S2A in Text S1 ) . Furthermore , utilizing a GFP-trap based pull-down assay with whole cell extracts from MagAG187S-GFP and a control strain ( cytosolic-GFP expressing strain ) , we found that MagAG187S physically interacts with Rgs1 ( Figure 2D ) . We conclude that Rgs1-containing vesicles and tubules represent sites of active G-protein signaling in M . oryzae . Furthermore , we infer that a preferential association exists between Rgs1 and the GTP-bound MagA on such dynamic structures during early stages of surface recognition and appressorium initiation . Pth11 , the CFEM-domain containing non-canonical receptor , is predicted to function upstream of the cAMP signaling pathway in M . oryzae [55] . Given that active MagA and Rgs1 colocalized on dynamic structures during early stages of appressorium initiation , we sought to investigate the subcellular localization of Pth11 at this crucial stage of pathogenic development . We found that in a manner identical to Rgs1 , a significant fraction of Pth11-mCherry localized to intracellular punctate and/or dynamic tubular structures . In addition , Pth11-mC also localized to the plasma membrane of the germ tube and incipient appressorium ( Figure 3A ) . Although predicted to be a GPCR involved in cAMP pathway [55] , there is no evidence supporting the interaction of Pth11 with requisite downstream G-protein signaling components . To assess such an interaction , we performed GFP-trap based pull-down assays using whole cell lysates from a Pth11-GFP strain . Relevant immuno-precipitated or wild type control samples were analyzed by western blotting using α-Rgs1 antibody . Remarkably , Pth11 physically interacted with Rgs1 during the hooking stage ( 4 hpi ) and not during vegetative growth ( Figure 3B ) . We further confirmed the physical interaction between Pth11 and Rgs1 by performing an RFP-trap based pull-down with total cell extracts from the Pth11-mCherry and a control strain ( cytosolic mCherry expressing strain ) ( Figure 3C ) . Additionally , in a BiFC ( Bimolecular Fluorescence Complementation ) assay Pth11-nYFP and Rgs1-cYFP interacted in vivo during early pathogenesis ( 2 and 4 hpi ) as judged by the YFP signal detected in the plasma membrane and punctate structures within the germ tube ( Figure 3D; Figure S3A in Text S1 ) . Such an interaction between Pth11-nYFP and Rgs1-cYFP was not apparent in the control strains ( Figure S3B ) or during vegetative growth ( Figure 3D ) . Taken together , we conclude that Pth11 , active MagA and Rgs1 localize to dynamic tubulo-vesicular compartments . In addition , Pth11 physically interacts with Rgs1 specifically during appressorium initiation or germ tube hooking . To determine if Rgs1-mC vesicles formed a part of the endolysosomal network , we performed a colocalization analysis with the fluorescent probe lysotracker green ( LTG ) that marks acidic endosomal compartments . A large number of structures with diverse morphologies were stained by LTG , whereas only a small population colocalized with Rgs1-mC , indicating that Rgs1 compartments formed a sub-population of the endosomal network ( Figure S4A in Text S1 ) . The phosphoinositide PI3P has been demonstrated to be a major constituent of endosomal membranes , limiting membrane of vacuoles and membranes of intraluminal vesicles of MVBs [57] . The GFP-FYVE ( Fab1 , YOTB , Vac1 and EEA1 ) probe has been extensively used to mark the PI3P rich structures [58] . Examination of the cellular distribution of GFP-FYVE and Rgs1-mC marked structures revealed a high degree of co-localization , as determined by correlation coefficient ( Pearson's coefficient = 0 . 34 ) and line scan analyses , suggesting a likely endosomal origin for Rgs1-containing vesicles and tubules ( Figure 4A; Figure S2A and B in Text S1 ) . Rgs1-mC structures were not related to filamentous mitochondria or vacuoles as judged by lack of colocalization between Rgs1-mC and mito-GFP , or Rgs1-mC and CMAC ( 7-amino-4-chloromethylcoumarin , a widely used membrane permeable dye that selectively stains fungal vacuoles , [59] , [60] , [61] , [62] , [63] , respectively ( Figure S4 B and C in Text S1 ) . The PI3 kinase , Vps34 , facilitates the synthesis of PI3P on endosomal membranes [64] . We disrupted Vps34 function , using the specific inhibitor LY294002 [65] , [66] , to determine if a persistent supply of PI3P was necessary to maintain the integrity/dynamic nature of Rgs1 compartments . In contrast to the solvent ( DMSO ) control , which displayed extensive dynamic tubulo-vesicular structures at the hooking stage , the LY294002-treated samples were devoid of such Rgs1-mC structures . Instead , there appeared to be an accumulation of Rgs1 protein in the plasma membrane of the developing appressorium as well as vesicular aggregates in the terminal cell of the conidium ( Figure 4B ) . Additionally , we confirmed the specificity of LY294002 by testing its effect on the GFP-FYVE strain . In contrast to the DMSO control , LY294002 treatment led to enlarged compartments containing an accumulation of the GFP-FYVE probe ( Figure S5 in Text S1 ) . Based on these results , we conclude that Rgs1 localizes to the P13P-rich endosomal network , in a manner dependent upon sustained PI3 kinase activity . Based on the selective presence of either of the specific small GTPases , namely Rab5 or Rab7 , the endosomal system is broadly classified into early and late compartments , respectively [67] . To identify the specific endosomal compartments that associate with active GαS and Rgs1 , we carried out a colocalization analysis between Rgs1-mC and GFP-tagged Rab5 or Rab7 ( Figure S6A and B in Text S1 ) . Rgs1-mC clearly colocalized with Rab7 positive late endosomes ( Pearson's coefficient = 0 . 48 ) ( Figure 4C; Figure S2A and B in Text S1; Figure S7 in Text S1; Video S2 ) , whereas no correlation was observed between the localization of Rgs1-mC and Rab5 marked early endosomes ( Pearson's coefficient = 0 . 11 ) ( Figure 4D; Figure S2A and B in Text S1 ) . To additionally assess whether the MagAG187S-GFP or Rgs1-mC harboring compartments were distinct from vacuolar compartments during the hooking stage ( 4 hpi ) , we carried out a co-staining analysis between MagAG187S-GFP or Rgs1-mC with the vacuolar dye CMAC . We found no colocalization between such GFP or mCherry marked compartments and vacuoles within the incipient appressorium . However , we did observe a distinct colocalization between CMAC and GFP or mCherry in the conidial cells ( Figure S8A and B in Text S1 ) . Taken together , these results clearly demonstrate that Rgs1 specifically localizes to the late endosomes . Furthermore , we propose that Rgs1 compartments , and by inference the late endosomes , likely represent sites of active G-protein signaling and/or anchoring in M . oryzae . In polarized cells , either the microtubule and/or the actin cytoskeleton facilitate the transport of a wide range of intracellular vesicular cargo/compartments . We determined the nature of the cytoskeletal tracks engaged by Rgs1 for mobility by treating samples at the hooking stage , with either MBC ( Methyl Benzimidazol-2-yl-Carbamate , to destabilize the microtubule tracks ) or latranculin A ( LatA , to disrupt actin ) [68] , [69] , [70] . Compared to the DMSO control , the dynamics and morphology of Rgs1-mC tubulo-vesicular structures were completely disrupted upon treatment with MBC ( Figure 5 ) . Additionally , MBC treatment caused an accumulation of the Rgs1-mC on the plasma membrane and as intracellular aggregates confined to the incipient appressorium and the terminal cell of the conidium . As expected , Rgs1-mC containing structures were coincident with microtubules ( GFP-α Tub ) , supporting the microtubule-dependent trafficking of Rgs1 ( Figure S9A in Text S1 ) in M . oryzae . In contrast , LatA-treated samples showed typical mobile/dynamic tubulo-vesicular Rgs1 compartments although at a slightly reduced extent compared to the control sample ( Figure S9B in Text S1 ) . Collectively , our results show that the dynamic mobility of Rgs1-containing late endosomal compartments requires an intact microtubule cytoskeleton . Vps39 ( Vacuolar Protein Sorting 39 ) , a key member of the HOPS ( homotypic fusion and vacuole protein sorting ) complex , plays a crucial role in the conversion of early endosomes into late endosomal compartments . Disruption of Vps39 function blocks endosomal conversion , producing merged organelles with characteristics of both early and late endosomes [71] . Our results so far support a model in which the late endosomes function as a likely scaffolding platform for active G-protein signaling . To verify our hypothesis and to understand the physiological relevance of late endosomes in cAMP signaling , we disrupted the VPS39 function in M . oryzae . The vps39Δ mutant showed significant defects in appressorium initiation , formation and maturation , when compared to the wild-type strain at identical time points ( Figure 6A ) . The vps39Δ conidia failed to form proper appressoria even upon prolonged incubation on inductive surface ( 24 h; Figure 6A ) . On an inductive surface , the vps39Δ germ tubes showed likely surface sensing and/or signaling defects ( multiple hooking but failed appressorium formation ) , indicative of a lack of proper activation and regulation of cAMP signaling . In addition , disruption of endosomal acidification and hence its function and organization using bafilomycin A1 ( an inhibitor of V-ATPases ) [72] resulted in a vps39Δ like phenotypic defect in the wild type M . oryzae ( Figure 6A ) . Remarkably , exogenous addition of 8-Br-cAMP completely suppressed such surface sensing and appressorium formation defects in the vps39Δ ( Figure 6A ) . The untreated vps39Δ strain failed to form proper appressoria even at 24 hpi , and instead showed aberrantly long germ tubes . Only 5% and 40% of vps39Δ conidia formed immature appressoria by 10 and 24 hpi , respectively ( Figure 6B ) . We carried out infection assays on barley as well as a susceptible rice cultivar ( CO39 ) to test the ability of the aberrant vps39Δ appressoria to cause blast disease . Disease symptoms were assessed seven and fourteen days post inoculation in barley and rice , respectively . Compared to the wild type , which caused blast disease with characteristic spindle shaped lesions with gray centers , the vps39Δ failed to cause disease or at best elicited only a mild hypersensitive response on rice seedlings ( Figure 6C ) and barley leaf explants ( Figure S10B in Text S1 ) . Furthermore , the vps39Δ strain was additionally compromised for radial as well as aerial growth ( Figure S10A in Text S1 ) . The number of conidia produced by the mutant was nearly ∼2 . 5 fold lesser compared to the wild type . Only 10% of the vps39Δ appressoria were functional or were successfully able to penetrate the host tissue ( Figure S10C in Text S1 ) . In conclusion , anchoring of cAMP signaling via Vps39 activity , and by inference the late endosomal/endolysosomal/HOPS function , is essential for timely and robust cAMP signaling . Additionally , late endosomal function/HOPS is necessary for pathogenesis and disease progression in M . oryzae . The suppression of vps39Δ defects by exogenous cAMP ( Figure 6A ) implies that the mutant is likely impaired for cAMP synthesis and furthermore compromised for Adenylate cyclase function . We therefore assessed the subcellular localization of Adenylate cyclase ( Mac1 or Adc ) in the wild type and the vps39Δ strain at 4 hpi . Disparate from the tubulo-vesicular localization in the wild type , Mac1-GFP was predominantly cytoplasmic and/or vacuolar in the vps39Δ mutant ( Figure 7A ) . Likewise , the subcellular localization of Rgs1-mC , as well as GFP-Rab7 , was compromised in the vps39Δ strain . Both the proteins localized as intracellular aggregates or accumulated in vacuoles present in the conidia ( Figure 7B and 7C ) . Collectively , our results clearly support a model in which the late endosomes/endolysosomes act as anchors or platforms facilitating proper cAMP signaling . We conclude that disruption of late endosomal integrity or stability results in breakdown of cAMP synthesis and/or regulation , and leads to defects in cell signaling necessary for appressorium initiation during M . oryzae pathogenesis .
Cell shape and the spatial segregation of signaling components therein collectively influence how such molecules interact to produce a timely cellular response . Spatial separation of interacting molecules , by localization to different subcellular compartments/organelles , is a widespread mechanism of regulating pathway activity [38] , [73] . Such a mechanism has been found to widely operate in mammalian cells , wherein signal compartmentalization is established via the anchoring of key molecules on scaffolding proteins called AKAPs . In this study , we demonstrate that in the absence of authentic AKAP–like anchors [74] , M . oryzae utilizes the dynamic late endosomal system as a common platform to anchor , transmit and directly regulate G-protein signaling during early stages of infection-related development . Upon conidial germination on an inductive surface , the germ tube tip switches from polarized to isotropic growth ( 3–4 hpi ) . In such a context of altering cell shape , the initial pool of less mobile Rgs1 structures transform into highly dynamic tubulo-vesicular compartments . We infer that such a change is likely in response to inductive host cues and successful activation of G proteins . The ability of the late endosomes to traverse the length of the germ tube and navigate in both directions likely facilitates the recycling and/or the transport of signaling components to or away from the sites of active signaling on the plasma membrane . However , the physical interaction and preferential localization of active MagA on Rgs1-containing dynamic vesicles , suggests that the late endosomes function as ancillary sites ( in addition to the plasma membrane ) of active G-protein signaling in the cytosol , facilitating signal transmission through the complex cytoplasmic milieu . We have previously demonstrated that cAMP levels in M . oryzae are regulated in two distinct phases , i . e . , upregulated during appressorium initiation and down regulated during host invasion [53] . In this context , anchoring on dynamic endosomal structures likely provides M . oryzae with the flexibility and the means to rapidly assemble G-protein signaling hotspots/complexes within defined regions of the cell , namely the tip of the germ tube or within the incipient appressorium; thus gaining the ability to compartmentalize and regulate the intensity and/or duration of cAMP signaling . Furthermore , given the inherent connection ( s ) of the endosomal network to the vacuolar pathway , such signaling hotspots may also be rapidly disassembled to down-regulate signaling by targeting such complexes to the vacuoles , either for sequestration or degradation . Based on the physical interaction with Rgs1 during early stages of appressorium initiation , we implicate Pth11 to be a bona fide GPCR ( albeit non-canonical ) for cAMP signaling in M . oryzae . Furthermore , based on the BiFC experiment , we propose that the interaction between Pth11 and Rgs1 likely occurs on the late endosomes , in the context of and in close proximity to the active MagA . It remains to be seen whether such an association requires post-translational modifications as has been demonstrated for the orthologous Sst2 ( GAP ) and Ste2 ( receptor ) in budding yeast during the mating response [75] . Incidentally , treatment with either MBC or the PI3Kinase inhibitor caused increased accumulation of Rgs1 on the plasma membrane of the incipient appressorium , suggesting that the trafficking of Rgs1 and by inference MagA likely initiates at the outer membrane upon receptor ( Pth11 ) activation . We found that a sustained supply of PI3P is necessary to maintain the integrity and dynamics of the G-proteins anchored on late endosomes . We have recently shown that the N-terminal DEP domain of Rgs1 facilitates specific targeting to such vesicular compartments [76] . Although we have not shown a direct interaction between Rgs1 and PI3P , a previous report has shown that the DEP domain of yeast Sst2 has the propensity to bind specific phospholipids in vitro [75] . Disruption of Vps39 , a member of the HOPS complex , resulted in complete deregulation of surface signaling and appressorium formation in M . oryzae , suggesting that the late endosomal/endolysosomal function and integrity is crucial for proper cAMP signaling in M . oryzae . In vps39Δ , a significant proportion of Mac1-GFP ( Adenylate cyclase ) as well as Rgs1-mC localized to the vacuole . This implies that an intact late endosomal HOPS function prevents key inducers of G-protein signaling from being rapidly targeted to the vacuole for degradation during appressorium initiation . However , Rgs1-mC was predominantly vacuolar during the late stages of pathogenic growth ( 24 hpi ) or in planta development ( 36 hpi ) ( Figure S11A and B in Text S1 ) , supporting our previous finding that cAMP signaling is down regulated during the late stages of pathogenic development , thus necessitating targeting of signaling proteins to the vacuole for either sequestration , recycling and/or degradation . The lack of proper appressorium initiation or an extensive delay in appressorium formation in the vps39Δ suggests that residual adenylate cyclase activity or G-protein activation likely occurs at other cellular locations , such as the plasma membrane . Such residual activity is likely insufficient to allow proper cAMP synthesis and appressorium formation , implying that the anchoring of G-proteins to the late endosomes allows sustained and robust activation of cAMP signaling in M . oryzae . In addition to its essential function in proper appressorium formation ( and cAMP signaling ) demonstrated here , we believe that endosomal fusion and maturation ( Vps39 function ) also contributes significantly to basic cellular functions such as proper colony growth , aerial hyphal development and conidiation ( Figure S10 in Text S1 ) . Our results add to a growing list of recent studies that implicate endosomes , and especially the late endosomes/MVB compartments as signaling scaffolds . In S . cerevisiae , the constitutive active G-alpha ( Gpa1Q323L ) interacts with PI3kinase on endosomes that also contain Snf7 [64] . In mammalian cells , the MVBs play a positive role in activating Wnt signaling , by sequestering the inhibitor ( GSK3 ) within the ILVs [44] . In another example , the free RIα subunit of cPKA is sequestered onto the surface of MVB's by AKAP11 in a cAMP dependent manner [42] . Finally , the integrity of the late endosomes is crucial for amino-acid and insulin stimulated mTOR signaling [43] . Our study revealed that dynamic tubulo-vesicular endosomes anchor active G-protein/cAMP signaling during appressorium initiation , whereas such signaling moieties are trafficked to the vacuole at late stages of infection-related morphogenesis ( Figure S11A in Text S1 ) . It was recently demonstrated that an Adenylate cyclase associated protein , Cap1 , localized as bright spots in the cytoplasm of the infection hyphae in M . oryzae [77] . However , we could not detect such punctate localization pattern for Rgs1-mC during infectious growth . Rgs1-mC was predominantly vacuolar in the infection hyphae ( Figure S11B in Text S1 ) . The detailed localization and function ( if any ) of G-protein/cAMP signaling in planta would form the basis of future research in blast disease . In conclusion , our findings clearly suggest that in addition to the canonical role ( s ) , the late endosomal compartments function as legitimate anchors ( Figure 7D ) that integrate trafficking and signal transduction for G-proteins and cAMP synthesis during pathogenesis in the devastating rice blast fungus .
The M . oryzae wild-type strain B157 was obtained from the Directorate of Rice Research ( Hyderabad , India ) . For growth and conidiation , wild type and mutant strains were grown on prune agar medium ( PA; per L: 40 mL prune juice , 5 g lactose , 5 g sucrose , 1 g yeast extract and 20 g agar , pH 6 . 5 ) as described [53] . Mycelial plugs were subcultured onto PA plates and incubated in a 28°C incubator in the dark for 7 days . For conidiation , cultures was incubated at 28°C in the dark for two days followed by light induction by exposing the plates to continuous fluorescent light at room temperature for 7 days . All the tagged strains used in this study were characterized for defects in vegetative growth , conidiation , appressorium formation , ability to respond to inductive/non-inductive surfaces , and pathogenicity ( Figure S12 and 13 in Text S1 ) . The wild-type strain was used as a control . The strains were also evaluated for protein expression and the size of the GFP- or mCherry fusion proteins was verified using western blot analysis ( Figure S12 and 13 in Text S1 ) . All the strains were found to be wild type like and the tagged proteins were found to be biologically functional . Experimental results were verified with a minimum of two strains of the same genotype . All strains used in this study were validated by Southern blotting , locus specific PCRs and requisite DNA sequence analyses . We have carefully compared and ascertained that constitutive promoter driven expression does not compromise the function , localization , and dynamics of the tagged Pth11 or Rgs1 compared to the natively expressed GFP-tagged versions . Only the intensity of the signal was enhanced thus allowing for better live-cell imaging . In the MagAG187S-GFP , MagA-GFP , Mac1-GFP , Rgs1-GFP and Pth11-GFP strains , the fusion protein in each instance is the sole copy of the given protein in that strain . For pathogenicity assays , leaves from two-week old barley seedlings were cut into small pieces ( 2–3 cm long ) and washed in sterile water , following which the leaf explants were rinsed for 45 seconds in 40% ethanol . The leaf pieces were then washed twice with sterile antibiotic containing distilled water . The washed leaves were placed on kinetin agar medium ( 2 mg/mL kinetin , 1% agar ) . Conidia were quantified and a dilution series of the conidial suspension was inoculated on detached barley leaves at the required concentrations . The samples were incubated in a humidified growth chamber with a 16 h light/8 h dark cycle at 22°C . Disease symptoms were assessed 5–7 days post inoculation . The RGS1 gene was tagged with mCherry at the C-terminus since the N-terminal DEP-DEP domain is known to specify membrane targeting [76] . The mCherry ORF was PCR amplified from pmCherry ( BD Biosciences Clontech , USA ) . The RGS1 gene ( MGG_14517 ) was amplified from genomic DNA of wild type M . oryzae . The TrpC terminator was amplified using pFGL275 as template . The PCR products were digested with appropriate restriction enzymes ( New England Biolabs , Beverly , MA ) and purified using the gel elution Nucleospin Extract II kit ( Machery-Nagel , Easton , PA ) . Using a three way ligation approach the mCherry ( SmaI/BamHI ) fragment along with the TrpC terminator as a BamHI/XbaI fragment were cloned into the SmaI/XbaI sites of pFGL44 vector [encoding hygromycin phosphotransferase gene ( HPH1 ) ] to obtain pFGL44-mCherry-TrpC terminator . The digested and eluted RGS1 ( EcoRI/SmaI fragment ) and Ccg1 ( EcoRI-EcoRI ) fragments were sequentially cloned into pFGL44-mCherry-TrpC , to yield pFGL44-RGS1-mCherry-TrpC and subsequently to give pFGL44-Ccg1-RGS1-mCherry-TrpC , representing the final construct . The orientation of the Ccg1 promoter fragment in the final constructs was confirmed using HindIII and KpnI restriction enzymes . The final clones were subjected to sequence analysis and then introduced into the wild type B157 strain via Agrobacterium T-DNA-mediated transformation . Fungal transformants were selected based on resistance towards hygromycin ( CM containing 250 µg/ml hygromycin , A . G . Scientific Inc , USA ) . Fungal transformants were screened for Rgs1-mCherry expression using an epifluorescence microscope and confirmed by sequencing of the requisite genomic region . The α-Tubulin gene was tagged at the N-terminus with eGFP driven by the RP27 promoter [55] , [78] . The eGFP was PCR amplified from pEGFP-N1 ( BD Biosciences Clontech , USA ) . The RP27 promoter and the α-tubulin gene ( MGG_06650 ) were amplified using genomic DNA from extracted from wild type M . oryzae as template . Plasmid pFGL275 was used as a template for the amplification of the TrpC terminator . The PCR products obtained were digested with suitable restriction enzymes ( New England Biolabs , Beverly , MA ) and purified using the gel elution Nucleospin Extract II kit ( Machery-Nagel , Easton , PA ) . The TrpC terminator as a SmaI/BamHI fragment was cloned into the SmaI/BamHI sites of pFGL97 vector [encoding ammonium glufosinate gene ( BAR ) ] to obtain pFGL97- TrpC terminator . The digested and eluted eGFP ( NcoI/SmaI fragment ) and RP27 ( EcoRI/NcoI ) fragments were cloned using a three way ligation approach into pFGL97-TrpC , to yield pFGL97-RP27-eGFP-TrpC . Finally the α-tubulin was cloned as a SmaI/SmaI fragment into pFGL97-RP27-eGFP-TrpC to subsequently yield pFGL97-RP27-α-tubulin-eGFP-TrpC , representing the final construct . The orientation of the α-tubulin fragment in the final construct was confirmed using BamHI restriction enzyme . The final clones were subjected to sequence analysis and were then introduced into the appropriate background ( Rgs1-mCherry ) strain via Agrobacterium T-DNA-mediated transformation . Fungal transformants were initially selected using ammonium glufosinate resistance ( BM containing 50 µg/ml ammonium glufosinate , Cluzeau Info Labo , France ) and confirmed by genomic DNA sequencing and GFP expression . The Rab5 ortholog in M . oryzae , Vps21 gene ( MGG_06241 ) was also tagged at the N-terminus with eGFP and driven by the RP27 promoter [55] , [78] . C-terminal tagging of Vps21/Rab5 is known to disrupt protein function . The eGFP was PCR amplified from pEGFP-N1 ( BD Biosciences Clontech , USA ) . The RP27 promoter and the Vps21 gene were amplified using genomic DNA from extracted from wild type as template . Plasmid pFGL275 was used as a template to amplify the TrpC terminator . The PCR products obtained were digested with suitable restriction enzymes ( New England Biolabs , Beverly , MA ) and purified using the gel elution Nucleospin Extract II kit ( Machery-Nagel , Easton , PA ) . First , the digested Vps21 gene , digested with SmaI/BamHI was cloned into pFGL97 vector [encoding ammonium glufosinate gene ( BAR ) ] at the SmaI/BamHI sites , followed by the TrpC terminator as a BamHI/BamHI fragment to obtain pFGL97-Vps21-TrpC terminator . The orientation of the TrpC terminator was confirmed using NcoI/PstI restriction enzymes . Next , the digested and eluted RP27 ( EcoRI/NcoI ) and eGFP ( NcoI/SmaI fragment ) fragments were cloned using a three way ligation approach into pFGL97-Vps21-TrpC , to yield pFGL97-RP27-eGFP-Vps21-TrpC , the final construct . The Rab7 ortholog Ypt7 ( MGG_08144 ) in M . oryzae was also tagged at the N-terminus , utilizing the pFGL97-RP27-eGFP-TrpC vector as the base for cloning . As mentioned above , the Ypt7 gene was PCR amplified using wild type genomic DNA as template . The Ypt7 amplicon was digested with SmaI enzyme and cloned into the pFGL97-RP27-eGFP-TrpC vector , at the SmaI site . This yielded the final construct pFGL97-RP27-eGFP-Ypt7-TrpC . The orientation of cloned Ypt7 fragment was checked and confirmed by digestion with HindIII . In both the cases , the final clones were subjected to sequence analysis and were then introduced into the appropriate background strains ( RGS1-mCherry ) via Agrobacterium T-DNA-mediated transformation . Requisite transformants were initially screened by epifluorescence microscopy for GFP expression , and candidate strains confirmed by locus-specific PCR and genomic DNA sequencing . The eGFP-2X FYVE fragment was released from pEGFP-2X FYVE ( a kind gift from Prof Harald Stenmark ) as an NcoI-SmaI fragment . This fragment ( NcoI-SmaI ) , along with the RP27 promoter ( EcoRI-NcoI ) was cloned into pFGL97 vector using a three-way ligation strategy to yield pFGL97-RP27-eGFP-2X FYVE . The TrpC terminator was cloned next into the same vector as SmaI/BamHI fragments to yield the final construct pFGL97-RP27-eGFP-2X FYVE-TrpC . Transformants were identified and confirmed as described above . Tagging at the C- or the N-terminus with fluorescent marker is known to compromise the function of the G-alpha subunit . In order to circumvent this problem , we introduced the eGFP coding sequence in frame into the alphaB-alphaC loop of M . oryzae MagA/Gαs ( between amino acids 113–120; as suggested by D . Siderovski and D . Bosch , USA based on an earlier study [56] . Based on an alternate strategy in D . discoideum [79] , the Gα subunits in N . crassa were recently tagged with TagRFP in the first alphaA-alphaB loop [80] . We used vector pFGL718 in which the fragment encoding ammonium glufosinate and eGFP is available as an HpaI-MluI fragment . Genomic DNA from the wild type was used as template to amplify the regions flanking either side of the alphaB-alphaC loop region in the MagA protein . The upstream flank was amplified as an EcoRI-MluI fragment and the downstream flank as an HpaI-HindIII fragment . Both the flanks were amplified such that they also incorporated short six amino acid linkers . In addition , another downstream flank incorporating the G187S point mutation was also amplified as an HpaI -HindIII fragment , using genomic DNA from the MagAG187S strain [54] . The sequence of the HpaI-HindIII fragment incorporating the G187S point mutation was sequenced and verified . The single 5′ upstream fragment and the two 3′ flanks were digested with appropriate restriction enzymes and cloned into pFGL718 , such that the upstream flank , the BAR-eGFP fragment and the downstream flank were all in frame . The constructs were confirmed and integrated into the Rgs1-mCherry strain such that the GFP-tagged MagA replaced the wild-type copy by homologous recombination . Site-specific replacement was confirmed using Southern analysis , locus-specific PCR and nucleotide sequencing . A BiFC strain expressing Pth11-YFP1–156 and Rgs1-YFP156–239 was generated to study at the interactions between the two proteins in-vivo . For this purpose , two constructs were generated , one in which PTH11 was cloned into pFGL44 and tagged C-terminally to YFP1–156 under the RP27 promoter; whereas the second construct involved cloning RGS1 into pFGL97 and tagged at the C-terminus to YFP156–239 under the Ccg1 promoter . The selected strains were confirmed by PCR analysis with appropriate controls . The VPS39 deletion mutant was generated using the standard one-step gene replacement strategy . Genomic DNA fragments ( about 1 kb each ) representing the 5′ and 3′ UTR of VPS39 ( MGG_01054 ) gene were amplified by PCR , ligated sequentially so as to flank the ammonium-glufosinate resistance gene cassette ( BAR ) under the TrpC promoter in pFGL97 to obtain plasmid vector pFGLvps39-KO . The sequence of pFGLvps39-KO gene replacement constructs was confirmed and introduced into Rgs1-mCherry , GFP-Rab7 and Mac1-GFP strains via Agrobacterium T-DNA-mediated transformation to specifically replace the VPS39 with BAR resistance cassettes . Resistance to ammonium glufosinate ( BM containing 40 ug/ml ammonium glufosinate ( Cluzeau Labo , France ) was used to select transformants . Total protein was extracted using non-denaturing NP40 buffer ( 10 mM Tris/Cl , pH 7 . 5 , 150 mM NaCl , 0 . 5 mM EDTA , 0 . 5% NP40 , 1 mM PMSF and 1× Protease Inhibitor Cocktail ) . Total protein was extracted from conidia inoculated on an inductive surface at the hooking stage ( 4 hpi ) or from vegetative growth from the requisite strains . Equal concentration of total protein lysate was incubated with GFP/RFP-Trap agarose beads ( ChromoTek , Germany ) and immuno-precipitation carried out as per manufacturer's instructions . Immuno-precipitated samples was fractionated by SDS–PAGE ( on 4–20% pre-cast gels , Bio-Rad , USA ) , transferred onto a PVDF membrane ( Millipore Corporation , USA ) and immuno blotted with α-Rgs1 ( DEP-DEP domain specific ) , α-GFP or α-mCherry antiserum ( 1∶1000 dilution ) . Secondary antibodies conjugated to horseradish peroxidase were used at 1∶10000 dilutions . The Super Signal kit ( Pierce , USA ) was used to detect the chemi-luminescent signal as instructed [54] , [76] . For appressorial assays and imaging using bright field optics , harvested conidia were resuspended in sterile water at a required concentration . Droplets ( 20–40 µl ) containing conidial suspension was placed on hydrophobic plastic coverslips or glass bottom dishes ( Mat Tek Corporation , USA ) and incubated under humid conditions at room temperature . The Rgs1-mC strain was inoculated on hydrophobic plastic coverslips and treated for 30 minutes with MBC ( final concentration 1 µM ) and LatA ( final concentration 10 µM ) at the hooking stage . PI3Kinase inhibitor LY294002 was added to Rgs1-mCherry conidia germinated on an inductive surface 4 hpi at a final concentration of 100 µM . The treatment was carried out for 180 minutes . In the case of GFP-FYVE strain , the LY294002 treatment was also carried out 4 hpi on an inductive surface , for 60–90 minutes . Rescue of the vps39Δ mutant was carried out by adding 8-Br-cAMP to a final concentration of 10 mM at 0 hpi . Wild type was treated with Bafilomycin-A1 2 hpi at a final concentration of 50 nM . A 0 . 1% DMSO solvent control was used in all assays . CMAC was used at a final concentration of 10 µM . Staining with CMAC was usually carried out 3–4 hpi , on an inductive surface , for 10–15 min at 37°C . The sample was washed and subsequently visualized . Lysotracker Green was used at a final concentration of 50 nM . Staining was carried out for 10 min at 37°C , 4 hpi , on an inductive surface . Droplets ( 20–40 µl ) containing conidial suspension were placed on hydrophobic plastic coverslips or glass bottom dishes ( Mat Tek Corporation , USA ) and incubated in a moist chamber at room temperature . Imaging was performed at the specified time points . Bright field images were captured at desired time points post inoculation using an Olympus BX51 microscope equipped with a PlanAPO 100×/1 . 45 or UPlan FLN 60×/1 . 25 objective with appropriate filter sets . Images were captured using a CoolSNAP HQ camera ( Photometrics , USA ) and processed using MetaVue ( Universal Imaging , USA ) . Time-lapse or live cell fluorescence microscopy was performed using a Zeiss Axiovert 200 M microscope ( Plan Apochromat 100× , 1 . 4NA objective ) equipped with an UltraView RS-3 spinning disk confocal system ( PerkinElmer Inc . , USA ) using a CSU21 confocal optical scanner , 12-bit digital cooled Hamamatsu Orca-ER camera ( OPELCO , Sterling , VA , USA ) and a 491 nm 100 mW and a 561 nm 50 mW laser illumination under the control of MetaMorph Premier Software , ( Universal Imaging , USA ) . Typically , z-stacks consisted of 0 . 5 µm-spaced planes for every time point . The maximum projection was obtained using the Metamorph built-in module . Alternatively , images were acquired using a Nikon TiE system ( CFI Plan Apochromat VC 100XH 1 . 4 N . A . objective ) equipped with a Yokogawa CSU-X1-A1 spinning disk unit , a Photometrics CoolSNAP HQ2 camera and a DPSS 491 nm 100 mW and DPSS561 nm 50 mW laser lines under the control of MetaMorph Premier Software , ( Universal Imaging , USA ) . Typically , single z-plane or image stacks that consisted of 0 . 5 µm-spaced sections were captured . GFP and LTG excitation were performed at 491 nm ( Em . 525/40 nm ) , FM4-64 excitation at 491 nm ( Em . 607/36 nm ) , mCherry excitation at 561 nm ( Em . 607/36 nm ) , YFP excitation at 515 nm ( Em . 542/27 nm ) and CMAC excitation at 405 nm ( Em . 452/45 nm ) . For quantitation of colocalization the raw . STK files were opened as stacks in Imaris , and the extent of colocalization between Rgs1-mCherry and the other GFP-tagged proteins or endosomal markers was measured using the Colocalization module of Imaris 7 . 5 , 64-bit version ( Bitplane AG , Saint Paul , MN , www . bitplane . com ) . Two-channel Z-stacks were assembled and the relative intensity of each wavelength determined for each voxel . The minimum intensity threshold for the analysis was determined for each channel automatically using the iterative algorithm defined by Costes et al . [81] . A time-dependent analysis of colocalization was run for each sample . Both the percentage of material colocalized ( as determined by the number of colocalized voxels relative to the total number of voxels with an intensity value of GFP signal above the minimum intensity threshold ) and the Pearson's colocalization coefficient was calculated from each sample for a minimum of four time points , a minimum of three different cells and at least two independent experiments performed on different days . The data were pooled and the results represent the average with error bars corresponding to the standard error of the mean . Image processing and figure preparation was performed using Bitplane Imaris for 3D surface rendering and colocalization analysis ( see above ) , Fiji ( http://fiji . sc/wiki/index . php/Fiji ) , and Adobe Photoshop and Microsoft Excel ( for figure preparation ) . Sequences for genes described in this article can be found under the following accession numbers: M . oryzae RGS1 ( MGG_14517 ) , MAG A ( MGG_01818 ) , PTH11 ( MGG_05871 ) , VPS21 ( MGG_06241 ) , YPT7 ( MGG_08144 ) , TUB ( MGG_06650 ) , VPS39 ( MGG_01054 ) and MAC1 ( MGG_09898 ) | Magnaporthe oryzae , a filamentous fungus , causes the devastating and economically important blast disease in rice and related species . It produces asexual spores , conidia , which can switch to a disease-causing mode of development in response to host and environmental cues . Activation of a highly conserved cell-signaling cascade , the G-proteins/Cyclic AMP pathway , has been shown to be important for pathogenic development in M . oryzae . In this study , we demonstrate that the late endosomal compartment is a highly dynamic tubulo-vesicular network that apart from membrane trafficking also anchors active G-protein signaling components in M . oryzae and therefore regulates the biogenesis and spatio-temporal dynamics of cAMP signaling . Furthermore , utilizing specific mutant strains compromised for late endosomal function we specifically demonstrate the importance of signal anchoring by late endosomes during the establishment and spread of blast disease in rice and barley . Thus , we propose that anchoring of G-protein signaling on late endosomes endows M . oryzae with the ability to specifically activate , integrate and achieve modularity of signaling and eventually control the robustness as well as the temporal duration of signaling responses critical for pathogenesis . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"biochemistry",
"plant",
"science",
"model",
"organisms",
"pest",
"control",
"genetics",
"biology",
"microbiology",
"molecular",
"cell",
"biology",
"agriculture"
] | 2013 | The Late Endosomal HOPS Complex Anchors Active G-Protein Signaling Essential for Pathogenesis in Magnaporthe oryzae |
The innate immune system is essential for controlling viral infections , but several viruses have evolved strategies to escape innate immunity . RIG-I is a cytoplasmic viral RNA sensor that triggers the signal to induce type I interferon production in response to viral infection . RIG-I activation is regulated by the K63-linked polyubiquitin chain mediated by Riplet and TRIM25 ubiquitin ligases . TRIM25 is required for RIG-I oligomerization and interaction with the IPS-1 adaptor molecule . A knockout study revealed that Riplet was essential for RIG-I activation . However the molecular mechanism underlying RIG-I activation by Riplet remains unclear , and the functional differences between Riplet and TRIM25 are also unknown . A genetic study and a pull-down assay indicated that Riplet was dispensable for RIG-I RNA binding activity but required for TRIM25 to activate RIG-I . Mutational analysis demonstrated that Lys-788 within the RIG-I repressor domain was critical for Riplet-mediated K63-linked polyubiquitination and that Riplet was required for the release of RIG-I autorepression of its N-terminal CARDs , which leads to the association of RIG-I with TRIM25 ubiquitin ligase and TBK1 protein kinase . Our data indicate that Riplet is a prerequisite for TRIM25 to activate RIG-I signaling . We investigated the biological importance of this mechanism in human cells and found that hepatitis C virus ( HCV ) abrogated this mechanism . Interestingly , HCV NS3-4A proteases targeted the Riplet protein and abrogated endogenous RIG-I polyubiquitination and association with TRIM25 and TBK1 , emphasizing the biological importance of this mechanism in human antiviral innate immunity . In conclusion , our results establish that Riplet-mediated K63-linked polyubiquitination released RIG-I RD autorepression , which allowed the access of positive factors to the RIG-I protein .
The innate immune system is essential for controlling virus infections , and several viruses have evolved strategies to evade host innate immune responses . Cytoplasmic viral RNA is recognized by RIG-I-like receptors , including RIG-I and MDA5 [1] , [2] . The RIG-I protein comprises N-terminal Caspase Activation and Recruitment Domains ( CARDs ) , a central RNA helicase domain , and a C-terminal Repressor domain ( RD ) [3] . RD consists of C-terminal RNA binding domain ( CTD ) and a bridging domain between CTD and helicase [4] . RIG-I CARDs are essential for triggering the signal that induces type I interferon ( IFN ) . In resting cells , RIG-I RD represses its CARDs signaling [3] . After viral infection , RIG-I RD recognizes 5′-triphosphate double-stranded RNA ( dsRNA ) , which results in a conformational change in the RIG-I protein [3] . This conformational change leads to the release of RD autorepression of CARDs , after which CARDs associate with an IPS-1 adaptor molecule ( also called MAVS , Cardif , and VISA ) localized at the outer membrane of mitochondria [3] , [5] , [6] , [7] , [8] . IPS-1 activates downstream factors such as TBK1 , IKK-ε , and NEMO [9] , [10] , [11] . NEMO forms a complex with TBK1 and IKK-ε and has a polyubiquitin binding region [12] . These protein kinases are essential for activating transcription factors such as IRF-3 to induce type I IFN production [13] . Several ubiquitin ligases are involved in regulating the RIG-I-dependent pathway , and RIG-I itself is regulated by ubiquitin chains [14] . Gack MU and colleagues firstly reported that TRIM25 ubiquitin ligase mediates K63-linked polyubiquitination of RIG-I N-terminal CARDs , which results in RIG-I activation [15] . Other groups also detected a RIG-I-anchored polyubiquitin chain after ligand stimulation or viral infection [16] , [17] . It was recently demonstrated that an unanchored polyubiquitin chain but not ubiquitination is essential for RIG-I activation [18] , [19] . However , RIG-I anchored K63-linked polyubiquitin chains are detected after viral infection [15] , [20] . Another E3 ubiquitin ligase , Riplet , binds RIG-I RD and mediates the K63-linked polyubiquitination of RIG-I RD [21] . In contrast , Chen DY and colleagues reported that Riplet ( also called Reul ) ubiquitinated RIG-I CARDs [22] . A study that used Riplet knockout mice showed that mouse Riplet is essential for RIG-I-mediated type I IFN production in response to vesicular stomatitis virus ( VSV ) , Flu , and Sendai virus ( SeV ) infections [23] . However , the functional difference between Riplet and TRIM25 remains unclear , and the molecular mechanism of how Riplet-mediated RIG-I ubiquitination activates RIG-I signaling remains unresolved . Hepatitis C virus ( HCV ) is a major cause of hepatocellular carcinoma ( HCC ) worldwide . HCV RNA is primarily recognized by RIG-I in vitro and in vivo [24] . The HCV protease NS3-4A can suppress type I IFN production [25] . NS3-4A cleaves IPS-1 to suppress RIG-I-mediated innate immune responses [7] , [26] . Human monocyte-derived dendritic cells recognize HCV RNA through Toll-like receptor 3 , and NS3-4A has the ability to cleave TICAM-1 , which is a solo adaptor molecule of Toll-like receptor 3 [27] , [28] . In this study , we found that Riplet was another target of NS3-4A . Here , we demonstrated the molecular mechanisms of how Riplet-mediated RIG-I polyubiquitination triggered the type I IFN production signal and showed that this mechanism was targeted by HCV .
In mouse embryonic fibroblasts ( MEFs ) , TRIM25 is essential for type I IFN production in response to SeV infection [15] . As with TRIM25 knockout , Riplet knockout abolished IFN-α2 and IFN-β mRNA expressions in response to SeV infection in MEF ( Figure 1A and 1B ) , which suggested that both ubiquitin ligases were essential for type I IFN expression in MEFs in response to SeV infection . Thus , we examined whether these two ubiquitin ligases played different roles in RIG-I activation . We performed reporter gene assays with p125luc ( IFN-β reporter ) using either RIG-I CARDs fragment or full-length RIG-I expression vectors . It is known that ectopic expression of RIG-I CARDs fragment activates the signaling even in the absence of stimulation with RIG-I ligand [2] , [3] and that the auto-activation is observed when full-length RIG-I is ectopically expressed in HEK293 cell [21] , [29] . As previously reported , TRIM25 ectopic expression efficiently increased RIG-I CARDs fragment-mediated signaling ( Figure 1C ) . However , TRIM25 expression only mildly increased full-length RIG-I signaling ( Figure 1D ) . In contrast , Riplet ectopic expression efficiently increased the full-length RIG-I-mediated signaling , although Riplet failed to increase the RIG-I CARDs-mediated signaling ( Figure 1C and 1D ) . Interestingly , when Riplet was co-expressed with TRIM25 , ectopically expressed TRIM25 could increase the full-length RIG-I-mediated signaling ( Figure 1F ) . We observed the same effects of TRIM25 and Riplet expressions on the reporter activation in the presence of stimulation with HCV 3′ UTR dsRNA ( a RIG-I ligand ) or VSV infection , which are recognized by RIG-I ( Figure 1E and 1F ) . These different effects of Riplet and TRIM25 on the signaling suggested that Riplet and TRIM25 ubiquitin ligases played different roles in RIG-I activation . We then investigated whether the two ubiquitin ligases associate with RIG-I after stimulation . First , we performed immunoprecipitation assay and found that endogenous Riplet and TRIM25 bound to endogenous RIG-I after stimulation with HCV dsRNA ( Figure 1G ) . Second , we investigated subcellular localizations of the proteins by confocal immunofluorescence microscopy . In resting cells , RIG-I was barely detectable , whereas RIG-I exhibited punctate staining in cytoplasm after VSV infection ( Figure 2A ) , and ectopically expressed Riplet was co-localized with RIG-I and TRIM25 in perinuclear region in VSV infected cells ( Figure 2A–2C and 2E ) . This is consistent with a previous observation that RIG-I and TRIM25 co-localized extensively at cytoplasmic perinuclear bodies [15] . Pearson's correlation coefficient values also suggested the correlation of the colocalizations of those proteins ( Figure 2D and 2F ) . Recently , it was reported that RIG-I recognized short polyI:C in stress granules [30] . G3BP is a marker of the stress granule [30] . In resting cells , G3BP dispersed in cytoplasm [30] , and thus barely detectable ( Figure 2G and 2H ) , whereas G3BP speckles were detected in cells stimulated with short polyI:C in the cytoplasm ( Figure 2G and 2H ) . Riplet and TRIM25 localizations within G3BP speckles were detected in the stimulated cells ( Figure 2G and 2H ) . Taken together , these data indicated that both Riplet and TRIM25 associated with RIG-I after stimulation . Next , we assessed the RIG-I regions that bind to the two ubiquitin ligases using RIG-I fragments . As previously reported , TRIM25 bound to RIG-I CARDs fragment ( Figure 2I ) . However , Riplet bound to RIG-I RD ( 735–925 aa ) , but not to CARDs fragment ( Figure 2J ) . Deleting the Riplet binding region ( RIG-I-ΔCTD ) abrogated Riplet effect on RIG-I signaling ( Figure 2K ) . This was contrast to TRIM25 , which affects RIG-I CARDs . Taken together , our genetic and biochemical data indicated that the two ubiquitin ligases associated with RIG-I after stimulation but showed different effects on RIG-I activation . Thus , we next focused on Riplet specific role in RIG-I activation . RIG-I CARDs harbor K63-linked polyubiquitination [15] . As RIG-I CARDs , RIG-I RD harbored K63-linked polyubiquitination ( Supplemental Figure S1 ) . Riplet expression increased the polyubiquitination of RIG-I RD but not that of CARDs ( Figure 3A ) . RIG-I RD has two functions . One is RNA binding activity and the other is autorepression of its CARDs signaling . Firstly , we tested whether Riplet affects RIG-I RNA binding activity . In a pull-down assay using biotin-conjugated dsRNA and streptavidin beads , we found that both polyubiquitinated and non-ubiquitinated RIG-I were recovered ( Figure 3B ) , which suggested that Riplet-mediated polyubiquitination was dispensable for RIG-I RNA binding activity . As RIG-I is known to form homo-oligomers [3] , it is possible that non-ubiquitinated RIG-I was recovered through ubiquitinated RIG-I by the pull-down assay . To exclude this possibility , we used RIG-I mutants , RIG-I 5KR and RIG-I K788R , which were barely ubiquitinated by Riplet ( described below ) . We compared the binding abilities of the RIG-I mutants to that of wild-type RIG-I by pull-down assay using cell lysate isolated from cells without stimulation with RIG-I ligand to avoid any ubiquitination . The results showed that the RIG-I 5KR and RIG-I K788R mutant proteins were recovered by pull down assay as wild-type RIG-I ( Figure 3C ) . This data also indicated that the polyubiquitination is dispensable for RIG-I RNA binding activity . Secondly , we investigated whether Riplet expression affects RIG-I RD autorepression of CARDs signaling . To test this possibility , we constructed RIG-I mutant proteins ( Figure 4A ) . We previously showed that Lys to Ala amino acid substitutions at Lys-849 , 851 , 888 , 907 , and 909 of RIG-I ( RIG-I 5KA ) severely reduced RIG-I polyubiquitination and activation [21] . However , it is possible that the substitutions of Lys with Ala affect other function of RIG-I because the substitution abolishes positive charge of Lys residues . Thus , we constructed the RIG-I 5KR mutant protein , in which the five Lys residues were substituted with Arg , and examined the signal activation ability . The results showed that the 5KR mutation reduced RIG-I signaling , however residual activation of RIG-I 5KR was still detected ( Figure 4B–4D ) . Thus , we assessed other Lys residues within RIG-I RD . Because Riplet is essential for RIG-I activation , it is expected that the Lys residues targeted by Riplet are conserved during evolution . Thus , we tested Lys residues within RIG-I RD conserved among vertebrate , and found that an RIG-I K788R mutation reduced RIG-I signaling at a level comparable to that by a K172R mutation , which abrogates TRIM25-mediated RIG-I activation [15] , [19] ( Figure 4B–4D ) . Interestingly , the 5KR and K788R mutations reduced the RIG-I polyubiquitination ( Figure 4E ) . RIG-I polyubiquitination in cells stimulated with HCV dsRNA was also reduced by the K788R mutation ( Figure 4F ) . Lys-788 is located within the 55-amino acid region ( 747–801 aa ) of RIG-I essential for RD autorepression of CARDs signaling [31] ( Figure 4A ) . It is possible that the reduction of RIG-I ubiquitination by the mutations was caused by the defect of RIG-I other functions , such as RNA binding activity . However , Figure 3C showed that the RIG-I 5KR and K788R mutant proteins efficiently bound to dsRNA as wild-type RIG-I . This data weakened the above possibility . Using the RIG-I mutants , we investigated whether Riplet releases RIG-I RD autorepression . Because Riplet ectopic expression activated full-length RIG-I signaling even in the absence of stimulation ( Figure 1D ) , we could assess whether Riplet can remove RIG-I RD autorepression by Riplet ectopic expression study . Due to autorepression , full-length RIG-I expression weakly activates the signaling compared with RIG-I CARDs expression [31] . Riplet ectopic expression increased full-length RIG-I signaling to a level comparable to that of RIG-I CARDs signaling ( Figure 5A ) , suggesting that Riplet released the autorepression . Moreover , the K788R substitution canceled this Riplet ectopic expression effect on RIG-I signaling ( Figure 5A ) . This suggested that the K788R mutation abrogated the release of RIG-I RD autorepression but not RNA binding activity . Expression of the RIG-I RD fragment is known to represses full-length RIG-I-mediated signaling [3] . Interestingly , Riplet ectopic expression removed the RD fragment repression effect on RIG-I signaling ( Figure 5B ) . This Riplet expression effect was canceled by the K788R amino acid substitution within full-length RIG-I ( Figure 5B ) . MDA5 C-terminal region does not function as an RD , and thus ectopically expressed MDA5 induced IFN-β promoter activity irrespective of viral infection [3] . If Riplet ectopic expression released RIG-I RD autorepression , it is expected that RIG-I and Riplet co-expression will induce IFN-β promoter activity irrespective of viral infection . As expected , Riplet and RIG-I co-expression induced IFN-β promoter activity irrespective of SeV infection ( Figure 5C ) . Taken together , these genetic data indicated that Riplet released RIG-I RD autorepression . If Riplet is essential for the release of RIG-I RD autorepression , it is expected that Riplet expression will increase the interaction between RIG-I and TRIM25 , because TRIM25 efficiently activated RIG-I CARDs but not full-length RIG-I ( Figure 1C and 1D ) . To test this possibility , we examined the interaction between TRIM25 and RIG-I in the presence or absence of Riplet ectopic expression and found that Riplet expression increased the interaction between TRIM25 and RIG-I ( Figure 5D and 5E ) . Moreover , this interaction was abolished by the K788R mutation ( Figure 5F ) . These data were also consistent with our model that Riplet affects RIG-I RD autorepression rather than RNA binding activity of RIG-I . If Riplet is essential for the release of RIG-I RD autorepression leading to the interaction between TRIM25 and RIG-I , it was expected that Riplet is essential for endogenous RIG-I K63-linked polyubiquitination that is mediated by both Riplet and TRIM25 . To test this possibility , we investigated endogenous RIG-I K63-linked polyubiquitination in mouse spleen cells infected with SeV . Endogenous RIG-I K63-linked polyubiquitination was increased after SeV infection in wild-type splenocyte , however knockout of Riplet abrogated the endogenous K63-linked polyubiquitination of RIG-I after SeV infection ( Figure 5G ) . Recently , it was reported that knockdown of Riplet strongly reduced endogenous RIG-I polyubiquitination in response to SeV infection in a mouse cell line Hepa 1 . 6 [32] . Based on our genetic and biochemical data in Figure 3–5 , we concluded that Riplet affects RIG-I RD autorepression rather than the RNA binding activity . Because Riplet-mediated release of RD autorepression increased the interaction between RIG-I and TRIM25 , we investigated whether the release of RD autorepression also increased the interaction of RIG-I with other factors . Interestingly , we found that ectopically expressed IKK-ε , TBK1 , and NEMO ubiquitin binding region co-immunoprecipitated with RIG-I RD ( Figure 6A–6D ) , and Riplet expression enhanced the physical interactions of RIG-I with TBK1 , IKK-ε , and the NEMO ubiquitin binding region ( Figure 6D–6G ) . The physical interactions between these proteins were not through IPS-1 , as IPS-1 did not co-immunoprecipitate with RIG-I RD ( Figure 6B ) . Microscopy analysis showed that the RIG-I was co-localized with TBK1 or NEMO in the cytoplasm . ( Figures 6H and S2A ) . RIG-I and TBK1 was detected in the region where there is no mitochondria ( Figure 6I ) , and the colocalization of RIG-I with TBK1 was also detected in the region where there is no mitochondria ( yellow stained region in Figure 6J ) . This was consistent with our immunoprecipitation results that RIG-I RD could bind TBK1 without IPS-1 . TBK1 is phosphorylated in its activation loop [33] . Surprisingly , the phosphorylated TBK1 ( p-TBK1 ) foci were exclusively localized on mitochondria ( Figure 6K and 6L ) . Co-localization of RIG-I with p-TBK1 was observed after dsRNA stimulation ( Figure S2B and S2C ) . We next assessed the role of endogenous Riplet in the interaction between RIG-I and TBK1 . Immunoprecipitation assay showed that Riplet KO reduced the interaction between endogenous RIG-I and TBK1 in mouse spleen cells during VSV infection , indicating that Riplet promoted the interaction between RIG-I and TBK1 ( Figure 6M ) . These data indicated that the Riplet function increased the interaction of RIG-I with TBK1 as well as TRIM25 . Several viruses have evolved strategies to escape the innate immunity . For instance , NS1 of influenza A virus inhibits TRIM25 function . This emphasizes the vital role of TRIM25 in modulating antiviral response [34] . To assess the biological significance of Riplet-mediated release of RIG-I RD autorepression in antiviral innate immune response , we investigated whether viral protein suppresses this mechanism . The endogenous Riplet protein level was not affected by polyI:C , HCV dsRNA stimulations , or by VSV infection ( Figure 7A–7C ) , however the Riplet protein level was severely reduced in a human hepatocyte cell line with HCV 1b full-length replicons ( O cells ) compared with hepatocyte cell line without these replicons ( O curred cells: Oc cells; Figure 7D ) , suggesting that viral protein reduced the Riplet protein level . Riplet knockout abolished the expression of type I IFN , IP-10 , and type III IFN in response to HCV RNA ( Figure 7E ) , indicating that Riplet was essential for type I IFNs expression in response to HCV RNA . Because HCV protease NS3-4A suppresses type I IFN expression in response to viral infection [7] , [25] , we examined whether NS3-4A could cleave the Riplet protein . N-terminal FLAG-tagged Riplet or C-terminal HA-tagged Riplet was expressed with or without NS3-4A , after which the Riplet protein levels were compared . NS3-4A expression severely reduced the FLAG-tagged and HA-tagged Riplet protein level but not that of FLAG-tagged RIG-I , whereas catalytically inactive NS3-4A* ( S139A ) failed to reduce the Riplet protein level ( Figure 7F , 7G , and S3A ) . This suggested that this protease's activity reduced the Riplet protein level . Although NS3-4A reduced IPS-1 protein level as previously reported , NS3-4A did not reduce the TRIM25 and IKK-ε protein levels ( Figure 7H , 7I , S3B–D ) . Within the Riplet RING-finger domain is a sequence that is similar to the NS3-4A target consensus sequence ( D/E-x-x-x-x-C/T-S/A; Figure 7J ) [35] . NS3-4A cleaves the target just after C/T within this consensus sequence . Acidic amino acids before the C/T site are conserved among the NS3-4A cleavage site within HCV polypeptide . The acidic amino acids from 16 to 18 aa within the Riplet sequence were substituted with Ala , and a Riplet-3A mutant protein was constructed ( Figure 7J ) . The Riplet-3A mutant protein was resistant to NS3-4A ( Figure 7I and 7K ) . Moreover , the Riplet-3A protein co-localized with NS3-4A in cytoplasm ( Figure 7L ) . Interestingly , recombinant NS3-4A that was purified from E . coli cleaved the immunoprecipitated FLAG-tagged Riplet protein and recombinant GST-fused Riplet protein purified from E . coli , and the cleaved fragments were detected at expected size ( Figure 7M and 7N ) . These data indicated that NS3-4A directly targeted the Riplet protein . Although Riplet digestion products were not observed in HEK293 cell lysate ( Figure S3E and S3F ) , it is known that the digestion products of TICAM-1 ( TRIF ) obtained by NS3-4A are not detectable because these products are unstable [28] . Cys-21 of Riplet corresponds to the C/T site in the NS3-4A target consensus sequence . The Cys-21 residue is the first Cys in the RING-finger motif; thus a C21A substitution causes the disruption of RING finger domain structure [36] . The Riplet-C21A mutant protein was unstable and barely detectable ( Figure S3G ) , which suggested that the loss of Cys-21 destabilized the Riplet protein . To determine if NS3-4A abolishes RIG-I activation by disrupting Riplet , we examined RIG-I ubiquitination and interaction between RIG-I and TRIM25 . Riplet-mediated ubiquitination of full-length RIG-I or RIG-I RD was abolished by NS3-4A expression ( Figure 8A and 8B ) . K63-linked polyubiquitination of RIG-I RD in SeV infected cells were also reduced by NS3-4A expression ( Figure 8C ) . Moreover , NS3-4A expression reduced the interaction between TRIM25 and full-length RIG-I ( Figure 8D ) . An IPS-1-C508A mutant protein is resistant to NS3-4A cleavage [7] . A reporter gene assay showed that NS3-4A expression severely reduced IFN-β promoter activation induced by RIG-I and Riplet expression even in the presence of the IPS-1 C508A mutant protein in HEK293 cell ( Figure 9A ) , and a catalytically inactive NS3-4A mutant failed to reduce this signal ( Figure 9B ) . These data were consistent with our observation that NS3-4A targeted the Riplet protein . There was endogenous IPS-1 in HEK293 cell , and thus the decrease of IFN-β promoter induction by NS3-4A in HEK293 cell could be due to the cleavage of endogenous IPS-1 by NS3-4A . To exclude this possibility , we next used IPS-1 KO mouse hepatocyte [37] . The IPS-1 C508A mutant protein was expressed in IPS-1 KO hepatocyte . NS3-4A expression reduced the reporter activation induced by RIG-I/Riplet expression or by stimulation with HCV dsRNA even in the presence of IPS-1 C508A ( Figure 9C and 9D ) . These results indicated that NS3-4A targeted both IPS-1 and upstream factors of IPS-1 and were consistent with our observation that NS3-4A reduced the Riplet protein level . We could not test whether NS3-4A failed to impair IFN-β promoter activity in the presence of the Riplet-3A and Riplet C21A mutants because the mutant proteins were not functional and failed to activate RIG-I ( Supplemental Figure S3H and S3I ) . Thus , we could not exclude the possibility that NS3-4A targeted another protein in addition to Riplet and IPS-1 . Endogenous RIG-I exhibited punctate staining in the human hepatocyte cell line HuH7 and HepG2 cells after simulation with HCV RNA ( Figure 9E ) but not in HuH7 . 5 cell line ( Figure 9E ) . In HuH7 . 5 cells , there is a T55I mutation within endogenous RIG-I gene that disrupts the interaction between RIG-I and TRIM25 [38] . We investigated whether Riplet is required for RIG-I to exhibit punctate staining , and found that knockdown of Riplet decreased RIG-I punctate staining induced by HCV RNA ( Supplemental Figure S4 ) . We next investigated whether HCV abrogated Riplet-dependent RIG-I punctate pattern in the cytoplasm . As expected , RIG-I failed to exhibit punctate staining in O cells with HCV replicons in NS3 positive cells and HuH7 cells infected with HCV JFH1 ( Figure 9F–9H ) . We confirmed that HCV disrupted IPS-1 in our experimental condition ( Figure 9G and 9H ) . To further assess whether HCV abrogates endogenous Riplet function , we observed endogenous K63-linked polyubiquitination of RIG-I in cells with HCV replicons . Although SeV infection induced endogenous K63-linked polyubiquitination of RIG-I in HuH7 cells , HCV replicons failed to increase the polyubiquitination ( Figure 9I ) . Next , we investigated the association of endogenous RIG-I with TRIM25 and TBK1 , which is promoted by Riplet as shown in Figure 5 and 6 . SeV infection induced the association of endogenous RIG-I with TRIM25 and TBK1 , whereas HCV replicons failed to induce the association ( Figure 9J ) . Taken together , these data indicated that HCV abrogated endogenous Riplet function . Although NS3-4A cleaves IPS-1 , a mutation within endogenous RIG-I gene increased the permissiveness to HCV infection in HuH7-derived cells [3] , indicating that RIG-I is required for antiviral response to HCV infection before NS3-4A cleaves IPS-1 . We used siRNA to knockdown endogenous Riplet in HuH7 cells , and then the cells were infected with HCV JFH1 . Interestingly , Riplet knockdown increased the permissiveness to HCV JFH1 infection ( Figure 9K ) , indicating that endogenous Riplet is required for antiviral response to HCV infection .
RIG-I activation is regulated by two ubiquitin ligases Riplet and TRIM25 [15] , [21] . The two ubiquitin ligases are essential for RIG-I activation [15] , [23] , however the functional difference had been unclear . It is known that TRIM25 is essential for RIG-I oligomerization and association with IPS-1 adaptor molecule [15] , [19] . Here , we demonstrated that Riplet was essential for the release of RIG-I RD autorepression of its CARDs , which resulted in the association with TRIM25 . This functional difference explained the reason why RIG-I requires the two ubiquitin ligases for triggering the signal . It has been reported that TRIM25 activates RIG-I signaling [15] , [19] . We confirmed that ectopic expression of TRIM25 increases RIG-I CARDs-mediated signaling . However , most previous studies used a RIG-I CARDs fragment but not full-length RIG-I [15] , [19] , [38] . Unexpectedly , we found that the increase of full-length RIG-I-mediated signaling by TRIM25 expression was much less than that of the CARDs-mediated signaling . It is intriguing that Riplet helped TRIM25 to activate full-length RIG-I . Riplet expression promoted the interaction between TRIM25 and full-length RIG-I , and this interaction was abrogated by an RIG-I K788R mutation , which reduced Riplet-mediated RIG-I ubiquitination . Thus , we propose that Riplet-mediated polyubiquitination of RIG-I RD is a prerequisite for TRIM25 to activate RIG-I ( Figure 10 ) . Ectopic expression of Riplet activated RIG-I without stimulation with RIG-I ligand . This is not surprising because ectopically expressed Riplet bound to RIG-I without stimulation with RIG-I ligand , whereas endogenous Riplet bound to endogenous RIG-I after stimulation with RIG-I ligand . RIG-I undergoes its conformational change after binding to a ligand [3] , [39] . The conformational change would allow the access of endogenous Riplet to RIG-I , which resulted in Riplet-mediated K63-linked polyubiquitination leading to the release of RD autorepression . This model is consistent with the observation that TRIM25 ectopic expression did not activate full-length RIG-I without Riplet expression , because TRIM25 hardly bound to full-length RIG-I without Riplet . Previously , we reported that the five Lys residues within RIG-I RD were important for Riplet-mediated RIG-I ubiquitination . We constructed the RIG-I 5KR mutant and indicated that the 5KR mutation reduced RIG-I ubiquitination and activation without loss of RNA binding activity . This is consistent with our previous conclusion . However , there is residual ubiquitination of RIG-I 5KR mutation , and we found that K788R mutation showed more sever phenotype . These data indicated that Riplet targeted the several Lys residues within RIG-I RD . This is not surprising , because TRIM25 targets not only Lys-172 but also other Lys residues within mouse RIG-I CARDs [32] . TBK1 and IKK-ε are downstream factors of IPS-1 . We found that TBK1 and IKK-ε could bind RIG-I RD . It is possible that RIG-I associates with TBK1 through IPS-1 . However , Hiscott J and colleagues demonstrated that IKK-ε could bind IPS-1 and that TBK1 did not bind IPS-1 [40] . Moreover , RIG-I RD did not bind IPS-1 , and RIG-I and TBK1 co-localization was detected in the cytoplasmic region where there are no mitochondria . These observations weaken this possibility . Our results indicated that RIG-I RD bound to the NEMO ubiquitin binding region . IRF-3 activation requires the ubiquitin binding domain of NEMO , and an endogenous K63-linked polyubiquitin chain plays a key role in IRF3 activation [41] . Thus , we prefer a model in which TBK1 associates with an RIG-I RD-anchored polyubiquitin chain through NEMO ( Figure 10 ) . Although Riplet knockout reduced the binding of RIG-I to TBK1 , residual binding was still detectable . Thus , there appears to be Riplet-dependent and independent associations between RIG-I and TBK1 . TRAF3 is an E3 ubiquitin ligase , and is involved in the RIG-I-mediated type I IFN production pathway [42] . Because there is residual activation of the type I IFN production pathway even in TRAF3 knockout cells [41] , it is possible that the RIG-I polyubiquitin chain may compensate for the TRAF3 defect in recruiting TBK1 to mitochondria . Further studies will be needed to determine the precise molecular mechanisms . Although TBK1 dispersed in the cytoplasm , p-TBK1 was exclusively localized on mitochondria . Considering that TBK1 is phosphorylated in its activation loop [33] , these results suggested that RIG-I RD associated with inactive TBK1 and that TBK1 was activated after loading on to mitochondria ( Figure 10 ) . HCV is a major cause of HCC and has the ability to evade host innate immune response [7] , [43] . HCV RNA is primarily recognized by the cytoplasmic viral RNA sensor RIG-I . Previous studies showed that the protease NS3-4A cleaves IPS-1 to shut off RIG-I signaling . However , our results indicated that there was another target of NS3-4A in RIG-I signaling . First , RIG-I failed to exhibit punctate staining in cells infected with HCV . Second , NS3-4A reduced RIG-I signaling even in the presence of an IPS-1-C508A mutant , which is resistant to the cleavage by NS3-4A . Third , the endogenous Riplet protein level was severely reduced in cells with HCV replicons . Fourth , NS3-4A targeted Riplet and abrogated Riplet-dependent RIG-I ubiquitination and complex formation with TRIM25 and TBK1 . These data support our model that NS3-4A targets not only IPS-1 but also Riplet to escape host innate immune responses ( Figure 10 ) . Recently it was reported that NS1 proteins of Influenza A virus inhibited Riplet function [32] . These findings indicated biological importance of Riplet in RIG-I activation during viral infection . In general , a ubiquitin ligase has several targets . We have performed yeast two-hybrid screening using Riplet as bait and found a candidate clone that encodes a tumor suppressor gene . Our pilot study showed that Riplet mediated K63-linked polyubiquitination of this tumor suppressor and suppressed retinoblastoma ( Rb ) activity . Thus , Riplet disruption by NS3-4A might be a cause of liver disease induced by HCV infection .
All animal studies were carried out in strict accordance with Guidelines for Animal Experimentation of the Japanese Associations for Laboratory Animal Science . The protocols were approved by the Animal Care and Use Committee of Hokkaido University , Japan ( Permit Number: 08-0245 and 09-0215 ) . HEK293 , Vero , and HepG2 cells were cultured in Dulbecco's modified Eagle's medium low glucose medium ( D-MEM ) with 10% heat-inactivated fetal calf serum ( FCS ) ( Invitrogen ) . HeLa cells were cultured in minimum Eagle's medium with 2 mM L-glutamine and 10% heat-inactivated FCS . HEK293FT cells were maintained in D-MEM high glucose medium containing 10% of heat-inactivated FCS ( Invitrogen ) . Human hepatocyte cell line with HCV 1b full-length replicons ( O cells ) and O curred cells ( Oc cells ) were kindly gifted from Kato N [44] . O cells were cultured in D-MEM high glucose with 10% of heat-inactivated FCS , G418 , NEAA , and L-Gln . VSV Indiana strain and SeV HVJ strain were amplified using Vero cells . To determine the virus titer , we performed plaque assay using Vero cells . HCV JFH1 was amplified using HuH7 . 5 cells . Generation of IPS-1 KO and Riplet KO mice were described previously [23] , [45] . Splenocyte was isolated from C57BL/6 wild-type and Riplet KO mice . Isolated cells were cultured in RPMI1640 containing 10% of heat-inactivated FCS . The preparations of wild-type and Riplet KO MEFs were described previously [23] . Preparation of IPS-1 KO mouse hepatocyte was described previously [37] . All mice were maintained under specific-pathogen free conditions in the animal facility of the Hokkaido University Graduate School of Science ( Japan ) . Expression vectors encoding for N-terminal FLAG-tagged RIG-I , N-terminal FLAG-tagged RIG-I CARDs ( dRIG-I ) , FLAG-tagged RIG-I ΔRD ( RIG-I-dRD ) , FLAG-tagged RIG-I RD , C-terminal HA-tagged TRIM25 , C-terminal HA-tagged Riplet , and Riplet-ΔRING ( Riplet-DN ) plasmids were described previously [21] , [23] . The amino acids substitutions from 16 to 18 with Ala was carried out by PCR-mediated mutagenesis using primers , Ripelt-3A-F and Riplet-3A-R and pEF-BOS/Riplet plasmid as a template . The primer sequence is Riplet-3A-F: TTC CCG TGT GGC TGG CCG CGG CCG CCC TCG GCT GCA TCA TCT GCC , and Riplet-3A-R: GGC AGA TGA TGC AGC CGA GGG CGG CCG CGG CCA GCC ACA CGG GAA . RIG-I K172R and RIG-I K788R expression vectors was constructed by PCR-mediated mutagenesis using primers , RIG-I K172R-F , RIG-I K172R-R , RIG-I K788R-F and RIG-I-K788R-R , and pEF-BOS/FLAG-RIG-I plasmid as a template . The primer sequences are RIG-I K172R-F: GGA AAA CTG GCC CAA AAC TTT GAG ACT TGC TTT GGA GAA AG , RIG-I K172R-R: CTT TCT CCA AAG CAA GTC TCA AAG TTT TGG GCC AGT TTT CC , RIG-I-K788R-F: TGC ATA TAC AGA CTC ATG AAA GAT TCA TCA GAG ATA GTC AAG AA , and RIG-I-K788-R: CTT GAC TAT CTC TGA TGA ATC TTT CAT GAG TCT GTA TAT GCA G . RIG-I 5KR expression vectors were constructed by PCR-mediated mutagenesis using primers , RIG-I 849 851 RR-F , RIG-I 849 851 RR-R , RIG-I 888R-F , RIG-I 888R-R , RIG-I 907 909 RR-F , RIG-I 907 909 RR-R , and pEF-BOS/FLAG-RIG-I plasmid as a template . The primer sequences are RIG-I 849 851 RR-F: AGT AGA CCA CAT CCC AGG CCA AGG CAG TTT TCA AGT TTT G , RIG-I 849 851 RR-R: CAA AAC TTG AAA ACT GCC TTG GCC TGG GAT GTG GTC TAC T , RIG-I 888R-F: GAC ATT TGA GAT TCC AGT TAT AAG AAT TGA AAG TTT TGT GGT GGA GG , RIG-I 888R: CCT CCA CCA CAA AAC TTT CAA TTC TTA TAA CTG GAA TCT CAA ATG TC , RIG-I 907 909RR-F: GTT CAG ACA CTG TAC TCG AGG TGG AGG GAC TTT CAT TTT GAG AAG , RIG-I 907 909RR-R: CTT CTC AAA ATG AAA GTC CCT CCA CCT CGA GTA CAG TGT CTG AAC . HCV cDNA fragment encoding NS3-4A of JFH1 strain was cloned into pCDNA3 . 1 ( - ) vector . The mutation on catalytic site of NS3-4A S139A was constructed by PCR-mediated mutagenesis using primers , NS3-4A S139A-F and NS3-4A S139A-R , and pCDNA3 . 1 ( - ) /NS3-4A plasmid as a template . The primer sequences are NS3-4A S139A-F: TTC GAC CTT GAA GGG GTC CGC GGG GGG ACC GGT GCT TTG C and NS3-4A S139A-R: AAG CAC CGG TCC CCC CGC GGA CCC CTT CAA GGT CGA AAG G . Total RNA was extracted with TRIZOL ( Invitrogen ) , after which the samples were treated with DNaseI to remove DNA contamination . Reverse transcription was performed using High Capacity cDNA Reverse Transcription Kit ( ABI ) . Quantitative PCR analysis was performed using Step One software ve2 . 0 . ( ABI ) with SYBER Green Master Mix ( ABI ) . HCV ss and dsRNA was in vitro synthesized with SP6 and/or T7 RNA polymerase using 3′ UTR of HCV cDNA as template as described previously [46] . Cells were plated onto microscope cover glasses ( matsunami ) in a 24-well plate . The cells were fixed for 30 min using 3% formaldehyde in PBS and permeabilized with 0 . 2% Triton X-100 for 15 min . Fixed cells were blocked with 1% bovine serum albumin in PBS for 10 min and labeled with the indicated primary Abs for 60 min at room temperature . Alexa-conjugated secondary Abs were incubated for 30 min at room temperature to visualize staining of the primary Ab staining . Samples were mounted on glass slides using Prolong Gold ( Invitrogen ) . Cells were visualized at a magnification of ×63 with an LSM510 META microscope ( Zeiss ) . Data collected with confocal microscopy were analyzed with ZEISS LSM Image Examiner software . NS3 , RIG-I , TBK1 , IPS-1 , and p-TBK1 were stained with anti-NS3 goat pAb ( abcam ) , anti-RIG-I mouse mAb ( Alme-1 , ALEXIS BIOCHEMICALS ) , anti-NAK ( TBK1 ) rabbit mAb ( EP611Y , abcam ) , anti-MAVS ( IPS-1 ) rabbit pAb ( Bethyl Laboratories Inc ) , and anti-p-TBK1 rabbit mAb ( Cell Signaling Technology ) , HEK293 cells were transiently transfected in 24-well plates using FuGene HD ( Promega ) or lipofectamine 2000 ( Invitrogen ) with expression vectors , reporter plasmids ( IFN-β: p125luc ) , and an internal control plasmid coding Renilla luciferase . The total amounts of plasmids were normalized using an empty vector . Cells were lysed in a lysis buffer ( Promega ) , and luciferase and Renilla luciferase activities were determined using a dual luciferase assay kit ( Promega ) . Relative luciferase activities were calculated by normalizing the luciferase activity by control . HCV dsRNA ( 3′ UTR polyU/UC region ) was synthesized using T7 and SP6 RNA polymerase as described previously [46] . RNA used for the assay was purchased from JBioS . The RNA sequences are as follows: ( sense strand ) AAA CUG AAA GGG AGA AGU GAA AGU G; and ( antisense strand ) CAC UUU CAC UUC UCC CUU UCA GUU U . Biotin was conjugated at the U residue at the 3′-end of the antisense strand ( underlined ) . Biotinylated dsRNA was phosphorylated by T4 polynucleotide kinase ( TAKARA ) . dsRNA was incubated for one hour at 25°C with 10 µg of protein from the cytoplasmic fraction of cells that were transfected with Flag-tagged RIG-I , Riplet , and/or HA-tagged ubiquitin expressing vectors . This mixture was added into 400 µl of lysis buffer ( 20 mM Tris-HCl pH 7 . 5 , 150 mM NaCl , 1 mM EDTA , 10% Glycerol , 1% NP-40 , 30 mM NaF , 5 mM Na3VO4 , 20 mM iodoacetamide , and 2 mM PMSF ) containing 25 µl of streptavidine Sepharose beads , rocked at 4°C for two hours , harvested by centrifugation , washed three times with lysis buffer , and resuspended in SDS sample buffer . Splenocytes ( 1×107 ) were infected with or without VSV at MOI = 10 for eight hours , after which cell extracts were prepared with lysis buffer ( 20 mM Tris-HCl pH 7 . 5 , 150 mM NaCl , 1 mM EDTA , 10% glycerol , 1% Nonidet P-40 , 30 mM NaF , 5 mM Na3VO4 , 20 mM iodoacetamide , and 2 mM phenylmethylsulfonyl fluoride ) . Immunoprecipitation used an anti-RIG-I Rabbit monoclonal antibody ( D14G6 , Cell Signaling Technology ) . To detect endogenous K63-linked polyubiquitin chain that is ligated to RIG-I , 6×107 of mouse splenocyte were infected with SeV at MOI = 0 . 2 for 24 hours . Immunoprecipitation was performed with anti-RIG-I mAb ( D14G6 ) . Anti-K63-linkage specific polyubiquitin ( D7A11 ) Rabbit mAb ( Cell Signaling ) was used for western blotting . HEK293FT cells were transfected with or without 0 . 8 µg of HCV dsRNA in a 6-well plate . HCV dsRNA ( HCV 3′ UTR polyU/UC region ) was synthesized using T7 and SP6 RNA polymerase as previously described [46] . Cell lysates were prepared at the indicated times . Immunoprecipitation was performed with an anti-RIG-I mouse monoclonal antibody ( Alme-1 ) . An anti-FLAG M2 monoclonal antibody ( Sigma ) was used for the immunoprecipitation of FLAG-tagged protein . An anti-TRIM25 rabbit polyclonal antibody ( abcam ) , an anti-p-TBK1 rabbit mAb ( Cell Signaling Technology ) , an anti-NAK ( TBK1 ) rabbit mAb ( EP611Y ) , and an anti-RNF135 ( Riplet ) pAb ( SIGMA ) , were used for western blotting . For ubiquitination assay , immunoprecipitates were washed three times with high salt lysis buffer ( ( 20 mM Tris-HCl pH 7 . 5 , 1M NaCl , 1 mM EDTA , 10% glycerol , 1% Nonidet P-40 , 30 mM NaF , 5 mM Na3VO4 , 20 mM iodoacetamide , and 2 mM phenylmethylsulfonyl fluoride ) to dissociate unanchored polyubiquitin chain [21] , and then washed once with normal lysis buffer described above for SDS-PAG analysis . Band intensity was semi-quantified using Photoshop software . siRNAs for human Riplet ( Silencer Select Validated siRNA ) and negative control were purchased from Ambion . siRNA sequences for Riplet are: ( sense ) GGA ACA UCU UGU AGA CAU Utt and ( anti-sense ) AAU GUC UAC AAG AUG UUC CCac . siRNA was transfected into cells using RNAiMax Reagent ( Invitrogen ) according to the manufacture's instructions . FLAG-tagged Riplet was expressed in HEK293FT cells , and cell lysate was prepared with the lysis buffer described above . The protein was immunoprecipitated with anti-FLAG antibody and protein G sepharose beads , and washed with Buffer B ( 20 mM Tris-HCl pH 7 . 5 , 150 mM NaCl , 10% glycerol , 1% Nonidet P-40 ) . The samples were suspended in 50 µl of Buffer B , and incubated with 400 ng of recombinant NS3-4A ( rNS3-4A ) protein at 37°C for one hour , and then subjected to SDS-PAGE analysis . The NS3-4A protein was purchased from AnaSpec Inc ( CA ) . N-terminal GST-fused Riplet ( 1–210 aa ) ( rRiplet ) was purchased from Abnova . 500 ng of rRiplet was incubated with or without 500 ng of rNS3-4A in 10 µl of reaction buffer ( 20 mM Tris-HCl ( 7 . 5 ) , 4% Glycerol , 5 mM DTT , 150 mM NaCl , 0 . 1% of Triton-X100 , 0 . 9% polyvinyl alcohol ) at 37°C for 30 min . The accession numbers are Riplet ( BAG84604 ) , TRIM25 ( NP_005073 ) , TBK1 ( NP_037386 ) , IKK-ε ( AAF45307 ) , IPS-1 ( BAE79738 ) , RIG-I ( NP_055129 ) , and G3BP ( CAG38772 ) . | The cytoplasmic viral RNA sensor RIG-I recognizes various types of pathogenic viruses and evokes innate immune responses , whereas several viruses have evolved strategies to escape the host innate immune responses . RIG-I triggers a signal to induce type I interferon and inflammatory cytokines . RIG-I activation is regulated by K63-linked polyubiquitin chains mediated by the ubiquitin ligases TRIM25 and Riplet; however , the functional difference between the two ubiquitin ligases remains unclear , and the molecular mechanism underlying Riplet-mediated RIG-I activation is unknown . We revealed sequential roles of the two ubiquitin ligases during RIG-I activation and found that Riplet-mediated polyubiquitination of the RIG-I repressor domain released RIG-I autorepression of its N-terminal CARDs responsible for triggering the signal , which resulted in an association with TRIM25 ubiquitin ligase and TBK1 protein kinase . Interestingly , we found that this mechanism was targeted by hepatitis C virus , which is a major cause of hepatocellular carcinoma . This result emphasizes the vital role of Riplet-mediated release of RIG-I RD autorepression in antiviral responses . Our results establish that Riplet releases RIG-I RD autorepression and demonstrated the biological significance of this mechanism in human innate immune responses . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"signal",
"transduction",
"immunity",
"mechanisms",
"of",
"signal",
"transduction",
"innate",
"immunity",
"immunology",
"biology",
"molecular",
"cell",
"biology"
] | 2013 | A Distinct Role of Riplet-Mediated K63-Linked Polyubiquitination of the RIG-I Repressor Domain in Human Antiviral Innate Immune Responses |
Expression of co-inhibitory molecules is generally associated with T-cell dysfunction in chronic viral infections such as HIV or HCV . However , their relative contribution in the T-cell impairment remains unclear . In the present study , we have evaluated the impact of the expression of co-inhibitory molecules such as 2B4 , PD-1 and CD160 on the functions of CD8 T-cells specific to influenza , EBV and CMV . We show that CD8 T-cell populations expressing CD160 , but not PD-1 , had reduced proliferation capacity and perforin expression , thus indicating that the functional impairment in CD160+ CD8 T cells may be independent of PD-1 expression . The blockade of CD160/CD160-ligand interaction restored CD8 T-cell proliferation capacity , and the extent of restoration directly correlated with the ex vivo proportion of CD160+ CD8 T cells suggesting that CD160 negatively regulates TCR-mediated signaling . Furthermore , CD160 expression was not up-regulated upon T-cell activation or proliferation as compared to PD-1 . Taken together , these results provide evidence that CD160-associated CD8 T-cell functional impairment is independent of PD-1 expression .
Co-stimulatory and co-inhibitory molecules play a major role in the regulation of antigen-specific T-cell responses [1] . Following T-cell receptor ( TCR ) engagement , activation or inhibition of T-cell responses depends upon the balance between stimulatory and inhibitory signals , on the type of molecules engaged or ligands involved and the availability of signaling molecules [2]–[4] . Co-stimulatory/co-inhibitory molecules are commonly divided into 4 families: 1 ) the B7 family including CD28 , Cytotoxic T-lymphocyte associated protein-4 ( CTLA-4 ) , Programmed Death receptor-1 ( PD-1 ) , Inducible T-cell Costimulator ( ICOS ) and B- and T-lymphocyte attenuator ( BTLA ) , 2 ) TNF-α receptor family including CD27 , 3 ) the CD2/SLAM family , including Signaling Lymphocyte Activation Molecule ( SLAM ) , 2B4 and CD48 and 4 ) the immunoglobulin ( Ig ) family including T-cell Immunoglobulin mucin-3 ( TIM-3 ) , lymphocyte Activation Gene-3 ( LAG-3 ) and CD160 [5]–[10] . Each co-inhibitory/stimulatory molecule interacts with one or several receptors expressed by one or various cell types ( reviewed in [2] ) . During the past decade , many studies performed in mice and humans have underscored the role of co-inhibitory molecules in the functional impairment ( also called “exhaustion” ) of antigen-specific T cells during chronic viral infections such as human immunodeficiency virus-1 ( HIV-1 ) or hepatitis C virus ( HCV ) [11]–[14] . In these virus chronic infections , the early functional impairment of T cells was marked by the loss of proliferation capacity likely resulting from reduced capacity to produce IL-2 and a deficient killing capacity of CD8 T cells . The ability to produce TNF-α was generally observed at an intermediate state of T-cell exhaustion while the loss of IFN-γ occurred in the advanced stage of T-cell exhaustion [15] , [16] . Recent studies have demonstrated that HIV-specific CD8 T cells co-expressing several co-inhibitory molecules such as PD-1 , CD160 and 2B4 were significantly more functionally impaired than CD8 T cells expressing only one co-inhibitory molecule [17]–[19] . However , the relative contribution of each co-inhibitory molecule has not yet been fully delineated . In the present study , we evaluated the impact of the expression of co-inhibitory molecules such as 2B4 , PD-1 and CD160 on CD8 T-cells specific to influenza ( Flu ) , Epstein Barr virus ( EBV ) and cytomegalovirus ( CMV ) . We demonstrated that CD160+ CD8 T cells had reduced proliferation capacity , IL-2 production and perforin expression regardless of PD-1 expression thus providing evidence that CD160-associated T-cell impairment is independent of PD-1 .
The expression of PD-1 , 2B4 , and CD160 co-inhibitory molecules was assessed by multiparametric flow cytometry in CMV- , EBV- and Flu-specific CD8 T cells from 22 healthy individuals ex vivo using peptide-MHC class I multimer complexes ( Table 1 ) . The gating strategy defining the positivity of co-inhibitory molecule expression was set by fluorescence minus one ( FMO ) for all co-inhibitory molecules tested ( Fig . S1 ) . The use of 2B4 , PD-1 and CD160 co-inhibitory molecules expression allowed the identification of 6 major CD8 T-cell populations i . e . 2B4+CD160+PD-1+ , 2B4+CD160+PD-1− , 2B4+CD160−PD-1+ , 2B4+CD160−PD-1− , 2B4−CD160−PD-1+ and 2B4−CD160−PD-1− CD8 T-cell populations . The 2B4−CD160+PD-1+ and 2B4−CD160+PD-1− CD8 T-cell populations are minor populations and represent less than 0 . 05% and 0 . 5% of virus-specific CD8 T cells , respectively , and were not included in the additional analyses shown below . Representative flow cytometry profiles and cumulative data ( n = 30 ) showed that CMV and EBV-specific CD8 T cells expressed significantly higher levels of 2B4 ( P = 0 . 0003 and P = 0 . 0015 , respectively ) and CD160 ( P<0 . 0001 and P = 0 . 0022 , respectively ) as compared to Flu-specific CD8 T cells ( Fig . 1A and B ) . Interestingly , no significant differences were observed in PD-1 expression between CMV , EBV and Flu-specific CD8 T cells in the cohort of subjects investigated in the present study ( Fig . 1A and B ) . However , the proportion of 2B4+CD160+PD-1+ and 2B4+CD160+PD-1− cells were significantly higher in CMV and EBV-specific CD8 T cells than in Flu-specific CD8 T cells ( P<0 . 025 ) ( Fig . 1C ) . Of note , mean fluorescent intensity ( MFI ) of CD2B4 , CD160 and PD-1 was not significantly different between FLU , EBV and CMV-specific CD8 T cells ( Fig . S2 ) . Flu-specific CD8 T cells were mainly composed of 2B4−CD160−PD-1− , 2B4−CD160−PD-1+ and 2B4+CD160−PD-1+ CD8 T-cell populations ( Fig . 1C ) . Interestingly , the co-inhibitory molecule expression profile of CMV and EBV-specific CD8 T cells was heterogeneous but not significantly different ( P>0 . 05 ) . As mentioned above , CD8 T-cell proliferation capacity is one of the first function lost during CD8 T-cell functional impairment [11]–[14] , [20] . The proliferation capacity of Flu , EBV and CMV-specific CD8 T cells was then assessed using the CFSE flow cytometry assay . To normalize according to the initial frequency of virus-specific CD8 T cells , virus-specific CD8 T-cell proliferation capacity was represented in proliferation index , defined as the ratio of the virus-specific CD8 T-cell proliferation capacity ( percentage of CFSE low CD8 T cells ) and the ex vivo frequency of virus-specific CD8 T cells ( measured by MHC class I multimer staining ) . The representative flow cytometry profiles and the cumulative data showed that the proliferation index of Flu-specific CD8 T cells was significantly higher than that of EBV and CMV-specific CD8 T cells ( Fig . 2 A–E ) . We then evaluated the potential association between the ex vivo expression pattern of co-inhibitory molecules expression and the proliferation index of virus-specific CD8 T cells . We showed that the ex vivo proportion of CD160 , 2B4 but not PD-1 inversely correlated with the proliferation index of virus-specific CD8 T cells ( r = −0 . 7605; P<0 . 0001 , r = −0 . 5248; P = 0 . 0049 , and r = 0 . 1002; P = 0 . 6192 , respectively ) ( Fig . S3 ) . We then assessed whether certain combinations of co-inhibitory molecules expression would be specifically associated with the proliferation index of virus-specific CD8 T cells . We showed that the ex vivo proportion of 2B4+CD160+PD-1+ or 2B4+CD160+PD-1− virus-specific CD8 T cells inversely correlated with the proliferation index of virus-specific CD8 T cells ( r = −0 . 72; P<0 . 0001 and r = −0 . 62; P = 0 . 0004 , respectively ) ( Fig . 2F ) . In addition , the proliferation index of virus-specific CD8 T cells did not inversely correlate with the proportion of CD160− CD8 T-cell populations ( Fig . 2F ) . These data suggest that the proliferation capacity of virus-specific CD8 T cells may be negatively regulated by CD160 expression independently of PD-1 expression . To determine the impact of co-inhibitory signals on virus-specific CD8 T-cell proliferation capacity , CD8 T-cell populations were sorted based on PD-1 , 2B4 and CD160 expression . Of note , naïve CD8 T cells ( CD45RA+CCR7+ ) were excluded from the analysis . Cell populations were then stimulated with ( viral ) peptides in the presence of antigen presenting cells ( CD8-depleted PBMCs ) for 6 days ( Fig . 3 ) . The proliferation capacity of Flu , EBV and CMV-specific CD8 T cells was then assessed using the CFSE flow cytometry assay and normalized according to the initial frequency of virus-specific CD8 T cells within each CD8 T-cell population ( i . e . 2B4+CD160+PD-1+ , 2B4+CD160+PD-1− , 2B4+CD160−PD-1+ , 2B4+CD160−PD-1− , 2B4−CD160−PD-1+ and 2B4−CD160−PD-1− CD8 T-cell populations ) . The representative flow cytometric profiles as well as the cumulative data show that the proliferation capacity of virus-specific 2B4+CD160+PD-1+ , 2B4+CD160+PD-1− , 2B4+CD160−PD-1+ , 2B4+CD160−PD-1− CD8 T-cell populations was significantly reduced as compared to 2B4−CD160−PD-1+ and 2B4−CD160−PD-1− CD8 T-cell populations ( P<0 . 05 ) ( Fig . 3A–B ) . Of note , the proliferation capacity of 2B4+CD160−PD-1− and 2B4−CD160+PD-1+ CD8 T cells was not significantly different ( P = 0 . 0628 ) ( Fig . 3B ) . These data suggest that the proliferation capacity of virus-specific CD8 T cells is influenced by co-inhibitory molecule expression . To determine the intrinsic CD8 T-cell proliferation capacity , CD8 T-cell populations were sorted based on PD-1 , 2B4 and CD160 expression . Of note , naïve CD8 T cells ( CD45RA+CCR7+ ) were excluded from the analysis . Cell populations were then stimulated with coated anti-CD3 and anti-CD28 MAbs for 6 days , and the proliferation capacity was assessed using CFSE flow cytometry based assay ( Fig . 4 ) . The representative flow cytometric profiles and the cumulative data showed that the proliferation capacity of CD160 and/or PD-1-expressing CD8 T cells was significantly reduced as compared to CD160 and PD-1 negative CD8 T-cell populations ( P<0 . 05 ) ( Fig . 4 ) . Interestingly , the proliferation capacity of 2B4+CD160+PD-1+ and 2B4+CD160+PD-1− CD8 T cells was not significantly different ( P = 0 . 6786 ) , suggesting that CD160-expessing CD8 T cells are endowed with reduced proliferation capacity , independently of PD-1 expression . In chronic virus infections the loss of CD8 T-cell proliferative capacity is commonly associated with a progressive reduction of IL-2 production [15] , [16] . Production of IFN-γ is affected generally at advanced stages of T-cell exhaustion [15] , [16] . IL-2 and IFN-γ production were then evaluated by flow cytometry in memory CD8 T-cell populations defined by PD-1 , 2B4 and CD160 expression . Of note , the naïve CD8 T-cell population ( CD45RA+CCR7+ ) was excluded from this analysis . The representative flow cytometric profiles as well as the cumulative data showed that the frequencies of IL-2 and IFN-γ-producing cells were significantly reduced in CD160+ CD8 T-cell populations as compared to CD160− CD8 T-cell populations ( P<0 . 05 ) ( Fig . 5A–C ) . The frequencies of both IL-2 and IFN-γ-producing cells were not significantly different between 2B4+CD160+PD-1+ and 2B4+CD160+PD-1− CD8 T-cell populations ( P = 0 . 998 and P = 0 . 470 , respectively ) and between 2B4+CD160−PD-1+ and 2B4+CD160−PD-1− CD8 T-cell populations ( P = 0 . 888 and P = 0 . 908 , respectively ) ( Fig . 5A–C ) . In addition , 2B4−CD160−PD-1+ CD8 T-cell populations contained higher frequencies of IL-2-producing cells than any other CD8 T-cell population ( P<0 . 05 ) ( Fig . 5A–C ) , demonstrating that CD160 and not PD-1 expression on memory CD8 T cells is strongly associated with reduced IL-2 and IFN-γ production capacity . Reduced IL-2 production capacity is commonly associated with increased differentiation state [21] . In this context , we assessed the potential association between co-inhibitory molecule expression and differentiation state . To that purpose , the differentiation state was evaluated using the expression of CD45RA and CCR7 . The representative flow cytometric profiles as well as the cumulative data showed that 2B4 , CD160 and PD-1 expression significantly increased with the differentiation state , confirming the recent findings from Legat and colleagues [22] ( Fig . 6A–B ) . However , in-depth analyses showed that CD8 T-cell populations defined by 2B4 , CD160 and/or PD-1 expression were heterogeneously distributed among the distinct differentiated CD8 T-cell subsets ( Fig . 6C ) . Indeed , the central memory compartment ( CM; CD45RA−CCR7+ ) was significantly enriched in 2B4−CD160−PD-1− and 2B4−CD160−PD-1+ CD8 T-cell populations , the effector memory compartment ( EM; CD45RA−CCR7− ) was significantly enriched in PD-1+ CD8 T-cell populations ( 2B4+CD160+PD-1+ , 2B4+CD160−PD-1+ and 2B4−CD160−PD-1+ CD8 T-cell populations ) and the terminally differentiated effector memory compartment ( TDEM; CD45RA+CCR7− ) was significantly enriched in the 2B4+CD160+PD-1− CD8 T-cell population ( P<0 , 05 ) ( Fig . 6C ) . These data suggest that co-inhibitory molecule expression is associated with differentiation state . However in depth analyses further indicated that significant differences are present within each CD8 T-cell populations , particularly among PD-1 and/or CD160-expressing populations . Since perforin expression is the hallmark of differentiated CD8 T cells [21] , we then assessed whether 2B4+CD160+PD-1+ and 2B4+CD160+PD-1− CD8 T-cell populations were enriched in perforin or granzyme B . Indeed , cytotoxic CD8 T cells exert their antiviral activity primarily through the secretion of cytotoxic granules containing perforin and granzymes [23] , [24] . Therefore , the expression of perforin and granzyme B was evaluated in CD8 T-cell populations defined by the expression of PD-1 , 2B4 and/or CD160 . Of note , the naïve CD8 T-cell population ( CD45RA+CCR7+ ) was excluded from this analysis . As shown in the representative flow cytometric profiles and the cumulative data perforin and granzyme B were significantly enriched in 2B4+ CD8 T-cell population as compared to 2B4− CD8 T-cell populations ( P<0 . 05 ) ( Fig . 7A–B ) consistently with a previous study [25] . Within 2B4+ CD8 T-cell populations , perforin expression was not significantly different between 2B4+PD-1−CD160+ and 2B4+PD-1+CD160− CD8 T-cell populations ( P = 0 . 1596 ) . However , perforin expression was significantly reduced in 2B4+PD-1+CD160+ , 2B4+PD-1−CD160+ and 2B4+PD-1+CD160− CD8 T-cell populations as compared to the 2B4+PD-1−CD160− CD8 T-cell population ( P = 0 . 0014 , P = 0 . 0047 and P = 0 . 0039 , respectively ) . Interestingly , perforin expression was significantly reduced in 2B4+PD-1+CD160+ as compared to 2B4+PD-1−CD160+ and 2B4+PD-1+CD160− CD8 T-cell populations ( P = 0 . 0092 and P = 0 . 0201 , respectively ) . Finally , the potential association between perforin and co-inhibitory molecule expression was assessed within each differentiated CD8 T-cell subset i . e . naïve , central memory , effector memory and terminally differentiated effector memory . The representative flow cytometric profiles and the cumulative data showed that perforin expression was significantly enriched within 2B4-expressing CD8 T cells in both EM ( P = 0 . 0011 ) and TDEM ( P<0 . 0001 ) ( Fig . 7C–D ) . Interestingly , perforin expression was significantly reduced in CD160-expressing CD8 T cells in both EM ( P = 0 . 0063 ) and TDEM ( P = 0 . 026 ) and in PD-1-expressing CD8 T cells in both EM ( P = 0 . 0391 ) and TDEM ( P = 0 . 00031 ) ( Fig . 7C–D ) . Taken together , these data demonstrate that CD8 T-cell populations expressing CD160 and/or PD-1 harbored reduced perforin expression , independently of the differentiation state ( Fig . 7C–D ) . Of note , the percentage of expression of EOMES , T-bet and CD57 ( immunosenescence [26] ) was not significantly different ( P>0 . 05 ) between 2B4+CD160+PD-1− and 2B4+CD160+PD-1+ CD8 T-cell populations ( Fig . S4 ) , suggesting that 2B4+CD160+PD-1+ CD8 T-cell population was not more immunosenescent than 2B4+CD160+PD-1− CD8 T-cell population . Taken together , these data demonstrate that the expression of CD160 and/or PD-1 on 2B4+ CD8 T cells is associated with a reduced capacity to express perforin as compared to the 2B4+CD160−PD-1− CD8 T-cell population , independently on the differentiation state . In order to assess the relative influence of CD160 signaling as compared to PD-1 signaling on the virus-specific CD8 T-cell proliferation capacity , the proliferation capacity of CMV , EBV and Flu-specific CD8 T cells was evaluated in the presence or in the absence of 1 ) blocking anti-CD160 monoclonal antibodies ( mAbs ) , 2 ) blocking anti-PD-L1/2 mAbs , 3 ) combined blocking anti-CD160 and anti-PD-L1/2 mAbs and 4 ) isotype controls . Of note , anti-CD160 MAbs did not show any agonistic potential either in absence or in presence of TCR signal ( data not shown ) , as assessed by CD69 , CD107a and BCL-2 expression or TNF-α and IFN-γ production ( data not shown ) . The representative flow cytometric profiles showed that both anti-CD160 and anti-PD-1 mAb treatments individually increased the proliferation capacity of EBV-specific CD8 T cells ( Fig . 8A ) . In addition , the cumulative data showed that both anti-CD160 and anti-PD-1 mAb treatments individually increased the proliferation capacity of CMV , EBV and Flu-specific CD8 T cells ( 2 . 06 fold increase; P<0 . 0001; and 7 . 6 fold increase; P<0 . 0001 , respectively ) ( Fig . 8B–C ) . The effects of the combined blockade of CD160 and PD-1 signaling pathways as compared to PD-1 or CD160 signaling blockade alone on CD8 T-cell proliferation were assessed on virus-specific CD8 T-cell populations harboring 1 ) both 2B4+CD160−PD-1+ and 2B4+CD160+PD-1− CD8 T-cell populations ( 2B4+CD160−PD-1+ and 2B4+CD160+PD-1− virus-specific CD8 T cells >10%; N = 5 ) ( Fig . 8D ) and 2 ) containing the highest proportion of 2B4+CD160+PD-1+ CD8 T-cell population ( 2B4+CD160+PD-1+ virus-specific CD8 T cells >30%; N = 6 ) ( Fig . 8E ) . The cumulative data show that the combined blockade of CD160 and PD-1 signaling pathways significantly increased CD8 T-cell proliferation as compared to PD-1 or CD160 signaling blockade alone , in virus-specific CD8 T cells containing both 2B4+CD160−PD-1+ and 2B4+CD160+PD-1− CD8 T-cell populations ( 2B4+CD160−PD-1+ and 2B4+CD160+PD-1− virus-specific CD8 T cells >10%; N = 5 ) ( P<0 . 05 ) ( Fig . 8D ) . However , consistently with the study by Peretz et al . [18] , the combined blockade of CD160 and PD-1 signaling pathways did not significantly increase CD8 T-cell proliferation as compared to PD-1 signaling blockade alone , in virus-specific CD8 T cells containing the highest proportion of 2B4+CD160+PD-1+ CD8 T-cell population ( 2B4+CD160+PD-1+ virus-specific CD8 T cells >30%; N = 6 ) ( P>0 . 05 ) ( Fig . 8E ) . In addition , PD-1 signaling blockade was significantly more potent on virus-specific CD8 T cells with dominant 2B4−CD160−PD-1+ CD8 T-cell population as compared to virus-specific CD8 T cells with dominant 2B4+CD160−PD-1+ or 2B4+CD160+PD-1+ CD8 T-cell populations ( 10 . 36 versus 4 . 06 and 1 . 95 fold increase; P = 0 . 0014 and P = 0 . 0023 , respectively ) ( Fig . 8F ) . Of note , PD-1 mean fluorescent intensity ( MFI ) was not significantly different between 2B4−CD160−PD-1+ and 2B4+CD160+PD-1+ CD8 T-cell populations ( Fig . S5 ) . Interestingly , however , the 2B4+CD160+PD-1− CD8 T-cell population expressed the highest levels of CD160 ( MFI ) and 2B4 ( MFI ) as compared to any other CD8 T-cell population ( Fig . S5 ) . Taken together , these data demonstrate that a ) CD160 and PD-1 signaling negatively regulate TCR-mediated CD8 T-cell signaling , b ) the functional restoration induced by the blockade of multiple co-inhibitory molecules may be incomplete , and c ) the cells expressing several co-inhibitory molecules are more profoundly impaired than cells expressing only one co-inhibitory molecule . Since the expression level of co-inhibitory molecules on virus-specific CD8 T cells is highly heterogeneous ( Fig . 1 ) , the influence of CD160 and PD-1 signaling blockade was evaluated with regard to the ex vivo expression levels of CD160 and PD-1 . The representative flow cytometric profiles as well as the cumulative data showed that CD160 signaling blockade was significantly more potent in cells harboring high level of CD160 ( percentage of virus-specific CD8 expressing more than 20% of CD160 ) versus cells harboring low level of CD160 ( percentage of virus-specific CD8 expressing less than 20% of CD160 ) ( P<0 . 0001 ) ( Fig . 9A–E ) . In addition , the degree of restoration of virus-specific CD8 T-cell proliferation capacity induced by CD160 blockade directly correlated with the ex vivo proportion of 2B4+CD160+PD-1+ or 2B4+CD160+PD-1− virus-specific CD8 T cells ( r = 0 . 6818; P = 0 . 0003 and r = 0 . 7141; P<0 . 0001 , respectively ) ( Fig . 9F–G ) . Interestingly , the effect of the blockade of PD-1/PD-L1/2 pathway on the degree of the restoration of virus-specific CD8 T-cell proliferation capacity was not associated with the ex vivo proportion of 2B4+CD160+PD-1+ , 2B4+CD160−PD-1+ or 2B4−CD160−PD-1+ virus-specific CD8 T cells ( r = −0 . 1611; P>0 . 05 , and r = −0 . 2566; P>0 . 05 and r = 0 . 2949; P>0 . 05 , respectively ) ( Fig . 9H–J ) . Previous studies have indicated that PD-1 expression increased following T-cell stimulation and activation [27] . However , the regulation of CD160 expression following T-cell activation or proliferation remains unclear . We then evaluated the expression levels of PD-1 and CD160 co-inhibitory molecules following TCR stimulation using flow cytometry . Briefly , cells were stained with CFSE , stimulated with anti-CD3 plus anti-CD28 Abs and the expression levels of PD-1 and CD160 were assessed by flow cytometry at day 0 , 1 , 2 , 3 and 5 . Representative flow cytometric profiles as well as cumulative data showed that CD160 expression was down-regulated in CD8 T cells upon T-cell activation and proliferation , while PD-1 expression increased ( Fig . 10A–B ) .
Several studies have demonstrated that the increased expression of co-inhibitory molecules including PD-1 , 2B4 and CD160 during chronic viral infections is associated with CD8 T-cell dysfunction [2] , [17] , [18] , [28] . However , the relative contribution of individual co-inhibitory molecules in the T-cell functional impairment remains unclear . In the present study , we have evaluated the impact of individual co-inhibitory molecules expression such as 2B4 , PD-1 and CD160 on virus-specific CD8 T-cell functions . To address this issue , we have performed both phenotypic and functional characterization of Flu , EBV and CMV-specific CD8 T cells in healthy individuals . We show that EBV and CMV-specific CD8 T cells contained higher frequency of cells co-expressing 2B4 , CD160 and PD-1 in different combinations while Flu-specific CD8 T cells had higher frequencies of triple negative , single PD-1 or dual PD-1/2B4 CD8 T-cell populations . In particular , the percentage of CD160 expression was significantly lower in Flu-specific CD8 T cells than in CMV or EBV-specific CD8 T cells . Interestingly , the level of total PD-1 was not significantly different in Flu , EBV or CMV-specific CD8 T cells . These data suggest that CD160 expression in virus-specific CD8 T cells is associated with chronic virus infections such as EBV and CMV . Of note , co-inhibitory molecule expression on virus-specific CD8 T cells are likely influenced by the pathogen characteristics , the type of viral infection ( cleared versus chronic viral infections ) , the cytokine milieu , the nature of the infected cell population . Proliferation capacity is one of the first CD8 T-cell functions lost during the progression to CD8 T-cell exhaustion [11]–[14] , [20] . In this regard , we assessed which pattern of co-inhibitory molecules expression had the greater impact on the proliferation capacity of virus-specific CD8 T cells . We showed that the proportion of total CD160 expression and in particular of 2B4+CD160+PD-1+ or 2B4+CD160+PD-1− CD8 T-cell populations inversely correlated with the proliferation capacity thus suggesting that CD160 expression was associated with reduced CD8 T-cell proliferation capacity independently of PD-1 expression . The reduced proliferation capacity of sorted virus-specific 2B4+CD160+PD-1+ , 2B4+CD160+PD-1− , 2B4+CD160−PD-1+ , 2B4+CD160−PD-1− CD8 T-cell populations as compared to 2B4−CD160−PD-1+ and 2B4−CD160−PD-1− CD8 T-cell populations suggest that the proliferation capacity CD8 T cells is influenced by co-inhibitory molecule expression . The intrinsic proliferation capacity of the different CD8 T-cell populations was then evaluated and showed that CD8 T cells expressing CD160 and/or PD-1 were endowed with reduced proliferation capacity as compared to CD160 and PD-1 negative CD8 T-cell populations . Interestingly , the proliferation capacity of 2B4+CD160+PD-1+ and 2B4+CD160+PD-1− CD8 T cells was not significantly different , suggesting that CD160 expressing CD8 T cells are endowed with reduced proliferation capacity , independently of PD-1 expression . We then showed that 2B4+CD160+PD-1+ or 2B4+CD160+PD-1− CD8 T-cell populations also had a reduced capacity to produce IL-2 , IFN-γ and perforin as compared to CD160− CD8 T-cell populations demonstrating that CD160+ CD8 T-cell populations were more functionally impaired than CD160− CD8 T-cell populations independently of PD-1 expression . Since reduced IL-2 production capacity is commonly associated with increased differentiation state [21] , the differentiation state was evaluated together with the expression of co-inhibitory molecules . We showed that co-inhibitory molecule expression is associated with differentiation state [22] . However in depth analyses indicated that significant differences are present within each CD8 T-cell populations , particularly among PD-1 and/or CD160-expressing populations and suggested that CD8 T-cell functions might be modulated by co-inhibitory molecule expression . The functions of CD160 are complex , and still remain under debate , with potentially various roles i . e . either stimulatory or inhibitory , depending on the cell type , the expression of the ligands ( HVEM or MHC class I ) [29]–[32] and the fact that CD160 exist in two isoforms ( spliced variants ) with and without transmembrane ( TM ) domain [33] , [34] . In order to determine whether CD160/CD160-ligand interactions could negatively regulate CD8 T-cell signaling , virus-specific CD8 T-cell proliferation was assessed in the presence or in the absence of mAbs that block CD160/CD160-ligand signaling . Since two molecules have been shown to interact with CD160 i . e . HVEM ( member of the TNF receptor superfamily ) and MHC class I molecules [29]–[32] , blocking anti-CD160 Abs that prevent both possible interactions were used [31] , [35] . Of note , HVEM is broadly distributed and expressed in hematopoietic and non-hematopoietic cells [36] . The ability of anti-CD160 Abs to increase virus-specific CD8 T-cell proliferation demonstrated that CD160 signaling negatively regulates TCR-mediated CD8 T-cell signaling . Interestingly , the level of restoration of virus-specific CD8 T-cell proliferation capacity induced by CD160 blockade was significantly more potent in cells harboring high level of CD160 and directly correlated with the ex vivo proportion of CD160 expression independently of PD-1 expression . These results suggested that 2B4+CD160+PD-1+ and 2B4+CD160+PD-1− CD8 T-cell populations were both rescued by CD160/CD160-ligands signaling blockade . We then showed that 2B4+CD160+PD-1+ , 2B4+CD160+PD-1− and 2B4+CD160−PD-1+ CD8 T-cell populations had reduced expression of perforin as compared to 2B4+CD160−PD-1− CD8 T-cell population , suggesting that the expression of CD160 and/or PD-1 on 2B4+ CD8 T cells is associated with a reduced capacity to express perforin . Interestingly , CD160 signaling is commonly associated with enhanced NK-cell cytotoxic activity [30]–[32] suggesting that CD160 might have distinct functions depending upon the cell subset considered . Indeed , CD160 is a GPI-anchored protein lacking both transmembrane ( TM ) and intracytoplasmic domains [33] , [34] ) . Therefore , the signaling cascade ( s ) triggered following CD160/CD160-ligand interactions might depend on the cell-type , the ligand ( either HVEM or MHC class I molecules ) , the density of CD160 expression , the potential co expression with CD160-TM and on the downstream adaptor proteins and/or signaling molecules , which are still not fully characterized [37] , [38] . We hypothesize that in CD8 T cells , CD160 might be associated with a specific phosphatase that might reduce the signal triggered following TCR/co-stimulatory receptors engagement . In addition , CD160 exists in two isoforms ( spliced variants ) with and without TM domain [33] , [34] . While NK cells can express both isoforms i . e . CD160 and CD160 TM , that are modulated following IL-15 stimulation [34] , CD8 T cells do not express CD160 TM directly ex vivo nor following IL-15 or TCR stimulation ( [34] and Fig . S6 ) . This feature may explain the dichotomic role of CD160 i . e . involved in CD8 T-cell negative regulation and NK-cell cytotoxic activity [17] , [18] , [30]–[32] . As previously demonstrated PD-1 signaling blockade by anti-PD-L1/2 Abs significantly increased virus-specific CD8 T-cell proliferation [12] , [13] , [17] . However , the PD-1 signaling blockade was significantly more potent on virus-specific CD8 T cells containing predominant 2B4−CD160−PD-1+ CD8 T-cell population as compared to virus-specific CD8 T cells with predominant 2B4+CD160−PD-1+ or 2B4+CD160+PD-1+ CD8 T-cell populations . These data suggest that cells expressing several co-inhibitory molecules are more profoundly impaired than cells expressing only one co-inhibitory molecule . Along the same line , PD-1 signaling blockade was on average more potent than CD160 signaling blockade in restoring virus-specific CD8 T-cell proliferation . Interestingly , the combined anti-CD160/anti-PD-L1/2 treatment showed additive effects only on virus-specific CD8 T cells containing both 2B4+CD160−PD-1+ or 2B4+CD160+PD-1− CD8 T-cell populations , but not on virus-specific CD8 T cells co-expressing PD-1 and CD160 . We postulate that cells expressing both co-inhibitory receptors might be intrinsically impaired , and their functions might be only partially reversible . Of note , PD-1 mean fluorescent intensity ( MFI ) was not significantly different between 2B4−CD160−PD-1+ and 2B4+CD160+PD-1+ CD8 T-cell populations ( Fig . S5 ) , suggesting that the difference in the restoration of proliferation capacity was not due to PD-1 expression level but likely on a different intrinsic capacity of the cells to proliferate . Taken together , these data suggest that the functional restoration induced by the blockade of several co-inhibitory molecules may be incomplete and that cells expressing several co-inhibitory molecules are more profoundly impaired than cells expressing only one co-inhibitory molecule . Consistently with a previous study [39] , we showed that PD-1 expression was up-regulated following T-cell activation and proliferation . However , CD160 expression was down-regulated following T-cell activation and proliferation . Thus , PD-1/PD-1-Ligand blockade may have an effect on both cells expressing PD-1 at the time of stimulation and on proliferating cells while CD160 blockade may only act on cells expressing CD160 at the time of stimulation . The differences in CD160 and PD-1 expression may contribute to understand 1 ) the higher potency observed of the PD-1/PD-1-Ligand blockade as compared to CD160 blockade in restoring virus-specific CD8 T-cell proliferation and 2 ) the absence of correlation between the level of restoration of virus-specific CD8 T-cell proliferation capacity induced by PD-1/PD-L1 blockade and the ex vivo proportion of PD-1-expressing CD8 T cells . Interestingly , the 2B4−CD160−PD-1+ CD8 T-cell population was significantly enriched in Flu-specific CD8 T cells as compared to EBV or CMV-specific CD8 T cells , and had higher IL-2 production capacity than all the other populations investigated . Furthermore , the 2B4−CD160−PD-1+ CD8 T-cell population expressed similar level of PD-1 ( MFI ) compared to the 2B4+CD160+PD-1+ CD8 T-cell population . The percentage of 2B4−CD160−PD-1+ CD8 T-cell population expressing EOMES , Tbet and CD57 was significantly lower than 2B4+CD160−PD-1+ and 2B4+CD160+PD-1+ CD8 T-cell populations . These data suggest that 2B4−CD160−PD-1+ CD8 T cells may be less functionally impaired than 2B4+CD160−PD-1+ and 2B4+CD160+PD-1+ CD8 T-cell populations . In conclusion , while accumulation of different immune check point blockers is clearly associated with progressive dysfunction , the present study demonstrates that CD160-mediated regulation of CD8 T-cell functions is independent of PD-1 expression thus providing evidence that CD160 contributes to the regulation of CD8 T-cell functions .
Forty subjects were recruited in this study . Blood samples were obtained at the local blood bank ( Centre de transfusion sanguine ( CTS ) , Lausanne , Switzerland ) . The use of samples used in this study was approved by the Institutional Review Board of the CTS , and all subjects gave written informed consent . Only individuals with no sign of HIV , HAV , HBV and HCV infections were included . Blood mononuclear cells were isolated as previously described [40] . The following monoclonal antibodies ( mAbs ) were used in different combinations . CD8-PB , CD3-APC-H7 , PD-1-PECy7 , PD-1-PB , IFN-γ-AF700 , IL-2-PE , CD4-PB , granzyme B-AF700 , perforin-APC were purchased from Becton Dickinson ( BD , San Diego , CA ) , CD45RA-ECD , CD4-ECD from Beckman Coulter ( Fullerton , CA , USA ) , CCR7-FITC from R&D Systems ( Minneapolis , MN , USA ) , 2B4-PECY5 . 5 , CD160-APC , CD160-PE from BioLegend ( San Diego , CA , USA ) , CD4-eFluor650NC , CD8-eFluor625NC from eBioscience . Cryo-preserved blood mononuclear cells ( 1–2×106 ) cells were washed , stained ( 30 min; 4°C ) for dead cells using the Aqua LIVE/DEAD stain kit ( Invitrogen ) and , when required , stained with appropriately tittered peptide-MHC class I multimer complexes at 4°C for 30′ in Ca2+-free media as described [24] . Cells were then washed and stained ( 30 min; 4°C ) with the following mAbs: CD3 , CD8 , CD4 , PD-1 , CD160 and 2B4 . Overnight-rested cryopreserved blood mononuclear cells ( 106 in 1 ml of complete medium ) were stained with 0 . 25 µM 5 , 6-carboxyfluorescein succinimidyl ester ( CFSE , Molecular Probes , USA ) as previously described [41] , and stimulated with CMV , EBV or Flu peptides ( 1 µg/ml ) , 200 ng/ml of SEB ( positive control; Sigma-Aldrich ) or left unstimulated ( negative control ) in presence or in absence of anti-CD160 ( 2 µg/ml; MBL ) and/or anti-PD-L1/2 ( 10 µg/mL ) [13] . At day 6 , cells were harvested and stained ( 4°C; 20 min ) using the violet or aqua LIVE/DEAD stain kit ( Invitrogen ) and Abs ( 4°C; 30 min ) to CD3 , CD4 , CD8 . Frequencies of proliferating CFSElow CD8 T cells were assessed by flow cytometry . Cryopreserved blood mononuclear cells ( 106 in 1 ml of complete medium ) were stained with 0 . 25 µM 5 , 6-carboxyfluorescein succinimidyl ester ( CFSE , Molecular Probes , USA ) as previously described [41] , and stimulated with anti-CD3/anti-CD28 magnetic beads ( Dynabeads , Invitrogen ) . The expression of CD160 and PD-1 on CD8 T cells was evaluated after 0 , 24 , 48 , 72 and 120 hours of stimulation by anti-CD3/anti-CD28 magnetic beads . CD8 T-cell proliferation was also evaluated by CFSE dilution . PBMC were stimulated overnight in complete media ( RPMI ( Invitrogen ) , 10% fetal calf serum ( FCS; Invitrogen ) , 100 µg/ml penicillin , 100 unit/ml streptomycin ( BioConcept ) ) with Staphyloccocus enterotoxin B ( SEB; 250 ng/mL ) or left unstimulated ( negative control ) in the presence of Golgiplug ( 1 µl/ml; BD ) and anti-PD1 , anti-CD160 and anti-2B4 mAbs . At the end of the stimulation period , cells were washed , stained ( 20 min; 4°C ) for dead cells using the Aqua LIVE/DEAD stain kit ( Invitrogen ) , permeabilized ( 20 min; 20°C ) ( Cytofix/Cytoperm , BD ) and stained ( 30 min; 20°C ) with mAbs to CD3 , CD8 , CD45RA , CCR7 , IFN-γ and IL-2 [42] . PBMC were washed , stained ( 20 min; 4°C ) for dead cells using the Aqua LIVE/DEAD stain kit ( Invitrogen ) . Cells were then washed and stained ( 30 min; 4°C ) with the following mAbs: CD3 , CD8 , PD-1 , 2B4 , CD160 , CCR7 and CD45RA . Cells were then permeabilized ( 20 min; 20°C ) ( Cytofix/Cytoperm , BD ) and stained ( 30 min; 20°C ) with mAbs to perforin and granzyme B [23] . PBMC were washed , stained ( 20 min; 4°C ) for dead cells using the Aqua LIVE/DEAD stain kit ( Invitrogen ) . Cells were then washed and stained ( 30 min; 4°C ) with the following mAbs in different combinations: CD3 , CD8 , PD-1 , 2B4 , CD160 and CD57 or CD3 , CD8 , PD-1 , 2B4 , CD160 , CCR7 and CD45RA . Cells were then permeabilized ( 1 h; 4°C ) ( Foxp3 Fixation/Permeabilization Kit; eBioscience ) and stained ( 30 min; 4°C ) with mAbs to EOMES and T-bet . Cryopreserved CD8 T cells ( 30–40×106 cells ) previously selected by MACS cell separation ( Miltenyi kit ) were washed , stained ( 20 min; 4°C ) for dead cells using the Aqua LIVE/DEAD stain kit ( Invitrogen ) . Cells were then washed and stained ( 30 min; 4°C ) with the following mAbs: CD8 , PD-1 , 2B4 , CD160 , CD45RA , CCR7 . CD8 T-cell populations were sorted by BD FACSAria III on the basis of 2B4 , PD-1 and CD160 expression after exclusion of naïve T cells ( CD45RA+CCR7+ ) . Sorted populations were labeled with CFSE and stimulated with viral peptides ( 1 µg/mL ) in the context of irradiated ( 40 Gy ) autologous CD8-depleted PBMCs ( ratio CD8/feeder cells 1∶10 ) or in anti-CD3 ( 10 ug/ml ) and anti-CD28 ( 0 . 5 ug/ml ) mAbs coated plate or left unstimulated . Proliferation was assessed at day 6 by flow cytometry . Cells were washed , stained ( 20 min; 4°C ) for dead cells using the Aqua LIVE/DEAD stain kit ( Invitrogen ) . Cells were then washed and stained ( 30 min; 4°C ) with anti-CD8 and anti-CD3 mAbs . CD160 and CD160-TM expression were assessed both ex vivo and after in vitro expansion . To assess their expression ex vivo PBMC were washed , stained ( 20 min; 4°C ) for dead cells using the Aqua LIVE/DEAD stain kit ( Invitrogen ) . Cells were then washed and stained ( 30 min; 4°C ) with the following mAbs: CD8 , CD3 , CD56 , CD160 , CD160-TM ( rabbit unconjugated ) . Cells were then washed and stained ( 30 min; 4°C ) with anti-rabbit secondary Ab . To assess their expression after in vitro expansion PBMC were labeled with CFSE and stimulated with IL-15 ( 50 ng/mL ) or with anti-CD3 ( 10 ug/ml ) and anti-CD28 ( 0 . 5 ug/ml ) mAbs coated plate . CD160 and CD160-TM expression were assessed at day 6 by flow cytometry . Cells were washed , stained ( 20 min; 4°C ) for dead cells using the Aqua LIVE/DEAD stain kit ( Invitrogen ) . Cells were then washed and stained ( 30 min; 4°C ) with the following mAbs: CD8 , CD3 , CD56 , CD160 , CD160-TM ( rabbit unconjugated ) . Cells were then washed and stained ( 30 min; 4°C ) with anti-rabbit secondary Ab . Cells were fixed with CellFix ( BD ) , acquired on an LSRII SORP ( 4 lasers: 405 , 488 , 532 and 633 nm ) and analyzed using FlowJo ( version 8 . 8 . 2 ) ( Tree star Inc , Ashland , OR , USA ) and SPICE 4 . 2 . 3 . When required , analysis and presentation of distributions was performed using SPICE version 5 . 1 , downloaded from <http://exon . niaid . nih . gov/spice> [43] . The number of lymphocyte-gated events ranged between 5×105 and 106 in the flow cytometry experiments . Statistical significance ( P values ) was obtained using one-way ANOVA ( Kruskal-Wallis test ) followed by Student's t test for multiple comparisons or a Spearman rank test for correlations using GraphPad Prism version 5 . 0 ( San Diego , CA ) . Statistical analyses of global cytokine profiles ( pie charts ) were performed by partial permutation tests using the SPICE software as described [43] . | T-cell immune response is regulated by a variety of molecules known as co-inhibitory receptors . The over expression of co-inhibitory receptors has been observed in several chronic viral infections such as HIV disease , and is found to be associated with severe T-cell dysfunction . Recent studies have demonstrated that the co-expression of several co-inhibitory receptors correlated with greater impairment of CD8 T cells . However , the relative contribution of individual co-inhibitory receptors to the regulation of T-cell functions remains unclear . In order to shed light on these issues , we have evaluated the influence of the expression of 3 major co-inhibitory receptors such as PD-1 , 2B4 and CD160 on CD8 T-cell functions such as proliferation , cytokines production and expression of cytotoxic granules . We demonstrate that CD160-associated CD8 T-cell functional impairment is independent of PD-1 expression and that the blockade of CD160 signaling may partially restore CD8 T-cell functions . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"biology",
"and",
"life",
"sciences",
"immunology"
] | 2014 | CD160-Associated CD8 T-Cell Functional Impairment Is Independent of PD-1 Expression |
Exhausted CD8+ T cell responses during chronic viral infections are defined by a complex expression pattern of inhibitory receptors . However , very little information is currently available about the coexpression patterns of these receptors on human virus-specific CD8+ T cells and their correlation with antiviral functions , T cell differentiation and antigen recognition . We addressed these important aspects in a cohort of 38 chronically HCV infected patients and found a coexpression of inhibitory receptors such as 2B4 , CD160 and KLRG1 in association with PD-1 in about half of the HCV-specific CD8+ T cell responses . Importantly , this exhaustive phenotype was associated with low and intermediate levels of CD127 expression , an impaired proliferative capacity , an intermediate T cell differentiation stage and absence of sequence variations within the corresponding epitopes , indicating ongoing antigen triggering . In contrast , a low expression of inhibitory receptors by the remaining HCV-specific CD8+ T cells occurred in concert with a CD127hi phenotype , an early T cell differentiation stage and presence of viral sequence variations within the corresponding epitopes . In sum , these results suggest that T cell exhaustion contributes to the failure of about half of HCV-specific CD8+ T cell responses and that it is determined by a complex interplay of immunological ( e . g . T cell differentiation ) and virological ( e . g . ongoing antigen triggering ) factors .
Virus-specific CD8+ T cells play a central role in the outcome of HCV infection . Indeed , several human and animal studies have shown associations between strong and multispecific T cell responses and viral clearance [1] . During chronic HCV infection , viral escape and an impairment of HCV-specific CD8+ T cell antiviral functions , e . g . the ability to proliferate or to secrete antiviral cytokines such as interferon-γ ( IFN-γ ) contribute to virus-specific CD8+ T cell failure . The underlying mechanisms for the functional impairment of HCV-specific CD8+ T cells have not been clarified in detail , although lack of CD4+ T cell help , action of regulatory T cells and expression of immunomodulatory cytokines , such as Il-10 , have been suggested to contribute [1] . In addition , expression of the inhibitory receptor PD-1 has been postulated to characterize a state of exhaustion of HCV-specific CD8+ T cells in chronic HCV infection in analogy to murine models of chronic viral infections [2] . Indeed , analysis of patients with chronic HCV infection identified high levels of PD-1 expression on HCV-specific CD8+ T cells in blood and liver [3] and blockade of PD-1 signaling resulted in the functional restoration of blood-derived HCV-specific CD8+ T cell responses in chronic infection [3] , [4] . However , the relevance of PD-1 in defining exhausted HCV-specific CD8+ T cells has not been unchallenged . For example , PD-1 blockade alone was unable to restore the function of liver-derived HCV-specific CD8+ T cells [5] while targeting additional inhibitory signaling pathways reinvigorated the antiviral function [6] . In addition , PD-1 expression did not necessarily identify exhausted HCV-specific CD8+ T cells during acute HCV infection in humans [7] and chimpanzees [8] . Thus , PD-1 expression alone may not be sufficient to determine exhaustion of HCV-specific CD8+ T cells during HCV infection . In this context , it is interesting to note that a recent study identified coexpression of additional inhibitory receptors next to PD-1 as a critical determinant of CD8+ T cell exhaustion in a murine model of chronic viral infection . For example , expression of several inhibitory receptors including 2B4 and CD160 next to PD-1 was identified on strongly exhausted virus-specific CD8+ T cells in severe LCMV infection [9] . 2B4 is a coregulatory receptor of the SLAM-receptor family that is capable of mediating both activatory and inhibitory signals upon ligand CD48 binding [10] . High levels of 2B4 expression are associated with 2B4 inhibitory receptor function [10] and indeed , 2B4 and CD160 , a glycoylphosphatidylinositol-anchored receptor that coinhibits T cells upon ligand HVEM binding [11] , both highly correlated in gene expression analysis with PD-1 expression [12] . The positive correlation with PD-1 expression was also observed for other inhibitory receptors , such as LAG-3 , a protein closely related to CD4 , and CTLA-4 , a member of the CD28 family of receptors [9] . Of note , expression of the inhibitory receptor Killer-cell-lectin-like receptor G1 ( KLRG1 ) showed no correlation with exhaustion in severe LCMV infection [9] . KLRG1 mediates proliferative dysfunction of differentiated CD8+ T cells upon ligand E-cadherin binding [13] , [14] , [15] and its upregulation is thought to be mediated by ongoing antigen triggering [16] . Despite these insights into inhibitory receptor expression on CD8+ T cells , very little information is currently available about the coexpression of these receptors on virus-specific CD8+ T cells during chronic human infection such as HCV and their association with T cell function , differentiation and antigen recognition . Here , we show that the coexpression of inhibitory receptors such as PD1 , 2B4 , CD160 and KLRG1 occurs in about half of HCV-specific CD8+ T cell responses and that it is linked to low or intermediate levels of CD127 expression , impaired proliferative capacity , an intermediate T cell differentiation stage and ongoing antigen-recognition . In contrast , the absence of inhibitory receptor expression is associated with a CD127 high phenotype and the presence of virus sequence variations in the respective epitopes , indicating viral escape . These results indicate that two different mechanisms contribute to the dysfunction of the HCV-specific CD8+ T cell pool in chronic HCV infection . These results also indicate that different factors are involved in the development of HCV-specific CD8+ T cell exhaustion .
In a first set of experiments , we screened HCV-specific CD8+ T cells for the expression of a large set of inhibitory receptors such as PD-1 , 2B4 , CD160 , KLRG1 , LAG-3 and CTLA-4 in a cohort of 10 chronically HCV infected patients . We compared the expression of these receptors on HCV-specific CD8+ T cells with the expression on FLU-specific CD8+ T cells that represent a memory population . As shown in Fig . 1 , we found an elevated expression of PD-1 , 2B4 , CD160 and KLRG1 by HCV-specific CD8+ T cells , however , no significant increase of LAG-3 and CTLA-4 expression was observed . Based on these results , we focused in the following on the analysis of the expression of PD-1 , 2B4 , CD160 and KLRG1 by HCV-specific CD8+ T cells in a cohort of 38 patients with chronic HCV that displayed HCV-specific tetramer responses ( Table 1 ) . As shown in Fig . 2A , we found a high expression of PD-1 in our cohort ( median: 88 . 9% ) . Although PD-1 expression was high on most HCV-specific CD8+ T cell responses , it was not detectable on all HCV-specific CD8+ T cells ( Fig . 2 ) . 2B4 was expressed by some HCV-specific CD8+ T cells with a median expression of 62 . 0% ( Fig . 2A and B ) , whereas CD160 was detectable only on a fraction of virus-specific CD8+ T cells ( median: 7 . 4% ) ( Fig . 2A and B ) . Expression of KLRG1 was found to be quite variable , with a median expression of 40 . 8% ( Fig . 2A and B ) . Previously , we have shown that HCV-specific CD8+ T cells consist of subsets with distinct phenotypical and functional properties that can be defined by CD127 expression [17] . Thus , we asked whether the expression of CD127 is also linked to the expression of the inhibitory receptors PD-1 , 2B4 , CD160 and KLRG1 by HCV-specific CD8+ T cells . To address this question , we costained the inhibitory receptors and CD127 on HCV-specific CD8+ T cells . Depending on the expression level of CD127 ( Fig . S1 ) , we defined three groups of HCV-specific CD8+ T cells as displayed in Fig . 3A: a CD127 high ( hi ) group with more than 80% of epitope-specific CD8+ T cells expressing CD127 , a CD127 intermediate ( mid ) group with 50–80% of CD127 expression and a CD127 low ( lo ) group with less than 50% expression of CD127 . Importantly , we found a clear association between the expression of CD127 and the expression of inhibitory receptors on HCV-specific CD8+ T cells . Indeed , CD127lo cells highly expressed PD-1 ( median 97 . 5% ) , 2B4 ( median 80 . 0% ) and CD160 ( median 19 . 3% ) ( Fig . 3B ) . CD127mid cells expressed slightly reduced levels of PD-1 ( median 86 . 7% ) , 2B4 ( median 66 . 4% ) and CD160 ( median 8 . 9% ) . In contrast , CD127hi cells showed a significantly lower expression of PD-1 ( median 48 . 3% ) , 2B4 ( median 35 . 7% ) and CD160 ( median 3 . 3% ) ( Fig . 3B ) . KLRG1 was also higher expressed on HCV-specific CD8+ T cells with a CD127lo phenotype compared to a CD127mid and CD127hi phenotype ( 67 . 1% vs 40 . 8% vs 25 . 0% ) ( Fig . 3B ) . In contrast to PD-1 , 2B4 and CD160 , however , the range of KLRG1 expression varied more strongly between both groups . Next , we analyzed the coexpression of inhibitory receptors on HCV-specific CD8+ T cells in 10 patients at the single cell level in a multi-inhibitory receptor staining ( Fig . S2 ) . As shown in Fig . 3C , a CD127lo phenotype was associated with a high frequency of HCV-specific CD8+ T cells coexpressing multiple inhibitory receptors . The inhibitory receptor coexpression profile of CD127mid HCV-specific CD8+ T cells was reduced compared to CD127hi cells . For example , in our cohort , more than 80% of CD127lo cells coexpressed 2 or more inhibitory receptors while coexpression was found in less than 55% of CD127mid cells ( Fig . 3C ) . In contrast , a CD127hi phenotype was associated with low or absent levels of inhibitory receptor coexpression . These results clearly illustrate that the coexpression of multiple inhibitory receptors by HCV-specific CD8+ T cells is linked to low levels of CD127 expression . Next , we asked whether the coexpression of multiple inhibitory receptors on CD127lo CD8+ T cells defines functionally impaired HCV-specific CD8+ T cells , indicating exhaustion . We addressed this by stimulating CD127lo , CD127mid and CD127hi HCV-specific CD8+ T cells antigen-specifically for one week and by analyzing the subsequent proliferation in the presence or absence of PD-L1 blockade . Of note , we found significant differences regarding the proliferative capacity and the effect of PD-L1 blockade between these groups ( Fig . 4 ) . Indeed , the proliferative capacity of CD127hi HCV-specific CD8+ T cells was much higher upon antigen stimulation compared to CD127mid and CD127lo HCV-specific CD8+ T cells . As shown in Fig . 4A and 4E , CD127lo HCV-specific CD8+ T cells that expressed multiple inhibitory receptors proliferated poorly upon antigen stimulation . However , these cells could be reinvigorated partially upon blockade of the inhibitory PD-1/PD-L1 signaling pathway , indicating that they were in a functional reversible state of exhaustion . CD127mid HCV-specific CD8+ T cells displayed an improved proliferative capacity upon antigen stimulation ( Fig . 4B ) . In contrast to CD127lo and CD127mid cells , CD127hi HCV-specific CD8+ T cells proliferated vigorously ( Fig . 4C ) upon stimulation with peptide alone . Additional PD-L1 blockade resulted in only a minor increase in proliferation . In a subsequent analysis of 15 patients , we found that the increase in HCV-specific CD8+ T cells due to PD-L1 blockade was about 2 . 2fold for CD127lo cells . This increase was significantly higher compared to CD127mid ( 1 . 5fold ) and CD127hi cells ( 1 . 2fold ) ( p<0 . 01 ) ( Fig . 4D ) . These results indicate that CD127lo HCV-specific CD8+ T cells are functionally impaired in vivo and are functionally dependent on inhibitory receptor blockade , indicating exhaustion . It has previously been shown that the expression of inhibitory CD8+ T cell receptors , e . g . PD1 is linked to the stage of T cell differentiation [18] , [19] . Thus , we asked whether the coexpression of CD127 and all inhibitory receptors analyzed in our study was linked to certain stages of T cell differentiation . For this analysis , we assessed post-thymic human CD8+ T cell differentiation using a linear model based on CD27 , CCR7 and CD45RA expression as displayed in Fig . 5A [18] . CD127 expression was highest in naïve and early T cell subsets ( subsets 1 and 2 ) , but reduced in late differentiated CD8+ T cells , indicating that CD127 expression was linked to early stages of T cell differentiation ( Fig . 5B ) . In agreement with a previous study , the expression of PD-1 was highest on intermediate differentiated CD8+ T cells ( subset 3 ) and lowest on naïve and late differentiated CD8+ T cells [19] ( Fig . 5C ) . Interestingly , the expression patterns of 2B4 , CD160 and KLRG1 did not mirror the expression pattern of PD-1 but displayed unique , individual profiles . 2B4 expression was lowest on naïve CD8+ T cells but progressively increased along the differentiation subsets reaching almost 100% expression on late-differentiated CD8+ T cells ( Fig . 5D ) . CD160 expression was lowest in naïve T cells and increased in further differentiated subsets with two expression peaks in intermediate subset 3 and late differentiated T cell subset 5 ( Fig . 5E ) while KLRG1 showed a similar expression pattern as 2B4 being low in naïve and early subsets but enriched in late differentiated subsets ( Fig . 5F ) . Overall , these results indicate that inhibitory receptors are increasingly expressed by further differentiated CD8+ T cells . However , since very distinct patterns of expression were observed for each marker , the expression of one inhibitory receptor does not necessarily indicate coexpression of others . Thus , we asked whether coexpression of the inhibitory receptors 2B4 , CD160 and KLRG1 together with PD-1 occurred in a specific T cell differentiation subset . Indeed , as shown in Fig . 6 , we identified the highest level of coexpression of all markers in the intermediate T cell subset 3 . These results indicate that inhibitory receptors exhibit individual and distinct expression patterns linked to T cell differentiation and that coexpression of 2B4 , CD160 and KLRG1 with PD-1 indicating exhaustion occurs in intermediate rather than late differentiated CD8+ T cell subsets . Finally , we asked whether the same association can also be found within the HCV-specific CD8+ T cell population . As shown in Fig . 7 and in agreement with previous findings , we observed that HCV-specific CD8+ T cells consisted mostly of early and intermediate differentiated T cells . Interestingly , the differentiation of HCV-specific CD127hi , CD127mid and CD127lo cells was significantly different ( p<0 . 0001 ) . The majority of CD127hi cells belonged to the early differentiation subset 2 , whereas CD127mid cells were distributed at comparable levels in subset 2 and subset 3 . In contrast , HCV-specific CD127lo cells were prominently found within the intermediate subset 3 . Keeping in mind that the coexpression of inhibitory receptors is highest in the intermediate subset 3 and associated with a CD127lo phenotype , these results clearly suggest that there is a complex link between CD127 expression , T cell differentiation and T cell exhaustion . In a final set of experiments , we set out to determine whether the differential expression of inhibitory receptors by phenotypically and functionally distinct HCV-specific CD8+ T cell populations is related to the autologous viral sequence present in a given patient . To address this issue , we sequenced the autologous viral sequences of genotype 1 patients corresponding to recognized CD8+ T cell epitopes ( Table 2 ) . Importantly , sequence variations from the genotype consensus sequence were significantly more prevalent within epitopes targeted by CD127hi HCV specific CD8+ T cells with a low expression of inhibitory receptors ( p<0 , 0001 ) ( Fig . 8A ) . We confirmed that these sequence variations represented viral escape in several patients by stimulating HCV-specific CD8+ T cells with peptides matched to both the autologous and consensus sequence . Indeed , as shown for six representative patients in Fig . 8B , HCV-specific CD8+ T cells produced high levels of IFN-γ upon stimulation with the consensus sequence peptide , but not when stimulated with the variant autologous peptide . Furthermore , proliferation of these HCV-specific CD8+ T cells was significantly reduced upon stimulation with the variant autologous peptide , indicating viral escape ( data not shown ) . These results indicate that viral escape mutations are linked to a CD127hi phenotype whereas a CD127mid or CD127lo phenotype is linked to ongoing antigen recognition of the autologous virus sequence . To further test this hypothesis , we analyzed HCV-specific CD8+ T cells from patients with multiple immunodominant T cell responses that harbored consensus and variant sequences in different epitopes . Indeed , different expression profiles of inhibitory receptors and CD127 could be observed within the same patient depending on the virus sequence ( patient 4 , Fig . 8C , Table 2 ) . Indeed , although both immunodominant responses were detectable at a comparable frequency ( Fig . S3 ) , all NS3-1406-specific CD8+ T cells expressed CD127 , but only low levels of PD-1 ( 14 . 3% ) , 2B4 ( 4 . 17% ) , CD160 ( 2 . 09% ) and KLRG1 ( 25 . 0% ) , whereas an opposite phenotype was observed on NS5-2594 specific CD8+ T cells , of which only 9 , 09% expressed CD127 , but high levels of PD-1 ( 86 . 7% ) , 2B4 ( 81 . 8% ) , CD160 ( 27 . 3% ) and KLRG1 ( 86 . 7% ) ( Fig . 8C , 8D ) . These results underline that the mechanisms responsible for inhibitory receptor coexpression operate in an epitope-specific manner depending on the autologous virus sequence . To further test the hypothesis that the low expression of inhibitory receptors and high expression of CD127 are indeed caused by absence of antigen triggering in vivo , we stimulated CD127hi HCV-specific CD8+ T cells with the consensus peptide in vitro . Importantly , this peptide specific stimulation induced a downregulation of CD127 and an upregulation of inhibitory receptor expression after a seven day culture ( shown in Fig . 9 ) . In sum , the downregulation of CD127 expression and upregulation of inhibitory receptors upon antigen stimulation suggest that the absence of antigen triggering in vivo is responsible for a CD127hi phenotype with a low expression of inhibitory receptors . In agreement with this hypothesis , we found a low expression of CD127 and a high expression of inhibitory receptors by virus-specific CD8+ T cells that targeted epitopes without sequence variations , indicating ongoing antigen recognition ( Fig . 8A ) . These results also underscore the importance of evaluating autologous viral sequences in studies aimed at investigating the mechanisms of virus-specific CD8+ T cell failure .
Here , we have investigated the expression of inhibitory receptors by virus-specific CD8+ T cells during chronic HCV infection and their association with CD127 expression , proliferative capacity , T cell differentiation and antigen recognition . We found a specific pattern of expression of inhibitory receptors by HCV-specific CD8+ T cells . Indeed , a large fraction of the HCV-specific CD8+ T cell responses detectable in our cohort coexpressed 2B4 , CD160 and KLRG1 next to PD1 . These results are interesting in the light of a recent study by Blackburn et al . demonstrating a clear association between a specific expression pattern of inhibitory receptors including 2B4 , CD160 , LAG-3 and CTLA-4 next to PD-1 and severe exhaustion of virus-specific CD8+ T cells in LCMV infection in mice [9] . Interestingly , we did not observe a significant expression of LAG-3 and CTLA-4 on HCV-specific CD8+ T cells in our study . However , these results are not surprising since it has previously been shown that CTLA-4 expression by HCV-specific CD8+ T cells is only detectable to some degree in the liver , but not in the blood of chronically infected patients [5] . In addition , a gene expression analysis of exhausted CD8+ T cells in mice revealed that LAG-3 and CTLA-4 coexpression showed a weaker association with PD-1 compared to 2B4 and CD160 coexpression [12] . Thus , these combined results indicate a dynamic hierarchical pattern of inhibitory receptor expression by exhausted CD8+ T cells with prominent roles of 2B4 and CD160 [12] . Accordingly , our results indicate that a significant fraction of HCV-specific CD8+ T cells display typical markers of exhaustion . Importantly , HCV-specific CD8+ T cell exhaustion defined as a loss of CD8+ T cell function , e . g . proliferation and cytokine secretion has been reported by several groups [4] , [5] , [20] , [21] , [22] . Our findings implicate that therapeutic targeting of exhausted CD8+ T cells during chronic HCV infection may require not only the blockade of one inhibitory receptor pathway , but rather a cocktail of antibodies targeting multiple inhibitory pathways [6] . In line with these findings , we observed only a 2 . 2fold increase in proliferation of exhausted CD127lo HCV-specific CD8+ T cells upon PD-1/PD-L1 blockade ( Fig . 4D ) . An even smaller increase in proliferation due to PD-1/PD-L1 blockade was seen in CD127hi HCV-specific CD8+ T cells , however , this may be explained by the finding that these cells proliferate vigorously in the absence of blockade , indicating that they are not significantly inhibited ( Fig . 4C–E and [17] ) . Clearly , additional studies should address the effect of 2B4 , CD160 and KLRG1 receptor blockade on exhausted T cells as soon as suitable blocking reagents are available . 2B4 is a coregulatory receptor capable of mediating activatory and inhibitory signals in CD8+ T cells [23] . It was previously considered as a predominantly activating receptor in humans [24] . However , a recent study addressed the dual roles of murine and human 2B4 receptor signaling and found that both human and murine 2B4 can exert both inhibitory or activatory receptor functions [10] . Importantly , high levels of 2B4 surface expression were linked to inhibitory 2B4 function , whereas low levels of 2B4 surface expression promoted activatory 2B4 signals [10] . Based on the data analyzed in our study , we cannot exclude that ligation of 2B4 on HCV-specific CD8+ T cells may result in a costimulatory signal . However , we used a strict gating strategy to determine 2B4 positivity only on cells with a high surface expression level of 2B4 ( Fig . 2B and data not shown ) , indicating that 2B4 may act as an inhibitory receptor on these cells . Further studies will be needed to clarify whether 2B4 expression on virus-specific CD8+ T cells , indeed , promotes inhibitory signals . Of note , we observed significant levels of KLRG1 expression on HCV-specific CD8+ T cells ( Fig . 1 , 2 ) . KLRG1 is up-regulated by virus-specific CD8+ T cells upon repetitive antigenic stimulation and frequently used as a marker for T cell differentiation [16] , [25] . Although KLRG1 expression was not shown to be strongly associated with PD-1 coexpression by exhausted CD8+ T cells in mice it was , however , identified on a significant fraction of exhausted CD8+ T cells [9] , [12] . Since the up-regulation of KLRG1 is induced by repetitive antigen stimulation , it is well possible that KLRG1 expression might not directly reflect T cell exhaustion but rather ongoing antigen recognition , that , however , by itself is thought to be the prerequisite for the development of CD8+ T cell exhaustion [26] , [27] . Our results also show that the expression level of the inhibitory receptors PD-1 , 2B4 , CD160 and KLRG1 is linked to a specific stage of cell differentiation . Recently , a linear model of human CD8+ T cell differentiation has been developed that distinguishes between naïve , early , intermediate and late differentiated cells [28] . The differentiation status of CD8+ T cells is influenced by the history of antigenic stimulation , clonal cell division , telomere length and proliferative capacity of CD8+ T cells [18] . Here , we found that the expression of PD-1 was highest on CD8+ T cells with an intermediate T cell differentiation phenotype . In contrast , 2B4 , CD160 and KLRG1 were differentially expressed by late differentiated CD8+ T cells ( Fig . 5 ) . However , the highest level of coexpression of all inhibitory receptors was observed within the intermediate T cell differentiation stage ( Fig . 6 ) , indicating that exhaustion of CD8+ T cells is most likely linked to this specific subset . It remains unclear whether the coexpression of inhibitory receptors 2B4 , CD160 and KLRG1 by late differentiated cells in the absence of PD-1 may also define a specific subset of exhausted CD8+ T cells . However , this is rather unlikely since coexpression of inhibitory receptors 2B4 and CD160 by late T cell differentiated T cells has been shown to define a functional cytotoxic CD8+ T cell population [24] , [29] . Another important finding of our study is that coexpression of inhibitory receptors on HCV-specific CD8+ T cells is linked to the expression level of CD127 . We have previously described the existence of subsets of HCV-specific CD8+ T cells with distinct functional properties that can be distinguished by the level of CD127 expression [17] . Here , we classified HCV-specific CD8+ T cells depending on the level of CD127 expression into three groups: CD127lo , CD127mid and CD127hi cells . CD127lo HCV-specific CD8+ T cells express markers associated with effector cells and are dysfunctional in terms of proliferation whereas CD127hi cells phenotypically and functionally resemble memory CD8+ T cells [17] . These memory-like properties are also reflected by an increased expression of antiapoptotic molecules [30] , [31] ( Fig . S3 ) . Here , we found that high coexpression of inhibitory receptors next to PD-1 was restricted to CD127lo HCV-specific CD8+ T cells with impaired proliferative capacity that could be reinvigorated upon PD-1/PD-L1 signaling blockade , thus , correlating well with the impaired functional state of these cells . We also found that HCV-specific CD8+ T cells do not represent a homogenously differentiated CD8+ T cell population . This novel finding was made possible through the use of polychromatic flow cytometry that allowed us to simultaneously stain HCV-specific CD8+ T cells for the set of differentiation markers recommended by Appay et al . [18] and extends previous reports regarding the differentiation of HCV-specific CD8+ T cells that used conventional 4-color flow cytometry [18] , [28] . Of note , the differentiation stage of HCV-specific CD8+ T cells was significantly different for CD127hi , CD127mid and CD127lo cells ( Fig . 7 ) . In addition , the intermediate differentiation of HCV-specific CD8+ T cells characterized by the coexpression of multiple inhibitory receptors ( as observed prominently for CD127lo cells ) correlates with the coexpression of inhibitory receptors in intermediate differentiation subsets observed on the general CD8+ T cell population . Since these results underline the existence of phenotypically and functionally very different HCV-specific CD8+ T cell populations during chronic HCV infection that can be defined by the differential expression of CD127 and inhibitory receptors , respectively , it will be important to distinguish between these populations in future studies analyzing HCV-specific CD8+ T cells . Another important finding of our study is that the expression of inhibitory receptors by virus-specific CD8+ T cells is associated with ongoing antigen recognition and the absence of viral escape . Since we could not analyze the virus sequence that initially infected our patients , and we thus could not observe the development of escape mutations as has been elegantly shown in recent longitudinal studies in acute HCV infection [32] , [33] , we assumed that the majority of patients was infected with the genotype consensus sequence ( table 2 ) . This approach has some limitations , for example it cannot be ruled out that remnants of a prior infection with an alternate genotype or varying epitope sequence confounded our analysis ( for example , the sequence of the autologous epitope matches the genotype consensus sequence in one CD127hi patient ( the NS3-1406 epitope in subject 9 ) ) . In addition , it is impossible to dissect whether a variant epitope that matches the heterologous genotype sequence that is targeted by a CD127hi T cell response represents viral escape or whether the T cell response represents the remnant of a prior resolved infection with the alternate genotype ( e . g . , the NS5-2594 epitope in subject 32 ) . Clearly , these issues can only be addressed in cohorts of HCV-infected patients , in which the sequence of the inoculum and information about the T cell response prior to infection is known . However , the clear association of high levels of CD127 expression on virus-specific CD8+ T cells with the presence of autologous sequence variations from the consensus epitopes , the reduced IFN-γ production and reduced proliferation of CD8+ T cells upon stimulation with variant but not with consensus peptides supports our approach in the majority of cases . Of note , we also found a clear association between the expression level of inhibitory receptors next to PD1 and the absence of sequence variations indicating ongoing antigen recognition . These results are in agreement with recent studies showing that high levels of specific antigen drive CD8+ T cell exhaustion [27] and that exhaustion of CD8+ T cells can be prevented by the emergence of viral escape mutations that abrogate epitope recognition [34] . In line with these findings , emergence of viral sequence variations was associated with a reduction of PD-1 expression and upregulation of CD127 expression in acute and chronic HCV infection [32] , [33] and HIV infection [35] . It is also interesting to note that we observed sequence variations in about half of the targeted T cell epitopes . Although the number of HCV-specific CD8+ T cell epitopes analyzed in our study was limited , and it is thus hard to make firm conclusions about the general frequencies of viral escape among HCV-specific CD8+ T cell epitopes , our results are in agreement with recent studies that also have shown the presence of viral escape mutations in about 50% of targeted CD8+ T cell epitopes [36] , [37] . Thus , both mechanisms of CD8+ T cell failure , viral escape and T cell exhaustion , defined by the coexpression of multiple inhibitory receptors , seem to contribute significantly to the ineffective viral control and they can be easily identified by specific surface expression patterns . In sum , our results show that the coexpression of inhibitory receptors by exhausted HCV-specific CD8+ T cells is linked to CD127 expression , proliferative capacity , the stage of T cell differentiation and ongoing antigen recognition . Thus , a complex interplay of immunological and virological factors determines T cell exhaustion in human chronic viral infection . These findings have also important implications for the rational design of immunotherapeutic treatment strategies in chronic HCV infection since only exhausted CD8+ T cells still recognize the present viral antigens and should thus be targeted by immunomodulatory strategies .
38 patients with chronic HCV infection presenting at the outpatient hepatology clinic of the University Hospital Freiburg with detectable HCV-specific CD8+ tetramer responses were included in the study after obtaining written informed consent from the patients and approval by the ethics committee of the Albert-Ludwigs-Universität , Freiburg . All investigations have been conducted according to the principles expressed in the Declaration of Helsinki . The characteristics of the study population are included in table 1 . Detection of HCV viral load and epitope sequencing was performed as previously described [37] . Lymphocytes were analyzed from patient blood as previously described [17] . Peptides corresponding to immunodominant HLA-A2- and HLA-B27- epitopes ( sequences CINGVCWTV , KLVALGINAV , ALYDVVTKL , ARMILMTHF ) and variant autologous patient sequences were obtained from Biosynthan , Berlin , Germany . These peptides were dissolved in 100% dimethyl sulfoxide ( Sigma-Aldrich , Germany ) at 20 mg/ml and further diluted to 1 mg/ml with RPMI 1640 ( Gibco ) before use . HLA-A2- and HLA-B27-tetramers were obtained from the National Tetramer Core Facility at Emory University , Atlanta . Influenza-specific pentamers containing the immunodominant GILGFVFTL peptide were purchased from ProImmune , Oxford , UK . The following reagents were used for polychromatic stainings: anti-2B4-FITC , anti-CD27-APC-eFluor780 , anti-CD45RA-PerCP-Cy5 . 5 , anti-PD-1-PE , anti-CD127-APC Alexa Fluor 750 , anti-CD127-APC-eFluor 780 , anti-CD127-Pacific Blue , anti-IgM-PerCPeFluor710 , streptavidin-eFluorV450 ( Ebioscience ) , anti-Bcl-2-PE , anti-CCR7-PE-Cy7 , anti-CD28-FITC , anti-CD38-PE , anti-CD8-APC H7 , anti-CD8 AmCyan ( BD Biosciences ) , anti-CD127-PE , anti-CD160-PE , anti-CD57-FITC ( Beckman Coulter ) , anti-CTLA-4-FITC ( R&D ) , anti-CD3-PerCP , anti-CD160-FITC , purified anti-CD160 ( BioLegend ) , anti-LAG-3-Atto488 ( Alexxis ) , anti-LAG-3 FITC ( LifeSpan BioSciences ) , anti-LAG-3-PE ( R&D ) . Viaprobe ( 7-AAD , BD Biosciences ) was used for dead cell exclusion . The anti-KLRG1-AlexaFluor488 and anti-KLRG1-biotin antibodies were generated as previously described [16] . Tetramer and antibody staining was performed as previously described [17] . For the multi-inhibitory receptor staining , the cells were first stained with the CD160-IgM-pure antibody , then with the anti-IgM-PerCP secondary antibody , then stained with the KLRG1-biotin antibody followed with the streptavidin-eFluorV450 antibody . Cells were then incubated with the tetramer followed by the staining with the fluorophore-conjugated antibodies ( CD8-AmCyan , 2B4-FITC , PD-1-PE , CD127-APC-eFluor780 ) . Incubation time was 15 min for each staining step and cells were washed twice in between the addition of staining reagents . Samples were acquired on a FACS Canto II flow cytometer ( BD Biosciences ) and analyzed with FlowJo v8 . 8 . 6 software ( TreeStar Inc . ) . Gates for positivity in polychromatic panels were determined by fluorescence-minus-one control stains , as recommended [38] . 2*10∧6 freshly isolated PBMCs were labelled with 40 µM Pacific Blue succinimidyl ester ( PBSE , Invitrogen ) , and stimulated with 10 µM peptide in the presence of 20 IU/ml rhIl-2 ( Roche ) in 1 ml complete medium ( RPMI 1640 containing 10% fetal calf serum , 1% streptomycin/penicillin , and 1 . 5% Hepes buffer 1 mol/L ) . Cells were incubated at 37°C in the presence or absence of 10 µg/ml anti-PD-L1 ( Ebioscience ) for one week , subsequently stained and analyzed by polychromatic flow cytometry . PD-L1 blockade-dependent gain of proliferation was determined by the ratio: frequency of HCV-specific CD8+ T cells after peptide stimulation and PD-L1 blockade divided by the frequency after peptide stimulation alone . The proliferation index was calculated by dividing the frequency of HCV-specific CD8+ T cells after stimulation by the frequency ex vivo . Statistical analysis was performed using GraphPad Prism 5 software ( GraphPad Prism Software , Inc . ) . The statistical tests used were ANOVA 1way analysis of variance followed by Newman-Keuls Multiple Comparison Test ( Fig . 3 , 4 , 7 ) , and Mann-Whitney U test ( Fig . 8 ) . | About 170 million people are infected with hepatitis C virus ( HCV ) , which may cause severe liver disease and liver cancer . Upon acute infection , only about 30% of patients are able to eliminate the virus spontaneously while about 70% of patients develop chronic infection . It is known that a successful immune response against HCV depends on virus-specific CD8+ T cells . However , during chronic infection , these cells are impaired in their antiviral function . In this study , we found that the exhaustion is characterized by the expression of multiple inhibitory receptors , such as PD-1 , 2B4 , CD160 and KLRG1 . Of note , the coexpression of these receptors depends on the ongoing recognition of the viral antigen and the maturation stage of the T cell . The remaining virus-specific T cell responses that are not exhausted do not recognize the virus present in the patients any more due to viral mutations , indicating viral escape . Thus , they fail to exert antiviral activity , although they share characteristics of fully functional memory T cells . In sum , we have found that T cell exhaustion contributes to the failure of about half of HCV-specific CD8+ T cell responses and that it is determined by a complex interplay of immunological and virological factors . These findings will be important to consider in the design of new antiviral vaccination strategies . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"immunology/antigen",
"processing",
"and",
"recognition",
"gastroenterology",
"and",
"hepatology/hepatology",
"immunology/immune",
"response",
"virology/immune",
"evasion",
"immunology/immunity",
"to",
"infections",
"infectious",
"diseases/gastrointestinal",
"infections"
] | 2010 | Coexpression of PD-1, 2B4, CD160 and KLRG1 on Exhausted HCV-Specific CD8+ T Cells Is Linked to Antigen Recognition and T Cell Differentiation |
Cytokinesis is powered by the contraction of actomyosin filaments within the newly assembled contractile ring . Microtubules are a spindle component that is essential for the induction of cytokinesis . This induction could use central spindle and/or astral microtubules to stimulate cortical contraction around the spindle equator ( equatorial stimulation ) . Alternatively , or in addition , induction could rely on astral microtubules to relax the polar cortex ( polar relaxation ) . To investigate the relationship between microtubules , cortical stiffness , and contractile ring assembly , we used different configurations of microtubules to manipulate the distribution of actin in living silkworm spermatocytes . Mechanically repositioned , noninterdigitating microtubules can induce redistribution of actin at any region of the cortex by locally excluding cortical actin filaments . This cortical flow of actin promotes regional relaxation while increasing tension elsewhere ( normally at the equatorial cortex ) . In contrast , repositioned interdigitating microtubule bundles use a novel mechanism to induce local stimulation of contractility anywhere within the cortex; at the antiparallel plus ends of central spindle microtubules , actin aggregates are rapidly assembled de novo and transported laterally to the equatorial cortex . Relaxation depends on microtubule dynamics but not on RhoA activity , whereas stimulation depends on RhoA activity but is largely independent of microtubule dynamics . We conclude that polar relaxation and equatorial stimulation mechanisms redundantly supply actin for contractile ring assembly , thus increasing the fidelity of cleavage .
The contractile ring is a dynamic , actomyosin-based structure whose constriction generates a cleavage furrow , ultimately producing two daughter cells . To ensure accurate distribution of chromosomes to each cell , the location of the division plane and the timing of cleavage must be tightly regulated to coordinate with chromosomal segregation . In animal cells , this coordination is mediated by the spindle apparatus , which both segregates the chromosomes during anaphase and signals the cortex as to where and when the contractile ring should be assembled ( reviewed in , e . g . , [1–4] ) . In addition to its structural components , the ring contains numerous other proteins , some of which are thought to recruit or retain ring components , while others function as regulatory or signaling molecules . For example , the GTPase RhoA and its guanine nucleotide exchange factor help regulate polymerization of actin and myosin within the ring , thus promoting its constriction ( reviewed in [5] ) . Some proteins are localized both to the central spindle and the equatorial cortex in a temporally dynamic manner . It is not known whether these proteins represent independent pools or whether the proteins diffuse or are somehow transported between locations . Spindle microtubules are known to play a critical role in specifying the cleavage plane in animal cells . Evidence from divergent organisms shows that the cleavage furrow can be initiated at the spindle equator by astral microtubules , central spindle microtubules , or both ( reviewed in [1] ) . Furthermore , microtubules have been shown to be the only structural constituent of the spindle apparatus required for furrow induction , regardless of their original location within the spindle [6] . However , it remains unclear whether microtubules induce furrow formation by mechanisms of polar relaxation and/or equatorial stimulation of the cell cortex ( reviewed in [1–4 , 7] ) . The polar relaxation model , as originally proposed by Wolpert [8] , postulates that inhibitory signals are sent from astral microtubules to the polar cortex to decrease cortical tension locally . Computer simulations indicate that high microtubule density at the poles may produce the highest surface tension at the equator , where the contractile ring forms [9 , 10] . In sea urchin eggs , just before cytokinesis , tensile forces become elevated globally , and thus relaxation of tension at the poles may lead to furrow formation at the equator [11 , 12] . Furthermore , in cells with depolymerized or genetically shortened microtubules , the overall cortical contractility is higher , albeit disorganized ( e . g . , [13–15] ) . In theory , microtubules could induce polar relaxation by a number of different mechanisms . Microtubules could inhibit assembly ( or promote disassembly ) of polar actin filaments . Alternatively , they could inhibit contractile activity of the polar filaments . For example , microtubules might inhibit Rho-dependent actin contractility through localized sequestration and inhibition of GEF-H1 , an activator of Rho [16] . Or instead , microtubules could induce polar relaxation by altering the distribution of cortical actin filaments , relocating them away from the poles in a process known as cortical actin flow . Cortical flow of actin has been observed in cultured mammalian cells [17] . The equatorial stimulation model proposes that furrowing is induced by stimulatory signals sent from the central spindle midzone and/or the spindle asters to the equatorial cortex . The classic “torus experiment” by Rappaport [18] demonstrated that astral microtubules from two opposing asters , not connected by a spindle , are sufficient to induce ectopic furrow formation in echinoderm embryos . Because the cell was topologically manipulated to eliminate the poles , the implication was that polar relaxation could not be invoked as the mechanism for furrowing—although technically , a role for cortical flow has not been excluded . Similarly , astral microtubules from two neighboring spindles can define such an ectopic furrow in a fused epithelial cell [19] . For monopolar spindles , it was proposed that furrow formation was stimulated by a subset of astral microtubules stabilized by chromosomes [20] , or by bundles of central spindle-like microtubules [21] . In cultured rat cells , when communication between the central spindle and equatorial cortex is disrupted by a small “perforation” ( i . e . , a deep indentation of the membrane , located peripherally within the equatorial plane ) , a furrow fails to form at the equatorial cortex [22]; the furrow initiates instead at the perforation site—as if the central spindle were now signaling to an “interior cortex” located on the spindle side of the perforation . Proteins localized to the central spindle , such as centralspindlin ( consisting of a RhoGAP and a kinesin ) and RhoGEF , are important for organizing midzone microtubules and signaling the equatorial cortex for furrow induction [5 , 23 , 24] . Recently , LET-99 and heterotrimeric G proteins were shown to play important roles in aster-positioned cytokinesis in Caenorhabditis elegans embryos [25] . In C . elegans zygotes , stimulatory signals from asters and the spindle midzone are both important for cytokinesis [26–28] . In Drosophila , it was very recently shown that F-actin–associated vesicles co-localize with central spindle and peripheral microtubules . Thus , it was hypothesized that the central spindle may somehow coordinate co-transportation of actin and new membrane to the ingressing cleavage furrow [29] . Although it has not been directly demonstrated that polar relaxation and equatorial stimulation coexist in a single cell during cytokinesis , other groups have worked to document the relationship between astral-derived and central spindle-derived cytokinetic signals . For example , in the Drosophila asterless mutant , a suboptimal version of cytokinesis can occur despite the absence of asters from the male meiotic spindle , demonstrating that astral microtubules are dispensable for inducing cytokinesis in male meiosis [30] . Bringmann and Hyman showed that astral and central spindle signals both help position the cytokinetic furrow during the first zygotic division in C . elegans [26]; however , cortical tension was not specifically addressed . The astral signal in C . elegans was later shown to be distinct from the central spindle signal and to negatively regulate distribution of cortical myosin [28] . In contrast , myosin II in cultured mammalian cells does not undergo cortical flow , as determined by kymography , whereas actin does [17] . Furthermore , the authors postulated that cortical actin is not the sole source of equatorial actin [17]; the other source was not identified . Hu et al . recently observed HeLa cells undergoing drug-induced monopolar cytokinesis . They postulated that feedback loops ( i . e . , bi-directional signaling ) between midzone-like spindle microtubules and various furrow components could promote polarization in these cells , and also that cortical flow could help organize microtubules [21] . Our work builds on , complements , and extends these ideas , using a novel experimental approach . If we assume that polar relaxation and equatorial stimulation can coexist in cytokinesis , it is not obvious how microtubules manage to provide opposing signals in the same cell , i . e . , to relax the polar cortex yet stimulate the equatorial cortex . To address this question , and to probe the mechanistic link between cortical properties and contractile ring assembly , we used fluorescently labeled cytokinetic cells of the silkworm Bombyx mori . These cells allowed us to examine the distribution of actin filaments as driven by micromanipulated microtubules , with or without drug treatments that affect either microtubule dynamics or actin assembly . We show that loss of cortical actin may occur at any region of the cell cortex adjacent to dynamic microtubules , due to cortical flow . Stimulation , as gauged by rapid de novo assembly and delivery of actin aggregates , may also occur at any region of the cell cortex , mediated by the bundled , overlapping plus ends of central spindle microtubules . Our data show that polar relaxation and equatorial stimulation mechanisms coexist in silkworm spermatocytes and are compatible with dual roles for microtubules . Furthermore , our work provides mechanistic advances and demonstrates how both mechanisms can independently supply actin for contractile ring assembly .
Spermatocytes cultured underneath a layer of halocarbon oil were relatively large ( ∼33 μm in diameter ) , optically clear , and amenable to micromanipulation . Digital-enhanced polarization microscopy on cells from late metaphase to early anaphase revealed a short but robust spindle ( Figure 1A and Video S1; n = 9 ) , whose relatively small size provided ample space for manipulation of spindle microtubules . To visualize the dynamics of microtubules and actin filaments during cytokinesis , we microinjected metaphase-stage spermatocytes with rhodamine tubulin and low-level Alexa Fluor 488 phalloidin ( Figure 1B and Video S2; n = 4 ) . The fluorescently labeled cytoskeleton was imaged using spinning disc confocal microscopy . Furrow initiation occurred during early anaphase after actin aggregates emerged at the interdigitating microtubule plus ends at the equator ( Figure 1B , between 0 and 4 ) . The aggregates gradually enlarged before coalescing with cortical actin filaments to assemble the contractile ring that constricts the cell ( Figure 1B , 4–45 ) . The time from anaphase onset to cleavage was not significantly different for the uninjected control cells ( 48 . 4 ± 2 . 5 min ) and the injected , double-labeled cells ( 49 . 5 ± 5 . 0 min ) , suggesting that the low-level phallodin was not perturbing actin function . Image stacks of fixed , immunostained cells revealed dense arrays of astral microtubules radiating from the centrosomes to the polar cortex ( Figure 1C , a–c; n = 15 ) . Notably , asters in both fixed ( Figure 1C , a–c ) and live ( Figure 1C , d and e; n = 15; also in Figure 1B , 0 ) cells appeared to be attached only loosely to the spindle , which is a natural phenomenon in silkworm spermatocytes [31] . Presumably due to the centrioles' motile flagellar axonemes ( Figure 1C , a–c; as seen in immunostained cells ) , both asters were mobile within the cell . In approximately 15% of the cells , the asters even moved to the same side or the same pole of the spindle , as documented in metaphase ( Figure 1C , a and d ) and anaphase ( Figure 1C , c and e ) stage cells . This natural phenomenon makes silkworm spermatocytes an ideal system for separating the role of asters from that of the central spindle in cytokinesis . We wanted to determine whether de novo assembly of actin aggregates at the plus ends of central spindle microtubules requires RhoA as a regulator . To do so , we microinjected late metaphase cells with C3 ribosyltransferase , a botulinum toxin that specifically inhibits Rho by ADP ribosylation [33] . We hypothesized that this inhibition would disrupt de novo actin assembly at the spindle midzone , without affecting cortical flow of actin . As expected , the C3 transferase-treated cell continued to divide but failed to accumulate actin aggregates in the central spindle ( Figure 7 ) . As seen in optical sections of control cells ( n = 15 ) , actin accumulated both at the equatorial cortex ( Figure 7A , top and bottom ) and at microtubule plus ends within the central spindle ( Figure 7A , middle ) . In contrast , C3 transferase-treated cells ( n = 9 ) accumulated actin at the equatorial cortex ( Figure 7B , top and bottom ) , but not within the central spindle ( Figure 7B , middle ) . This result implies that RhoA inactivation inhibits actin assembly at the plus ends of central spindle microtubules ( i . e . , the initial step of equatorial stimulation ) , but does not inhibit relocation of cortical actin from the poles to the equatorial cortex ( i . e . , polar relaxation ) . Additional tests confirmed that in C3 transferase-treated cells , actin filaments were still excluded from the cortex adjacent to the dislocated spindle and incorporated into the contractile ring ( Figure 7C , n = 8 ) . However , when late metaphase cells were microinjected with C3 transferase , and subsequently treated with paclitaxel at anaphase onset , both pathways for actin redistribution were affected ( Figure 7D , n = 5 ) ; specifically , delivery to the equatorial cortex from both the polar cortex and the midzone was impeded . Cells failed to divide long after entering anaphase ( Figure 7D , 0–73 min ) , due to inhibition of both de novo assembly of nascent actin aggregates , and relocation of cortical actin filaments to the contractile ring .
By mechanically manipulating the spindles of double-labeled silkworm spermatocytes , we show how polar relaxation ( as gauged by redistribution of cortical actin ) and equatorial stimulation of the cortex ( as gauged by de novo actin assembly and transport ) both contribute to contractile ring assembly and furrow induction . We demonstrate that spindle microtubules deliver both inhibitory and stimulatory signals to the cell cortex during furrow formation , depending on their configuration and location within the cell . The dynamic astral microtubules drive actin filaments from the polar region to the equatorial cortex , which relaxes the polar cortex , while increasing cortical tension at the actin-dense equator . Meanwhile , bundles of interdigitating microtubules across the entire central spindle stimulate de novo assembly of actin aggregates at their overlapping plus ends , and deliver the aggregates to the equatorial cortex . These dual signaling mechanisms ensure that both cortical actin and central spindle-associated actin are delivered to the equatorial cortex , resulting in a robust contractile ring , and providing redundancy to safeguard an essential cellular process . Videos S4 and S9 , as well as previous studies [34 , 35] , provide hints that cytoplasmic actin filaments ( perhaps vesicle bound; [29] ) may constitute a third source of actin that could ultimately be delivered to the equatorial cortex by microtubules . We show that in the absence of equatorial stimulation , contractile ring assembly and furrowing can still be induced by means of actin redistribution driven by cortical flow ( Figure 3A and 3B and Figure 7C ) . We accomplished this by displacing the entire spindle apparatus , bearing “stimulatory” furrow cues that were reorganized ( via spindle collapse ) , or C3 transferase-suppressed , to an arbitrary region of the cortex . The existence of a cortical flow of actin filaments away from the displaced spindle microtubules was established by our ability to intercept the flow locally , using a microneedle ( Figure 3D ) . We found that the flow of cortical actin during cytokinesis can be induced by any microtubules positioned near the cortex , provided their plus ends are dynamic and accessible ( i . e . , neither interdigitating nor otherwise bundled or stabilized ) . We hypothesize that collapsing the spindle can mechanically free a subset of plus ends; this could allow those formerly interdigitating central spindle microtubules to become more dynamic and to take on properties held by astral microtubules , such as the ability to contact the cell cortex and induce cortical flow of actin filaments . Interdigitating microtubules in an intact central spindle would be able to contact the cortex only laterally , and as incorrectly oriented bundles , which should not be conducive for promoting cortical flow . As predicted by this finding , furrow formation can occur between a collapsed asterless spindle and its two detached asters in a remodeled cell ( Figure 3C ) . We have also shown that this microtubule-driven actin flow begins shortly after anaphase onset and persists through telophase ( Figure 3A and 3B ) , contributing to both contractile ring assembly at the equator and furrow inhibition at the poles throughout the furrow induction and ingression stages . Cortical flow during cytokinesis has previously been inferred or observed for contractile elements including myosin II [28 , 36 , 37] and pre-existing actin filaments [17 , 35 , 38] , as well as for a membrane-bound receptor-ligand complex , membrane domains , and cell surface proteins [39–41] . As the movement of certain surface proteins mirrors that of cortical actin , it has been suggested that the actin network could serve as a scaffold for these proteins , thus engineering their co-transport to the furrow region [41] . How might microtubules interact with actin filaments to effect their exclusion from the poles ? Since paclitaxel inhibits cortical flow ( Figure 6B–6D ) , actin filaments are likely to be driven from the polar region towards the equatorial cortex by elongating microtubules . Following anaphase onset , the plus ends of astral microtubules can elongate toward and contact the cortex [20 , 42] . Such microtubules would be anchored to the spindle pole at their minus ends , and conceivably constrained at their plus ends by indirect attachment to a mobile cortical component , namely filamentous actin ( analogous to cortical capture via Bim1/Kar9/Myo2 [43 , 44] ) . Given the microtubule's length , flexibility , geometrical orientation within the cell , and constraints on its direction of movement , its elongation may be sufficient to drive cortical actin toward the equator . Elongation of tethered microtubules against a fixed barrier can produce force of up to 10 pN , as gauged by microtubule buckling [45] . Presumably , this force could be harnessed to allow tip-attached cortical actin to “hitchhike” to the equatorial cortex , where it would be released , possibly by binding to myosin II [38] or anillin [46] . Alternatively , or in addition , actin filaments could be transported along the subset of microtubules that are contiguous to the cortex , and longitudinally aligned , with appropriate polarity . Motor proteins would provide the driving force for transport , with microtubule elongation serving only to extend the track toward the equator . Actin filaments could either be cargo for a plus-end–directed kinesin ( attached via an adaptor protein ) , or else the actin could be coupled to the microtubule track via kinesin/myosinV-type hetero-motors ( as modeled in [44 , 47] ) , and slide along the stationary microtubule toward the equatorial cortex , with kinesin providing directionality for the flow . The tracks could consist of the cortically apposed section of centrosome-attached astral microtubules and/or those released from the centrosomes [42 , 48–50] that become cortically apposed [34 , 47 , 51 , 52] . We also show that actin aggregates can be assembled de novo within the central spindle in the absence of cortical-flow based redistribution of actin , by using paclitaxel to impede the flow . Similarly , paclitaxel-stabilized microtubules can induce a cleavage furrow in mammalian cells [53] , perhaps implying mechanistic conservation . We demonstrate that cortical stimulation can be induced at any region in the cortex by microtubule plus ends of the central spindle ( Figure 6 ) . Our data reveal that de novo–assembled cytoplasmic actin aggregates can comprise a supply of actin sufficient to form the contractile ring and induce cell cleavage ( Figure 6D ) . Newly assembled actin aggregates are precursors of the contractile ring , since they ultimately merge with the ring ( Figure 5 ) . The existence of de novo assembly of actin during contractile ring formation is consistent with evidence from diverse cell types , including grasshopper spermatocytes ( unpublished data ) , yeast [54 , 55] , Xenopus eggs [56] , and mammalian cells [17 , 38]; in addition , there is precedent for the presence of actin nucleating protein formin at microtubule plus ends [57] . Assembly of actin aggregates is sensitive to the Rho GTPase inhibitor , C3 transferase ( Figure 6B , middle ) , supporting the existence of a microtubule-dependent zone of active RhoA during cleavage plane specification [5 , 23 , 58–62] . Our findings also raise the possibility that equatorial stimulation signals could be transported along stationary central spindle microtubule tracks from the spindle pole region to the midzone ( Figure 5B ) , before ultimately reaching the equatorial cortex , perhaps via lateral transport . Actin aggregates are assembled across the entire spindle midzone ( Figure 5A and 5C ) ; hence they must be transported laterally from the plus ends of the bundled microtubules to the equatorial cortex . We have captured this novel means of delivery using time-lapse sequences ( Figures 5B , 6E , and 6F ) . Bundles of central spindle microtubules , some tipped with actin aggregates , splay laterally outward as the spindle changes shape , thereby conveying their cargo to the incipient furrow ( Figure 5B ) . The actin cargo remains associated with the tips of the microtubule bundles until the bundles are within about 0 . 4 μm of the cortex , less than the length of the aggregate itself ( Table 1 ) . At this point , the microtubule-associated aggregates appear to be released from the microtubule bundles , in that the costaining at the microtubule tips disappears . The dissociation of the actin from the bundles may be mediated by kinesins bound to the bundle , as the aggregate appears to slide off the end of the bundle as it approaches the cortex ( Table 1 ) . This transport of actin aggregates may be largely independent of microtubule dynamics; when a post-anaphase , paclitaxel-stabilized spindle with mechanically exposed , bundled plus ends is brought close to the cortex by micromanipulation , it remains functional for delivery of actin aggregates to the cortex ( Figure 6E and 6F ) . The induction of an ectopic furrow following cortical deposition of actin aggregates from microtubule plus ends ( Figure 6D ) indicates that actin-tipped microtubules may contain sufficient furrow constituents ( i . e . , structural constituents along with any co-transported stimulatory cues ) to permit contractile ring assembly . This idea is consistent with work by Hu et al . , showing that the tips of monopolar spindles contain critical stimulatory components ( e . g . , the Rho GEF Ect2 ) [21] . We propose that in the bipolar spindle , other spatially dynamic midzone components may also utilize this microtubule-based lateral transport mechanism—either co-transported with actin aggregates , or else delivered via a distinct set of midzone microtubules . By ensuring timely relocation of cytokinetic components , lateral transport could play a role in processes such as midzone signaling of the equatorial cortex , assembly and maintenance of the contractile ring and abscission . For example , the midzone component centralspindlin becomes localized to the equatorial cortex during anaphase . Bundled interdigitating microtubules , tipped with centralspindlin , may become docked at the contractile ring via Anillin [46 , 63] . We suggest that these tipped microtubules may originate from the central spindle and deliver their cargo to the cortex via lateral transport . Similarly , lateral transport could provide a means of delivery to the cortex for a recently characterized subset of vesicles . These vesicles are associated both with actin and with the central spindle , and are thought to contribute to furrow ingression [29] . In fixed cells , the authors saw thin microtubule bundles potentially connecting the central spindle with the periphery [29] , consistent with a lateral transport mechanism . As another example , by early anaphase , the chromosomal passenger complex ( CPC ) is at the mitotic spindle midzone , whereas later in anaphase it is also present at the equatorial cortex ( reviewed in [64] ) , perhaps arriving there by lateral transport . Interestingly , in Dictyostelium , the CPC component INCENP has been shown to bind actin [65] . It is also worth determining whether a subset of RhoA may reach the equatorial cortex via lateral transport , given that RhoA localization may depend on microtubule organization [58] . Our results explain how a single spindle component , the microtubule , can play opposing roles in two complementary pathways of contractile ring assembly . The seemingly incongruous stimulatory and inhibitory effects on the cortex are produced by different configurations of microtubules that populate different regions of the cytokinetic cell . We propose a microtubule induction model to describe how microtubules perform their dual signaling at the poles and the equator to induce contractile ring formation and cleavage furrow initiation . During early anaphase , dynamic astral microtubules at the poles inhibit ectopic furrow formation by excluding pre-existing actin filaments from the poles ( Figure 8A , red dotted arrows and red lines at cortex ) , ultimately resulting in actin accumulation at the incipient equatorial furrow . Meanwhile , at the plus ends of the relatively stable central spindle microtubules , actin aggregates are assembled de novo , and enlarge over time ( Figure 8A , red lines in midzone ) . As the cell progresses toward telophase , preexisting cortical actin filaments continue to flow from the polar cortex to the equator ( Figure 8B , red dotted arrows ) . At the same time , the plus ends of bundled central spindle microtubules splay laterally to deliver the actin aggregates to the equatorial cortex ( Figure 8B , red solid arrows ) . The newly delivered aggregates coalesce with the actin excluded from the polar cortex to assemble the contractile ring ( Figure 8C ) . In summary , we show that actin is redistributed into the contractile ring via two microtubule-dependent pathways that coexist during cytokinesis of the silkworm spermatocyte: equatorial stimulation and polar relaxation . These dual signaling pathways ensure the fidelity of cytokinesis , which fails only if both mechanisms are inhibited ( Figure 7D ) . With some modification , our model could be generalized to multiple cell types . For example , cytokinetic cells of varying sizes and spindle geometries might rely more heavily on one mechanism than the other to achieve a global balance between tension and relaxation . We have demonstrated that the entire cell cortex is potentially responsive to both stimulatory and inhibitory cues from microtubules during cytokinesis in silkworm spermatocytes . How might the cell restrict the contractile ring to a localized band of equatorial cortex ? By stimulating only the equatorial cortex while concomitantly relaxing other regions of the cortex , a cell could ensure cytokinetic fidelity by building in redundancy . Polar relaxation could be thought to include a component of equatorial stimulation in that pre-existing actin filaments are transported to the equatorial cortex . Equatorial stimulation , with its de novo actin assembly , recruits cytoplasmic G-actin to the equator . Hence , the seemingly disparate models ultimately result in redistribution of all available actin resources to the equator for ring assembly . Our work with spermatocytes of the silkworm B . mori has enabled us to document a number of novel events essential for understanding contractile ring assembly . These spermatocytes are highly amenable to mechanical manipulations and imaging techniques , making them ideal for cytokinetic studies . Equally important , the silkworm genome has been sequenced [66] , paving the way for molecular and genetic studies , such as RNAi inhibition of particular genes . Thus , we believe the silkworm spermatocyte has great potential as a model system for research in cell biology , cell physiology , and developmental biology—given its innate attributes coupled with the power of molecular genetics .
The spermatocytes of silkworm Bombyx mori were isolated from the testes of a 5th instar larva and spread under inert halocarbon oil ( 400 oil , Halocarbon Products ) to form a monolayer of cells in a glass chamber slide . Only cells in the first meiotic division were used . Cells were observed with an inverted Zeiss Axiovert 135 microscope modified for both digital-enhanced polarization [6] and spinning disc confocal microscopy ( CARV , Carl Zeiss ) . The microscope is equipped with a 1 . 4 NA achromatic-aplanatic condenser and a 1 . 45 NA/100X α-Plan objective ( Carl Zeiss ) . An EM-CCD digital camera ( Hamamatsu C9100–12 ) , Simple PCI software ( C-image ) , and Photoshop software ( Adobe Systems ) were used to record and process images . All images in a given category “n” are fully representative of all unshown data in that category . Micromanipulation needles were pulled from glass tubing ( outer diameter: 1 . 0 mm; inner diameter: 0 . 58 mm , World Precision Instruments ) as described [67] using a microforge ( Narishige , Model MF-830 ) and maneuvered with a Burleigh MIS-5000 series piezoelectric micromanipulator [6] . Glass tubing ( outer diameter: 1 . 0 mm; inner diameter: 0 . 75 mm ) with an internal capillary ( World Precision Instruments ) was pulled on a Flaming/Brown P-87 micropipette puller ( Sutter Instrument Company ) to produce micropipettes with a tip diameter ∼ 0 . 1 μm . The injectant was back loaded into the micropipette using a Hamilton syringe . Microinjection was conducted at 60 psi using a custom-built pneumatic injector maneuvered with a Burleigh MIS-5000 series piezoelectric micromanipulator . Alexa Fluor 568 tubulin . Alexa Fluor 568 ( Invitrogen ) was conjugated to porcine tubulin ( Cytoskeleton ) following the protocol adapted from Peloquin et al . [68] . The labeled tubulin was resuspended in an injection buffer ( 20 mM PIPES , 0 . 25 mM MgSO4 , 1 mM EGTA , 1 mM GTP , pH 6 . 8 ) to a final concentration of 3 mg/ml for microinjection . Phalloidin . Alexa Fluor 488 phalloidin or rhodamine phalloidin ( Invitrogen ) in methanol was concentrated using a SpeedVac concentrator ( Savant ) and resuspended in the injection buffer to a final concentration of ∼6 . 6 μM for microinjection . Although the intracellular concentration of phalloidin was low , controls were performed to ensure that the phalloidin did not adversely affect the function of the actin filaments . The timing of cell division was not slowed down in a phalloidin-treated cell relative to uninjected controls ( as in Table S1 ) . C3 transferase . C3 transferase ( Cytoskeleton ) was stored as 1 mg/ml aliquots at −70 °C in a buffer containing 500 mM imidazole ( pH 7 . 5 ) , 50 mM Tris HCl ( pH 7 . 5 ) , 1 . 0 mM MgCl2 , 200 mM NaCl , 5% sucrose , and 1% dextran . Immediately before microinjection , the C3 transferase was mixed with Alexa Fluor 568 tubulin and Alexa Fluor 488 phalloidin to a final concentration of 0 . 5 mg/ml in the micropipette . Tubulin Tracker ( Invitrogen ) was used as directed , with the Oregon Green 488 Taxol ( paclitaxel ) diluted to 50 μM in insect Ringer's solution . The diluted Tubulin Tracker was micropipetted around the target cells , resulting in a further dilution of at least 1000 fold . Hoechst 33342 ( Invitrogen ) was stored as 10 mg/ml stock aliquots at −20 °C and diluted in Tubulin Tracker buffer to 0 . 5 mg/ml before micropipetting . Silkworm spermatocytes were fixed and stained as previously described [6] . Microtubules were stained with tubulin primary antibody ( Chemicon ) and Alexa Fluor 488 conjugated secondary antibody ( Invitrogen ) ; actin filaments , with rhodamine phalloidin ( Invitrogen ) ; and chromosomes , with DAPI . In all figures , microtubules are shown in green , and actin in red ( false colored , if necessary ) . Chromosomes , when labeled , are blue ( Hoechst for live cells , DAPI for fixed ) . The degree of stiffness at one region of the spermatocyte cortex relative to another region was assayed as follows . The spindle of a cell in anaphase was collapsed ( by pushing the spindle poles together ) and mechanically repositioned near the cortex to induce cortical flow of actin filaments . After waiting 5–10 min to allow for redistribution of actin , the first region of interest was probed with the side of a flexible microneedle , tangential to the cortex . Depending on the flexural rigidity of the needle and the local cortical stiffness , a given needle would either deform the cortex by flattening it , without deflecting the needle , or the cortex would bend the needle . The cell was repositioned ( by rotating the microscope slide by 180° ) to bring the second region of interest in proximity with the needle , and then reprobed . | In animal cells , the last step of cell division , or cytokinesis , requires the action of a contractile ring—composed largely of actin and myosin filaments—that cleaves the cell in two . Before the cell divides , it first duplicates its genome and separates the chromosomes into the two newly forming daughter cells , a task carried out by a structure called the spindle apparatus , which is composed mostly of long polymers called microtubules . The site of cleavage must occur between the segregating chromosomes—at the spindle equator—to ensure that each cell receives the proper number of chromosomes . In addition to separating the chromosomes , microtubules are also essential for inducing cytokinesis—but how they do this is controversial . For example , the “polar relaxation” hypothesis proposes that the astral microtubules , which radiate outward , cause contractile elements to flow from the polar cortex toward the equator , resulting in furrowing . In contrast , the “equatorial stimulation” hypothesis proposes that the spindle microtubules directly stimulate cleavage exclusively at the equator . Using a novel approach , we demonstrate that both mechanisms are in fact functioning together to recruit actin filaments to the nascent ring , providing redundancy that increases fidelity . Specifically , we were able to mechanically alter the distribution of actin filaments in living , dividing cells by using a microscopic needle to manipulate microtubules while perturbing the cytoskeleton with chemical compounds . Our high-resolution microscopy data advance the understanding of both proposed mechanisms . We also documented a novel , microtubule-based mechanism for transporting actin aggregates to the equatorial cortex . These results help to resolve a long-standing dispute concerning this fundamental cellular process . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"cell",
"biology"
] | 2008 | Redundant Mechanisms Recruit Actin into the Contractile Ring in Silkworm Spermatocytes |
The bacterium Neisseria meningitidis is commonly found harmlessly colonising the mucosal surfaces of the human nasopharynx . Occasionally strains can invade host tissues causing septicaemia and meningitis , making the bacterium a major cause of morbidity and mortality in both the developed and developing world . The species is known to be diverse in many ways , as a product of its natural transformability and of a range of recombination and mutation-based systems . Previous work on pathogenic Neisseria has identified several mechanisms for the generation of diversity of surface structures , including phase variation based on slippage-like mechanisms and sequence conversion of expressed genes using information from silent loci . Comparison of the genome sequences of two N . meningitidis strains , serogroup B MC58 and serogroup A Z2491 , suggested further mechanisms of variation , including C-terminal exchange in specific genes and enhanced localised recombination and variation related to repeat arrays . We have sequenced the genome of N . meningitidis strain FAM18 , a representative of the ST-11/ET-37 complex , providing the first genome sequence for the disease-causing serogroup C meningococci; it has 1 , 976 predicted genes , of which 60 do not have orthologues in the previously sequenced serogroup A or B strains . Through genome comparison with Z2491 and MC58 we have further characterised specific mechanisms of genetic variation in N . meningitidis , describing specialised loci for generation of cell surface protein variants and measuring the association between noncoding repeat arrays and sequence variation in flanking genes . Here we provide a detailed view of novel genetic diversification mechanisms in N . meningitidis . Our analysis provides evidence for the hypothesis that the noncoding repeat arrays in neisserial genomes ( neisserial intergenic mosaic elements ) provide a crucial mechanism for the generation of surface antigen variants . Such variation will have an impact on the interaction with the host tissues , and understanding these mechanisms is important to aid our understanding of the intimate and complex relationship between the human nasopharynx and the meningococcus .
N . meningitidis ( the meningococcus ) colonizes the nonciliated columnar mucosal cells of the human nasopharynx as a harmless commensal organism and , as such , is carried by five to ten percent of the adult population [1 , 2] . Some strains are able to cross the mucosa into the bloodstream from where they can cause septicaemia or meningitis and , as a result , are a major cause of disease worldwide [2] . Several genetic loci have been associated with disease [3 , 4] , but for most strains the mechanism of virulence is not well defined . The close interaction with the human host is reflected in enriched diversity and variability at the bacterial cell surface . There are 12 different polysaccharide capsules , which are the basis of serogrouping , some of which are virulence determinants [5–7] . Vaccines targeted to the capsule types most commonly associated with disease have been successful , though capsule switching is a cause of concern [8] . Many meningococcal surface-exposed proteins and carbohydrates are also highly variable , creating a major challenge in the development of a universal meningococcal vaccine [9 , 10] . Current models of bacterial populations describe a spectrum of structures ranging from clonal , where lineages are derived from a common ancestor and horizontal genetic exchange plays no role , to nonclonal ( or panmictic ) , where rates of horizontal genetic exchange are so high that genetic differences between isolates are effectively randomised and individual genetic lineages are undetectable [11] . Extremes are rare with many bacteria having a semiclonal structure where horizontal exchange is common , but groups of clonally related bacteria exist . Multilocus sequence typing has played a major role in defining bacterial population structure and shows N . meningitidis to have a fundamentally nonclonal population due to the natural competence and high rates of recombination that characterise the species [12–14] . However , multilocus sequence typing is able to resolve N . meningitidis into groups of related sequence types known as clonal complexes , and studies have shown that while there is enormous diversity in the population as a whole there are relatively few lineages associated with the ability to cause disease [15 , 16] . Most disease causing strains belong to serogroups A , B , or C , but it is clear that membership of one of the hyperinvasive lineages is equally predictive of the ability to cause disease . The genomes we analyse here represent three ( of around ten ) such disease associated lineages ( Table 1 ) , and it is hoped that comparative genomics will help to unravel the paradox of devastating virulence in an organism that relies on asymptomatic carriage and person-to-person transmission for its proliferation . The genome sequences of N . meningitidis strains Z2491 [17] and MC58 [18] , which belong to serogroups A and B respectively , have been reported previously and allowed the initial identification of both known and potentially novel mechanisms for variation of surface structures , including the transfer of coding information from silent gene cassettes , phase variation through slippage-like mechanisms , local recombination , and the presence of arrays of short noncoding repeats throughout the chromosome . These repeat arrays were postulated to increase the variability of the associated genes through enhanced recombination with externally acquired DNA [17] . Here we report the genome sequence of N . meningitidis serogroup C strain FAM18 , a medically important representative of the ET-37/ST-11 complex , which has been a major cause of meningococcal disease worldwide throughout the last century [19] and , despite low carriage rates , continues to be associated with sporadic outbreaks [20–23] . Strain FAM18 was isolated from the cerebrospinal fluid of a child suffering from meningitis by Dr . Janne Cannon and colleagues in North Carolina in the 1980s and remains capable of infection . The genomes of these three strains have been incorporated into pan-Neisseria DNA microarrays and used in three separate comparative genomic hybridisation studies to compare the gene contents of a range of isolates including N . gonorrhoeae strains , invasive and carriage strains of N . meningitidis , and commensal species of Neisseria [24–26] . These studies highlighted differential acquisition of islands between strains , evidence of horizontal DNA transfer between N . meningitidis and N . lactamica , and the potential to define virulence-specific gene sets . Here , we have compared the FAM18 genome data with those of strains Z2491 and MC58 , to specifically focus upon local sequence divergence and provide evidence for mechanisms whereby recombination of exogenous DNA with specific chromosomal loci is promoted to generate variation in cell surface antigens .
The three sequenced N . meningitidis genomes are largely colinear with only three apparent reciprocal inversions around the origin of replication ( Figure 1 ) . Each genome appears to represent one of these inversions relative to the other two , suggesting that these inversions may have occurred once in each lineage . However , the foci of the inversions show a high degree of variability relative to the strong synteny across the chromosome , suggesting that they may be subject to frequent additional inter- and intra-genomic recombination . The inversion event closest to the origin of replication ( IE1; 3′-adjacent to NMA0220/NMB0050/NMC0034 ) seems to be due to recombination between repeat arrays in Z2491 . Interestingly , one of the arrays flanks a pilin gene , pilC2 , while the other is adjacent to genes involved in pilus retraction ( pilTU ) . Thus , in MC58 and FAM18 , pilC2 is adjacent to pilTU , while in Z2491 they are distant . Neisserial PilC proteins are important components of the type IV pilus machinery involved in adhesion to host cells , promotion of piliation , and transformation competence [27 , 28] . The PilT protein is essential for pilus retraction [29] , and it has been shown that PilC1 regulates PilT-mediated pilus fibre retraction [28] . Although there is no direct published evidence for coregulated transcription of pilTU and pilC2 , it seems plausible that this rearrangement may have an effect on pilus phenotype . The FAM18 pilus regulon also differs from Z2491 and MC58 because of the deletion of much of the pilE/S locus and the insertion of the class II pilin-encoding pilE2 ( see below ) . Further variability at IE1 can be seen with the insertion of a copy of a meningococcal disease associated ( MDA ) island in the repeat array ( see below ) directly upstream of pilC2 in FAM18 . The MDA island encodes a filamentous bacteriophage that is secreted via the type IV pilus and is specifically associated with strains that have the potential to cause disease [30] . The second inversion , IE2 ( 5′-adjacent to NMA2200/ NMB0287/NMC0293 ) , is probably due to recombination between copies of IS1106 in FAM18 and is also associated with the insertion of a locus encoding a putative restriction-modification system in FAM18 . Restriction-modification systems coordinate the recognition and destruction of “non-self” DNA from sources lacking the same system and for N . meningitidis have been associated with specific lineages [31] . The third , IE3 , is the most complex of the three inversion events and seems to be due to recombination between loci encoding a large repetitive surface protein and its associated secretion system ( NMA0688 , NMB0497 , NMB1779 , and NMC0444; see below ) . These loci appear to encode two-partner , or type V , secretion systems [32] and are similar in sequence and genetic arrangement to those of Bordetella species where fhaC and fhaB , respectively , encode a secretion accessory protein and a filamentous haemagglutinin important in virulence [33] . Z2491 and FAM18 have a single copy of this locus , while MC58 has two copies that are approximately equidistant from the origin of replication and are the foci of the rearrangement . Prior to duplication the locus has also acquired a novel set of two-partner secretion protein genes and an MDA-related prophage . Duplication of the whole locus and subsequent recombination involving another MDA island may have lead to the current genomic arrangement , which would appear to have benefited MC58 with a greater potential variety of surface protein expression . All three of the reciprocal inversion foci that we have described here seem to affect genetic loci with the potential to modulate interaction with the host and/or other strains of N . meningitidis . Despite frequent inter-strain recombination , N . meningitidis genomes maintain a high level of colinearity , so it may be the case that the rearrangements observed in this three-genome comparison have added significance . The predicted amino acid sequences of the coding sequences ( CDS ) from each of the genome annotations were compared by three-way reciprocal Fasta analysis to assess the numbers of orthologous and unique CDS . The latter were defined as CDS where a reciprocal match was not detected in either of the other two translated genome sequences . Visualisation and manual curation of the results of this analysis using the Artemis Comparison Tool revealed limitations of the test . This analysis methodology did not take into account the relative chromosomal position of the genes , so the best matches between genes of the different genomes could be those that are in different chromosomal contexts and , therefore , likely to be paralogues ( genes of similar sequence in the same genome ) rather than true orthologues . Examples of characteristic features in N . meningitidis that confound the reciprocal match test include CDS within loci encoding variable surface proteins such as adhesins or haemagglutinins . In some cases multiple paralogous loci exist within each genome and may be exchanging DNA by intra- and/or inter-genomic recombination . The result is that syntenic loci ( those in the same position ) are equally diverged from one another as they are from nonsyntenic loci ( see below ) . For convenience we have designated such genes as “variable” to distinguish them from simple orthologues . Paralogous CDS at nonsyntenic loci are also designated as variable . Variable genes tend to occur in clusters , and there is a clear correlation between these gene clusters and regions of low % G + C content . Viewed on a whole genome scale , eight of the nine most prominent GC troughs across the genome of FAM18 coincide with variable loci with the one exception being the ribosomal protein operon ( NMC0129–NMC0159 ) ( Figure S1 ) . It was observed in Helicobacter pylori that genes that are not universally present across a number of strains , and are therefore likely to be laterally acquired , tend to have a lower than average GC content [34] , and a similar bias has been seen in related enteric genomes [35] . It has been suggested that accessory genes ( those variably present in different strains within a species ) may be subject to different selective pressures to the core genes , and that low % G + C content is one of the results of this difference [36] . It is therefore possible that the low % G + C nature of the variable genes in N . meningitidis may be a consequence of selection for exchange within the species . In addition , the three meningococcal genome sequences were compared using ACEDB , as described previously [37] , to identify unique coding sequences , regardless of their annotation in their respective genomes . The results of these two analysis methodologies were combined and , following manual curation , 240 unique genes were identified; 83 ( 4 . 1% ) in Z2491 , 97 ( 4 . 8% ) in MC58 , and 60 ( 3 . 0% ) in FAM18 . Table 2 summarizes the types and numbers of regions containing unique genes and Table S1 details the individual CDS functional annotations . The majority encode hypothetical proteins of unknown function . This is to be expected , because strain-specific genes are generally poorly studied , and largely do not form part of the common and core metabolic functions that have been most studied , and are most readily identifiable through comparison with other well-studied species and biochemical pathways . There are some unique restriction-modification systems and these would be expected to have an impact on the uptake of DNA from N . menigitidis strains in the same niche; such systems have previously been shown to be associated with different lineages [5 , 38 , 39] . The majority ( 39 of 56 ) of the unique gene clusters contain three or fewer consecutive genes , and 30 of these ( 68 genes ) correspond to known or candidate Minimal Mobile Elements [40] with alternative unique loci present at syntenic locations across the three genomes . With the exception of dam in FAM18 , all of the unique restriction-modification system genes are within MMEpheST or MMErfaDclpA . Larger unique regions are often associated with insertion sequence elements ( nine of 56 clusters; 53 genes; Table S1 ) and with a Mu-like prophage ( pnm2 ) present at the same location in all three genomes ( between; NMA1280 and NMA1323 , NMB1077 and NMB1112 , NMC1041 and NMC1056 ) . The IS-associated unique CDS are often small and lie in low % G + C troughs . They are also mostly annotated as “hypothetical proteins” with little information available to allow prediction of the effect of their differential presence . MC58 carries the largest version of prophage pnm2 , which includes CDS-encoding cell surface antigens able to induce bacteriocidal antibodies in mice [41] . The presence of unique genes in each genome within these prophage could be due to independent phage insertions , differential gene loss from a larger prophage inserted in a common ancestor , or intergenomic recombination between prophage . Z2491 contains a large unique region of 63 annotated genes ( NMA1821–NMA1885 ) , which constitutes a Mu-like prophage ( pnm1 ) shown to be conserved among epidemic serogroup A strains [42] , though an association with virulence has not been demonstrated . FAM18 contains a region ( NMC0852–NMC0895 , IHT-E ) that includes genes homologous to lambdoid bacteriophage genes and a transposon carrying a type I secretion system [26] . As with other members of the species , the N . meningitidis FAM18 genome contains many hundreds of repetitive sequence elements ranging from simple sequence repeats associated with phase variable genes ( see below ) , to complete gene cluster duplications ( Table 3 ) . DNA uptake sequences ( 5′-GCCGTCTGAA-3′ ) are the most abundant repeats and are distributed throughout the genome [43] . Concordant with their % G + C-rich sequence , they are less frequent in low % G + C regions , which often coincide with important genetic loci including those for ribosomal proteins , capsule biosynthesis , pilus biosynthesis , Maf adhesins , prophage , Iga protease , cytolysin transport , and RTX-family exoproteins . The next most abundant repeat types are the “neisserial intergenic mosaic elements” ( NIMEs ) , which consist of 20-bp inverted repeats ( ATTCCCNNNNNNNNGGGAAT , dRS3 elements ) flanking over 100 families of ∼50–150-bp repeat sequences ( RS elements ) [17] . Also frequent are the “Correia repeat enclosed elements” ( known as CREE or Correia elements ) , which comprise a conserved repeat sequence ( 156 bp full length or 51 bp internal deletant ) bounded by a 51-bp inverted repeat . CREEs are often located upstream of genes [44] , have been shown to affect gene expression [45 , 46] , and may be transposable or mobilisable [47] . The numbers of each major repeat type are comparable in the three complete N . meningitidis genomes ( Table 3 ) . Comparison of repeat elements between the three genomes revealed no repeat types unique to one genome though it did identify RS element diversity . For example , repeat sequence clustering analysis for Z2491 and FAM18 showed that of the 611 RS elements in FAM18 , there are 80 FAM18-specific versions that group into 27 subfamilies , suggesting novel repeat development , possibly generated by recombination . NIMEs are often clustered into long arrays of multiple dRS3s separated by different RS elements . These arrays may also contain other repeats such as CREEs and insertion sequence elements , which may be opportunistic insertions . We have previously suggested that these NIME arrays may encourage sequence variation in neighbouring genes by increasing the frequency of recombination with exogenous DNA , and thus exchange of adjacent sequences , either by acting as substrates for homologous recombination , or as targets for a specific recombinase [17] . The chromosomal position of these repeat arrays is generally consistent between the three genomes suggesting that they were initially introduced in a common ancestor . However , comparison of syntenic repeat arrays reveals considerable differences in repeat number and array length , indicating that the arrays themselves are dynamic ( Figure 2 ) , as would be expected if they were substrates for recurring recombination . To study the correlation between repeat arrays and coding sequence divergence , the three pairwise genome comparisons were combined to measure amino acid identities between orthologous CDS . This showed that the average percentage identity between orthologous CDS flanking repeat arrays is significantly ( p-value = 5 . 2 × 10−6 using a single-tailed t-test ) lower than the average percentage identity of orthologues not flanking repeat arrays , supporting the hypothesis that the arrays are associated with increased diversity in flanking genes . Despite this strong association , other measures of the diversifying affect of repeat arrays are less clear cut . The relationship between array length and flanking gene diversity is displayed in Figure 3A and shows that for positionally orthologous genes immediately adjacent to repeat arrays there is a tendency for sequence identity to decrease as array length increases and that the same pattern is seen in all three pairwise comparisons . The low combined R2 value indicates that the association with repeat length is not strong . However , a strong correlation with array length may be obscured since array size is dynamic ( Figure 2 ) and may itself change rapidly . We further analysed the orthologue sequence identities to test whether the diversifying affect of the repeat arrays could be detected beyond the immediate adjacent CDS . Figure 3B plots the orthologue identities for the CDS within ten genes of an array and shows a weak association between othologue diversity and distance from the array . Figure 3C shows that this effect is almost entirely due to the diversity seen in the CDS immediately adjacent to arrays . The distance to which the diversifying effect of the arrays extends should be limited by the length of DNA fragment that can be recombined through a double crossover event where one crossover is within the repeat array . A study focused on allele replacement of the tbpB gene ( Figure 2 ) encoding an immunogenic cell surface protein in N . meningitidis found that the recombining DNA fragments ranged from 1 . 5 to 9 . 9 kb with a median size of 5 . 1 kb [48] . This size range would suggest that recombination fragments could extend several CDS away from a repeat array , although the effect of this is not seen in our comparisons . Interestingly , of the 19 fragments mapped to tbpB allele replacements , 13 had at least one end point within or very close to a flanking NIME repeat array [48] , so while immune selection may be driving the retention of variants , the majority of the required recombination events are associated with NIME repeat arrays . Based on the above findings , we hypothesise that the relative positions of CDS and repeat arrays are under selective pressure such that genes where increased variation is beneficial are more likely to be associated with arrays . The repeat arrays serve to promote recombination with exogenously acquired DNA , increasing the rate of gene exchange at the adjacent loci . Although this does not directly cause increased variation in these genes , it should enhance the exchange of variants , and therefore increase the apparent rate of variation . This correlates with the pattern seen in Figure 3A , where some genes are highly variable in some comparisons but not in others—exchange is a stochastic process—sometimes a highly variant gene will have been acquired , and sometimes a gene more similar to the comparators used here . Clearly any exchange will only be fixed if it is selected for , and this may explain why variation from the array is not retained . In support of this hypothesis , there is a clear association between repeat arrays and CDS encoding cell surface associated proteins where increased sequence variation may be an advantage in host interactions ( Figure 4; Table S2 ) . At any given repeat array size range , the majority of flanking genes code for cell surface or exported proteins . Moreover , the proportion of this class of genes increases as array size increases . A clear example of the diversifying effect of repeat arrays is the tbpAB locus ( Figure 2 ) , which lies at a syntenic position in all three genomes flanked by large repeat arrays . The dynamic nature of the repeat arrays is reflected in a wide variation in array size and content . The tbpA and tbpB DNA sequence identities ( 74 . 1%–94 . 9% and 50 . 8%–77 . 8% , respectively ) are far below the typical genome figure of >90% and the flanking gene identities of 96 . 1%–98 . 5% . Clearly , sequence divergence is focused on tbpAB leaving the surrounding genes largely unaffected further implying that selection for advantageous variation is operating . Analogous arrangements can be seen for the lbpAB locus , encoding a lactoferrin-binding protein and the porA locus , encoding a major outer membrane protein , and the pilC1/pilC2 loci . Another consideration is that there may be a reciprocal relationship between the genes undergoing repeated recombination to generate antigenically variable mosaics and the flanking repeats . Since there is little or no selective pressure for accurate recombination within the flanking repeat regions , the recombination within these regions is likely to be more “error prone” than that within the coding regions . So , the recombination of these genes may serve to create variation and growth of the flanking repeats , which in turn may favour further recombination within the adjacent coding regions . The NIME arrays themselves display a striking and regular wave profile for % G + C content with troughs corresponding to RS elements and peaks corresponding to dRS3 . Figure 5 shows one example but the profile is apparent for all large regular NIME arrays in the genome . The profile is sometimes less obvious in smaller arrays or less regular arrays containing other repeat types such as CREEs , but the % G + C profile is often even detectable for isolated dRS3-RS-dRS3 units . The profile does not appear to be simply due to the high % G + C nature of the dRS3s , but is rather a combination with the lower % G + C RS elements . Average % G + C values for FAM18 are 55 . 78% for dRS3s , 45 . 18% for RS elements , and 51 . 62% for the whole chromosome . Closer analysis shows considerable internal variation for the NIME subunits with dRS3 % G + C ranging from 35% to 70% and RS elements ranging from 28% to 76% . The large overlap between these ranges implies that to maintain the wave profile the dRS3 and RS element sequences are somehow interdependent . Since dRS3s are only 20 bp long and have a conserved 6-bp terminal inverted repeat , their % G + C variation is limited to the central eight bases . RS elements are essentially defined as regions flanked by dRSs , so may not necessarily be functional units and could be largely inert , or have a structural role . Under this definition , RS elements vary in size from 19 to 214 bp and form multiple sequence families some of which tend to have short ( 5-bp ) terminal inverted repeats . Within RS elements % G + C content is often “patterned” with either a central low % G + C trough or a side-to-side slope . Maintenance of the % G + C wave profile in NIME arrays suggests that the structure may be related to function , potentially participating in , or promoting , recombinational events . We hypothesise that dRS3 sequences within NIME arrays are binding sites for a site-specific recombinase that enhances recombination between these sequences and exogenously acquired DNA containing other dRS3 elements , thereby promoting variation at a number of genes associated with NIME arrays . If the dRS3-mediated recombination formed an initial cross-over event , then the insertion of linear DNA could be completed by RecA-mediated homologous recombination in the adjacent sequences , ensuring replacement with similar genes . Alternatively , pairs of arrays surrounding genes could both participate in dRS3-mediated recombination , or array-flanked genes on acquired DNA could be inserted into chromosomal arrays . Continued recombination between chromosomal and acquired dRS3 elements , with the functionally selected consequence of exchanging adjacent genes , could have the effect of building up repeat arrays containing “spacer” regions ( the RS elements ) with specific physical or conformational properties . Bille et al . [30] and Kawai et al . [49] have recently described a type of neisserial filamentous prophage whose presence in meningococcal genomes is associated with the ability to invade host tissues . They have also showed that these bacteriophage integrate into dRS3 repeats by the action of a phage-encoded transposase/recombinase . This protein is therefore a plausible candidate for the specific recombinase predicted by our hypothesis . This phage is a member of a larger family of neisserial phage , and it is therefore reasonable to suppose that this recombinase has been present in the neisserial genome for some time . N . meningitidis genomes contain several loci where transcriptionally silent gene cassettes can be used as sources of variation for expressed surface structures and proteins . Comparison of such variable loci from different strains reveals detail of different genetic arrangements and may be useful for understanding the mechanisms for generation of variants . The best-described example is the pilin-encoding pilE/S system where the expressed pilin ( PilE ) can be altered by incorporation into the pilE CDS of DNA from 5′-adjacent promoter-less pilS genes [50] . Much of the pilE/S locus has been deleted in FAM18 , which is associated with the previously recognized insertion , and conversion to the sole expression , of a class II pilin-encoding gene ( pilE2 ) elsewhere on the chromosome ( which is not present in Z2491 and MC58 ) [51] . Variation of the pilE gene using pilS sequences has been extensively studied [50 , 52 , 53] , the efficiency of which has been shown in N . gonorrhoeae to involve a short DNA sequence ( the Sma/Cla repeat ) located downstream of the pilE gene . In N . meningitidis the silent pilS loci are embedded within NIME arrays , and it is possible that the specific dRS3-mediated recombination postulated above may contribute to generating silent variation within pilS sequences . A different mechanism of variation appears to exist for several loci encoding putative haemagluttinins ( fhaB ) and adhesins ( mafB ) . Downstream of these genes are what appear to be silent cassettes encoding alternative C termini for the encoded proteins . These cassettes contain short repeats that are identical to sequences only present within the upstream genes . We have previously suggested that these repeats could be the substrates for direct recombination , replacing the 5′ end of the gene [17] . The three-way comparison provides more evidence in support of this view , including examples where the C terminus of one of the expressed genes in one genome is identical to a silent cassette in the same locus in another genome . The maf loci are generally comprised of tandem mafA and mafB genes , both of which are thought to encode adhesins , followed by a number of putative silent cassettes , and many genes of unknown function ( Figure 6A ) . Each of the three N . meningitidis genomes has three maf loci designated here as maf1 , maf2 , and maf3 ( based on their order relative to the origin of replication in the FAM18 genome sequence ) . The loci are at syntenic positions but occur in a different order within each chromosome due to the chromosomal inversion events described above ( Figure 1 ) . For all three genomes the maf2 and maf3 loci start with a mafA gene , which is absent from maf1 . However , maf1 from MC58 ( maf1_MC58 ) and FAM18 ( maf1_FAM18 ) have a truncated mafA mid way through the locus , and maf1_Z2491 has been largely deleted . There is also a possible mafA remnant in the middle of maf3_FAM18 . Clearly there is considerable variation encoded within the maf loci , which would be expected to manifest as variations in adhesin structure at the cell surface . Although all three maf loci have a similar structure and appear to be encoding a similar product , at the sequence level maf1 and maf2 are more similar to each other , with maf3 having only localised similarity . DNA identities between maf1 and maf2 mafA genes are high ( >97% ) , but their identities to maf3 mafA genes are much lower ( ∼65% ) . Identities between maf3 mafA genes are greater than 98% . An analogous situation exists for the mafB sequences . The encoded maf3 MafAs have an N-terminal extension relative to the others but all appear to have an intact signal sequence and are likely to be exported . At 41 . 2% , the average % G + C content of maf loci is markedly lower than the genome average of 51 . 5% . Furthermore , there is a distinct profile of % G + C content across maf loci . Generally mafA and mafB CDS , including the downstream alternative CDS , correspond to % G + C peaks with some mafA CDS as high as 59% GC . Although the intervening % G + C troughs have been annotated with potential CDS , they have no similarity to genes in the database , and their role is unclear . It is notable that they do not contain any of the repeats associated with repeat arrays . FAM18 and Z2491 have single syntenic fha loci , while MC58 has two that are associated with a genome inversion event ( IE3 ) as described above ( Figure 1 ) . These loci are analogous to the fha loci in Bordetella pertussis with an upstream gene ( fhaC ) encoding a two-partner secretion system outer membrane transporter [54] , though the B . pertussis loci do not have downstream silent cassettes . The N . meningitidis fha loci have similar structure to the maf loci with silent cassettes directly downstream from a gene encoding a large low complexity surface protein and a distinct % G + C profile ( Figure 6B ) . The size of the repeats shared by the silent cassettes and the functional gene is comparable for both fha and maf . DNA identities for N . meningitidis fhaC and the 5′ portion of fhaB are greater than 98% . The degree of similarity between fhaB genes is variable with the two copies in MC58 having a high level of identity over almost the full length , while for Z2491 and FAM18 gene identity is largely confined to the 5′ end with the remainder either unique or having similarity in the downstream “silent” region . The maf and fha loci show considerable potential for generation of multiple versions of the expressed coding sequence and , together with surface structures such as pilus , capsule , and other surface proteins , are likely to be major contributors to cell surface diversity . The presence of multiple syntenic maf loci is striking and suggests an important role . Previously , a number of potential phase variable genes have been identified based on the presence of potentially slippage-prone short repetitive sequences , and these lists have been progressively refined , through analysis of neisserial genome sequences , first in N . meningitidis strains MC58 [18] and Z2491 [17] and then in comparative studies using both of these and N . gonorrhoeae strain FA1090 [37] , and subsequently in a study of the commonly used experimental N . gonorrhoeae strains [55] and a partial study of N . meningitidis [56] . Based upon these studies , and those published by others on specific genes , and a four-way comparison using the N . meningitidis FAM18 genome sequence , a revised and updated phase variable gene list is presented here ( Table S3 ) . There are now 24 known phase variable genes in the Neisseria spp . and a further 25 strong candidates , counting members of established gene families such as Opa proteins only once . Over half of these encode surface proteins , enzymes that modify surface proteins , or are LPS biosynthesis proteins . This mechanism therefore has a vast capacity to vary the surface-exposed structures and epitopes of N . meningitidis . Based upon the comparisons of the three meningococcal genomes , we further characterized a number of known and putative mechanisms for the generation of diversity within and between strains of this highly adaptable and variable species . Many of these mechanisms involve random variation , which is locally increased due to the presence of repeats , either generating local instability or serving as substrates for homologous recombination , resulting in altered expression of specific genes ( phase variation ) or generating allelic diversity within particular surface proteins ( pilE , mafB/fhaB families , NIME array-associated genes ) While phase variation through homopolymeric tracts has been noted in several other genera , the NIME arrays and cassette-mediated variation described here seem to be specific to Neisseria and may be important characterising features of the genus . NIME arrays are found at syntenic positions in the genomes of N . gonorrhoeae and N . lactamica ( http://www . sanger . ac . uk/Projects/N_lactamica ) but have not been observed in non-neisserial genomes . Although two-partner secretion systems analogous to the fha locus are found in other bacterial genomes [33] , they only include the two essential components and lack the silent cassettes that enhance variability in N . meningitidis . Notably , the N . gonorrhoeae and N . lactamica genomes both have three maf loci syntenic with those in N . meningitidis , N . gonorrhoeae has two extra maf loci , and neither have fha loci . These genomic differences will affect the cell surface and may relate to niche differences and interactions with the host . Variation of the bacterial cell surface is a common theme in host–pathogen interactions and appears to be important for colonisation of new niches and avoidance of the immune system . Such variation may be even more important for commensal organisms , such as Neisseria , that remain associated with their hosts for long periods . The genome of FAM18 , and its comparison with MC58 and Z2491 , highlights N . meningitidis as a paradigm of genomic variability linking a combination of DNA uptake and recombination , minimal mobile elements , intergenic repeat arrays , and phase variation to generate and maintain phenotypic diversity focused at the cell surface .
N . meningitidis strain FAM18 genomic DNA was prepared as previously described [57] . An approximately 8×-shotgun sequence was produced from a total of 68 , 352 end-sequences from pUC clones with 1 . 4–2 . 0kb inserts using the Big Dye Terminator Cycle Sequencing kit from Applied Biosystems ( http://www . appliedbiosystems . com ) . Reactions were run on Applied Biosystems 3700 sequencers . An approximately 1× coverage was produced from 1 , 152 end sequences from 10–20 kb inserts cloned into pBACe3 . 6 and used to scaffold contigs and bridge repeat sequences . The sequence was finished to standard criteria [17] . Sequence assembly , visualisation , and finishing were performed using PHRAP ( P . Green , unpublished data; http://www . phrap . org ) and Gap4 [58] . Putative orthologues were identified by reciprocal-best-match FASTA searches between the meningococcal strains Z2491 , MC58 , and FAM18 protein sequences with cutoffs of 80% sequence length and 30% identity . The orthologue list was manually curated and annotation was transferred for orthologues common to strains Z2491 and FAM18 . All other genes were annotated using standard criteria [17] , and the complete genome annotation was then manually curated in Artemis [59] . The strain Z2491 EMBL entry has also been resubmitted to reflect the annotation updates generated during this study and the rotation of the sequence to place the origin of replication at the start . Genome comparisons were visualised using the Artemis Comparison Tool [60] . Repeats were defined and annotated using a combination of BLAST [61] and HMMer [62] . In an independent analysis , the complete genome sequences of N . meningitidis strains FAM18 , MC58 , and Z2491 and N . gonorrhoeae strain FA1090 were analysed using ACEDB ( R . Durbin , J . T . Thierry-Mieg , unpublished data , http://www . acedb . org ) as described previously [37 , 55 , 63 , 64] . Perfect sequence repeats characteristic of phase variable genes were identified using ARRAYFINDER [65] . Repeats , the annotations from all four neisserial genome sequences , and other sequence features were displayed in their sequence context within ACEDB . Analysis of the potential for simple sequence repeats to generate transcriptional or translational phase variation was determined through analysis of the repeat in the sequence context , as has been done previously [37 , 55 , 63 , 64] . Unique genes were identified as those for which no homology was displayed , the display parameters within ACEDB being set to 1e−50 for DNA identity , and 1e−4 for amino acid similarity . In cases of large paralogous gene families , genes that displayed low homology to only a portion of the gene with an annotated feature from another genome sequence were considered in their wider chromosomal context to determine if the allele is unique or divergent . The results of these independent analyses were combined and curated .
The EMBL Nucleotide Sequence Database ( http://www . ebi . ac . uk/embl ) accession numbers for the genomes discussed in this paper are N . gonorrhoeae ( AE004969 ) , N . meningitidis strain FAM18 ( AM421808 ) , and N . meningitidis strain Z2491 ( AL157959 ) . | Human surface tissues , including the skin and gut lining , are host to many different species of bacteria . N . meningitidis is a species of bacteria that is only found in humans where it is able to colonise mucosal surfaces of the nasopharynx ( nose and throat ) . This association is normally harmless and at any one time around 15% of the population are carriers . Some strains of N . meningitidis can cause disease by invading the host tissue leading to septicaemia or meningitis . We aim to gain understanding of the mechanisms by which these bacteria cause disease by studying and comparing genomes from different strains . Here we describe specific genes and associated repetitive DNA sequences that are involved in variation of the bacterial cell surface . The repeat sequences encourage the swapping of genes that code for variant copies of cell surface proteins . The resulting variation of the bacterial cell surface appears to be important in the close interaction between host and bacteria and the potential for disease . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] | [
"infectious",
"diseases",
"genetics",
"and",
"genomics",
"microbiology",
"eubacteria"
] | 2007 | Meningococcal Genetic Variation Mechanisms Viewed through Comparative Analysis of Serogroup C Strain FAM18 |
Mounting evidence suggests that glycans , rather than merely serving as a “shield” , contribute critically to antigenicity of the HIV envelope ( Env ) glycoprotein , representing critical antigenic determinants for many broadly neutralizing antibodies ( bNAbs ) . While many studies have focused on defining the role of individual glycans or groups of proximal glycans in bNAb binding , little is known about the effects of changes in the overall glycan landscape in modulating antibody access and Env antigenicity . Here we developed a systems glycobiology approach to reverse engineer the complexity of HIV glycan heterogeneity to guide antigenicity-based de novo glycoprotein design . bNAb binding was assessed against a panel of 94 recombinant gp120 monomers exhibiting defined glycan site occupancies . Using a Bayesian machine learning algorithm , bNAb-specific glycan footprints were identified and used to design antigens that selectively alter bNAb antigenicity as a proof-of concept . Our approach provides a new design strategy to predictively modulate antigenicity via the alteration of glycan topography , thereby focusing the humoral immune response on sites of viral vulnerability for HIV .
Env glycoproteins on the surface of enveloped viruses , such as HIV [1–4] , Dengue [5 , 6] , Ebola [7] , hepatitis C [8] , influenza [9] , Lassa [10] , and Zika [6 , 11] , are the primary vaccine targets for the induction of protective , broadly neutralizing antibodies ( bNAbs ) . However , many of these viruses evade the evolution and activity of bNAbs via sequence diversification and the masking of critical Env epitopes by glycosylation . Various molecular engineering approaches have been applied to generate HIV immunogens , such as stablizing a closed conformation of a native like-trimeric Env [12–17] or creating minimal target sites of neutralizing vulnerability on nanoparticluate structures [18] . These efforts have successfully elicited autologous neutralizing antibodies ( Abs ) in rabbits as well as in macaques [19 , 20] and have been shown to guide the first steps of germline bNAb precursor activation [21–26] . Unfortunately , these immunogens have yet to prove sufficient for driving the evolution of broadly cross-neutralizing antibody ( Ab ) responses [24] , indicating that new immunogen engineering strategies are urgently required to improve antigenic profiles of Env immunogens for selective generation of Abs against sites of neutralizing vulnerability . Glycans represent more than half the mass of the HIV Env glycoprotein , obscuring nearly the entire surface of the Env trimer . While these glycans were originally believed to shield against an Ab response , over the past decade a number of bNAbs have been identified that actively recognize these glycan themselves . Interestingly , these antibodies usually emerge following extensive evolutionary selection enabling them to generate unusual antigen-recognition domains ( Fabs ) that are able to reach through , and even utilize glycans , to access the underlying protein surface [27–36] . Moreover , through high-resolution imaging approaches , including cryo-electron microscopy ( Cryo-EM ) , it is becoming apparent that bNAb:glycan interactions are common across nearly all bNAb classes . Through linked viral evolution studies in subjects who evolve bNAbs , it is clear that while some glycans are essentially part of the Ab-epitope , other glycans actively block binding , thereby giving rise to a complex network of potential agonists/antagonists [3 , 4 , 37 , 38] . Antigenic opportunities for manipulating glycans on Env were first elucidated in studies focused on either altering the overall glycans or certain glycans specifically [39–43] . These approaches resulted in enhanced antigenicity , as well as the evolution of Abs to sites of neutralizing Ab vulnerability [44] . Removal of particular glycans in the C-terminus of the V2 loop enabled the induction of neutralizing Abs with some breadth in non-human primates ( NHPs ) [45 , 46] . Similarly , the induction of autologous tier 2 neutralization following immunization with native-like SOSIP trimeric Env was reproducibly attributable to induction of nAbs against a “glycan hole” in the shield [47] . Elimination of the glycan located in loop D of the CD4-binding site resulted in enhanced germline B cell activation of VRC01 and NH45-46 precursors [44 , 48] . Moreover , recent reports point to the acquisition or loss of particular glycans in driving affinity maturation of specific bNAb lineages [49 , 50] . However , while certain glycans may help shape individual epitopes [30 , 31 , 33 , 34 , 36 , 51] , glycans interact with one another , dynamically reshaping the exposed protein surface , coordinately influencing Ab binding access [37 , 52–56] . Collectively , these data all highlight the critical role of glycans , as individuals or as groups , contributing to the overall antigenic profile of the Env glycoprotein . Nonetheless , there remains inadequate understanding of systematic principles that could offer an actionable path for the development of immunogens that can exploit the large mass of sugars decorating viral protein surfaces . While causal relationships between individual glycosylation sites , or groups thereof , and bNAb binding have been defined [27 , 30–36 , 38 , 44–46 , 49–56] , it is conceivable that modulating glycan interactions more broadly may represent a further means to focus the humoral immune response [57–59] . The Env glycans extend beyond the protein surface and accordingly yield a topographical landscape of the macromolecule , influencing access to the underlying protein . Changes in glycosylation at one site , such as addition or removal of a particular glycan , may have a substantial impact on epitope availability even at relatively distal sites [37 , 56] . Therefore , methods that can exploit both the protein and the glycans from a multi-site , landscape perspective , may provide a benefit for improved HIV immunogen design . In this work we describe a multi-variate , combined experimental/computational glyco-engineering strategy to rationally alter HIV envelope glycosylation with a goal of inducing desirable topographical changes in epitope accessibility . First , a deconvolution model was constructed to reverse-engineer the data in a panel of 94 recombinant HIV Env antigenic profiles screened against a battery of broadly neutralizing and non-neutralizing Abs . Glycan occupancy at each n-linked glycan site ( ‘sequon’ ) was assessed by mass spectrometry , and both protein sequence and glycan occupancy were used as variables in the model to elucidate how Ab binding is dependent on them . Next , based on this model , forward-engineering principles aimed at changing the overall antigenic profile of any given HIV Env were ascertained . Using a Bayesian machine learning algorithm permitted determination of key sequon sites that positively or negatively influence Ab binding . Finally , these results were then used in a de novo , in silico protein antigen design model using an iterative antigenicity-guided sequence evolution framework , seeking to predict alterations in the antigen that could selectively improve or impair target antibody binding . These predictions were then used to synthesize novel antigens that successfully bound to target bNAbs with enhanced and selective antigenicity . All together , our work provides proof-of-concept , in an initial HIV Env application , for the advance of multi-variate glycosite-based engineering of immunogens that can focus the humoral immune response on sites of neutralizing vulnerability .
Over the past decade , significant effort has been invested in the generation of stabilized , native , trimeric HIV Env proteins or minimal scaffolds of sites targeted by bNAbs , aimed at focusing the immune response to sites vulnerable to antibody mediated neutralization . While exciting results continue to accrue , these proteins alone have not proven sufficient to drive neutralizing Ab breadth . Strategies to further optimize the HIV Env glycoprotein for induction of bNAb breadth thus continue to be urgently needed . Given our emerging appreciation for the antigenic nature of the large array of glycans that decorate the HIV Env glycoprotein , here we hypothesized that the glycan shield itself might be strategically modified and exploited in a rational manner , offering a potentially generalizable glyco-engineering approach that could be applied to either trimeric or monomeric Env proteins . For proof-of-concept purposes , we focused our study on a diverse panel of Env monomers . Because of the remarkable sequence diversity among HIV viral variants , we assessed glycan site occupancy across a panel of recombinant HIV Env monomers ( gp120 ) . Traditional assays for HIV Env glycosylation have relied on enumeration of N-linked glycosylation sites ( sequons; Asn-X-Ser/Thr , X ≠ Pro ) , within a sequence ( Figure A in S1 File ) . However , accumulating data suggests that the presence of a sequon is necessary but not sufficient to guarantee the presence of the associated glycan [60–64] . To address this discrepancy , we first conducted a mass spectrometric N-linked glycoproteomic analysis of the glycan occupancy profiles for each of 94 recombinant gp120 proteins that span clades A , B , and C with additional AE , AG , and BC variants ( Fig 1 , Figure B in S1 File ) . Site-specific , N-linked glycan occupancy was determined using proteolytic digestion , peptide deglycosylation , occupied site labeling , and MS/MS fragment quantification and site-mapping . The glycan occupancy level was estimated by calculating the ratio of the spectral counts of the glycosylated peptide in relation to the overall incidence of the peptide , where the glycosylated peptide was quantified through 18O isotopic labeling [65–67] . To validate linkage of the spectral count to the relative peptide abundance , we tested spectral counting of a pre-mixed peptide population ( Table A in S1 File ) ; this peptide mixture contained two nearly identical sequences: one with a particular sequon and another with a Asn-to-Asp substitution within that sequon , to represent glycosylated and non-glycosylated-peptide variants . The ratio of the spectral counts of the two peptides provided consistent quantitation of the peptide relative distributions , giving confidence in this method . From this N-glycoproteome analysis , 83% of the possible sequons across the panel of 94 gp120 variants were identified as quantitatively present , with 92% of the sequons exhibiting full or partial occupancy ( Table B in S1 File ) . The glycan occupancy profiles strikingly revealed heterogeneity at nearly all sites ( Fig 1A , S2 Table ) , and the likelihood of a site being occupied was independent of its sequon frequency ( Fig 1B and 1C , Figure C in S1 File ) . For example , for some highly-conserved sequon sites ( e . g . , N156 , N197 , N241 , N301 ) only moderate glycan occupancy was observed , while other highly-conserved sequon sites demonstrated nearly complete occupancy across the 94 proteins ( e . g . , N88 , N234 , N262 ) . Given the significant variation in viral sequences among the tested gp120s ( Figure A in S1 File ) , the relationship between glycan occupancy and sequon sequence was evaluated to determine whether protein sequence alone could predict sequon occupancy . Classic sequons ( N-G-S/T ) showed the highest degree of glycan occupancy , whereas N-E-S/T exhibited the lowest glycan occupancy ( Figure D in S1 File ) and a number of sequons demonstrated intermediate occupancy profiles . We also inspected glycan occupancy variance across all the sites ( Fig 1D ) and mapped the sites with the top 15% highest variance into a Env trimer structure . Interestingly , sites with high glycan occupancy variance were enriched within known bNAb epitopes ( Fig 1E ) , except for the CD4 binding site epitope . This result indicates that the observed heterogeneity over our 94 proteins is not simply related to unusual glycosylation on monomeric gp120 , but rather points to the possibility that HIV may actively exploit variation in both protein sequence and glycan occupancy as a means to escape bNAb detection selectively . In turn , sequons , with reproducibly variable glycosylation , provide an opportunity to engineer glycoproteins . Previous efforts to define the role of glycan occupancy in shaping antibody activity have focused largely on the impact of removing or adding one or a few glycans . We continued our multi-variate approach by using ELISA to interrogate the impact of glycans on Ab binding , for 13 bNAbs and 3 non-NAbs ( F105 , A32 , 447-52D ) ( Figure E in S1 File ) against the 94 gp120s for which glycan occupancy profiling was obtained ( Fig 1A ) . These Abs covered four well-defined epitope clusters: N332 & V3 glycan ( PGT121–128 ) , N332 & outer domain glycan ( PGT135-136 , 2G12 ) , V3 loop ( 447-52D ) , and CD4 binding site-like ( VRC01 , B12 , F105 , A32 ) [12 , 68] . While all Abs bound to clade B variants more effectively , the various Abs demonstrated unique binding profiles across clade A and C and recombinant variants ( Fig 2A ) . Despite the observed binding profile heterogeneity , clusters of binding profiles were observed across the proteins , pointing to clear epitope families , with clusters of binding including PGT125–128 , PGT121–124 , PGT135–136 , and the CD4 binding site–targeting Abs ( B12 , VRC01 , F105 , A32 ) . Thus , related-monoclonals bound preferentially to similar sets of gp120s , as expected based on epitope availability of target epitopes on particular gp120 structures ( Fig 2B ) . Based on the disparate Ab binding profiles ( Fig 2A ) and the known glycan occupancy differences ( Fig 1A ) , we ascertained the impact of glycan site occupancy on Ab binding . The correlation matrix , which depicts the pair-wise covariance between glycan site occupancy and Ab binding across the gp120 panel , illustrates a strong relationship between glycan occupancy and Ab binding profiles ( Fig 2C ) . A striking bimodal pattern was observed , where glycan occupancy at specific sites either agonized or antagonized Ab binding . Moreover , intriguing complex patterns also emerged from these data , including the observation of “super antagonist” sites ( e . g . , N149 and N139 ) and “super agonist” sites ( e . g . , N332 and N362; Fig 2C , highlighted in black squares ) that affected binding of all tested Abs , potentially acting through glycoprotein stability rather than direct antigenicity . Additionally , sites exhibiting epitope-specific modulatory effects were found ( Fig 2C , highlighted in red/green/pink squares ) . Notably , while the N332 is known for its critical role in shaping V3-family binding , it is not known for its role in shaping other bNAb responses . However , the data here suggest that the N332 glycan plays a much broader role in bNAb binding , including shaping the epitope of all CD4–binding site Abs . This observation indicates that proximal as well as distal glycans can contribute to antigenicity , potentially via alterations in the topographical remodeling of the overall glycan shield . Thus , groups of both positively-influencing and negatively-influencing glycans were associated with individual classes of Abs , pointing to glycans that either selectively decrease binding to non-NAbs ( red boxes ) along with glycans that either increase or decrease binding to target bNAbs ( PGT121 and PGT128 –in green and pink , respectively ) . These data provide new insights into the directional influence ( i . e . , agonistic or antagonistic ) of individual glycans as well as clusters of glycans in shaping Ab binding to HIV gp120 . To more rigorously model the integrative contribution of proximal and distal glycans in collectively shaping Ab binding profiles , and to define the minimal glycans required to shape the overall glycoprotein topology , we developed a supervised Bayesian machine-learning algorithm to capture relationships—linear as well as non-linear—between site occupancy profiles and Ab binding fingerprints . This computational framework consisted of two key parts . First , a support vector regression ( SVR ) model [69] involving incorporated kernels , to uncover relationships between glycan sites ( individually and in combination ) and Ab binding . And , second , a Bayesian Markov Chain Monte Carlo ( MCMC ) sampler [70 , 71] , to approximate how the various sites and combinations of glycans impact binding and ultimately capture the minimal set of optimal glycan sites that predict binding . As a proof of concept , we applied this approach to defining the minimal combination of glycan predictors for PGT121 binding . We built models based on 4 different types of input variables: sequons alone ( Figure A in S1 File ) ; protein sequences alone; glycan occupancies alone ( Fig 1A ) ; protein sequences and glycan occupancies inputs . 10-fold cross-validation was performed for each model to test robustness and prediction accuracy . The models based on sequons or sequences alone were correlated relatively poorly to the experimental ELISA binding results , with r = 0 . 53 and 0 . 75 , respectively ( Fig 3A and 3B ) . The model based on glycan occupancy alone was better correlated with binding activity ( r = 0 . 81 , p < 0 . 001; Fig 3C ) , suggesting that glycan occupancy was more predictive that sequence alone in predicting antibody binding . However , the fourth model , combining both glycan occupancy and protein sequence , showed the best performance with robust predictive power ( r = 0 . 94 , p < 0 . 001 ) and a very low MSE ( 0 . 17 ) ( Fig 3D ) , confirming that combination of glycan occupancy and sequence information can substantively account for the complex interaction between Abs and Env proteins . Given the superior improvement in model performance when both glycan occupancy and sequence were included , we next constructed models for additional Abs , incorporating directionality of glycan on antibody binding ( i . e . , positive or negative influence on binding ) to deconvolute and define the specific glycosylation sites ( Fig 3E and Figure F in S1 File , q < 0 . 01 ) and amino acid residues ( S3 Table ) that most strongly contribute to binding . Hierarchical clustering consistently showed that Abs recognizing related epitopes share similar glycan-utilization profiles ( Fig 3E ) . Glycans previously appreciated to be involved in shaping antibody binding , including N332 , were clearly found to be important across all the PGT Abs along with 2G12 , but were not among the most critical predictors of binding for models for the CD4bs-dependent Abs . Importantly , critical glycans for all Abs were observed across the entire protein sequence , indicating that glycans proximal as well as distal to the Ab binding site contribute to Ab binding profiles . Nonetheless , it must be considered that our models , based on the limited 94-monomer dataset , may not completely rule out additional sites . To gain deeper understanding of how proximal and distal glycans influence Ab binding , the linear glycan sites and protein residues were projected onto a 3-dimensional structure of gp120 ( Fig 3F , Figure G in S1 File ) . Using PGT121 , PGT128 , and VRC01 as examples , both positively-influencing and negatively-influencing glycan sites were found within the PGT121/128 binding site but not within the VRC01 binding site , as expected . However , both categories of glycan sites were also observed outside the immediate Ab binding region for all Abs . Interestingly , critical protein residues defined by the model were located both within and outside the VRC01 epitope but were largely scattered outside the PGT121/128 binding sites . These findings highlight broad distribution of influential glycans and amino acids across the surface of the HIV Env protein beyond those within the Ab binding site itself; we imagine that this influence likely transpires through alterations in the overall topography of the viral protein surface . Finally , projection of PGT122-specific glycan and amino acid determinants on the co-crystal structure of the PGT122 Fab and BG505 SOSIP gp140 Env trimer [4] further illustrates how these sites coordinately impact Ab binding ( Fig 3G ) . Sites facilitating PGT122 recognition comprised glycans that participate directly in Ab binding ( N332 and N295 ) , proximal glycans located on the nearby V1/V2 loops ( N130 , N138 , N143/147 and N188 ) , and distal glycans located on the outer domain of gp120 ( N262 , N362 , N401 , N410 and N462 ) that may reduce conformational flexibility of glycan clustering and maintain epitope accessibility [37 , 53] . Conversely , antagonistic glycan sites were mostly located immediately around the binding site ( N139/141 , N149 , N186/187 , N334 and N407/408 ) . This arrangement could reshape protein topography to limit epitope access or increase local steric hindrance preventing further glycan processing into the correct glycan structures [53] , in either case interfering with PGT122 binding . Additionally , key amino acid determinant , residue 323 , was identified within the epitope [4] , whereas others were scattered across the protein , highlighting both the proximal and distal influence of both glycans and amino acid changes in shaping the protein surface . Thus , as the majority of crystal structure studies have focused mainly on proximal glycans that contribute directly to Ab access , the importance of distal inhibitory or agonizing glycans may also be exploited to optimize immunogen design . With an appreciation for the contribution of both proximal and distal glycans in shaping Ab binding , we hypothesized that glycan occupancy signatures could be exploited to guide the design of gp120 proteins with improved binding properties for specific bNAbs . We employed an optimization algorithm for this purpose , based on dynamic evolution based on glycan occupancy , using the same Bayesian machine learning core as above . This algorithm iteratively evolved a given gp120 sequence in silico by perturbing defined individual glycan sites ( mutating one glycan site at a time ) , aimed at driving enhanced or impaired bNAb binding for a given bNAb ( Fig 4A ) . This serial computational mutagenesis approach sampled the antigenicity landscape of any given combination of glycan occupancy profiles , ultimately generating a particular combination of glycans that could enhance or impair target bNAb binding most effectively . During iterative glycan site perturbations , the model assumes that the changes ( e . g . , knock-in or knock-out ) of target sites do not influence glycan occupancy of other sites . This approach was applied with the goal of generating immunogens optimized for PGT121 and/or PGT128 binding , due to their related but distinct antigenic fingerprint profiles ( Fig 2A ) . The clade A MF535 . W0M . ENV . D11 gp120 sequence was used as a starting sequence , because of its negligible binding profile to both bNAbs ( Fig 2A , Figure E in S1 File ) . Antigens were first designed to exhibit enhanced binding to one of these two bNAbs , based on the addition of bNAb-specific agonistic glycan sites and removal of bNAb-specific antagonistic glycan sites ( Fig 4B and 4C , +PGT121 and +PGT128 ) . The engineered proteins were expressed ( Figure H in S1 File ) and proper folding was verified based on binding by VRC01 ( Figure I in S1 File ) . In successful validation of our model predictions , both antigens exhibited enhanced binding to their respective bNAb: PGT128 binding was enhanced to the +PGT128 antigen , and PGT121 binding was enhanced to the +PGT121 antigen ( Fig 4D and Figure I in S1 File ) . We also engineering a third antigen , with the goal of concomitantly optimizing glycosylation to enhance binding to both PGT121 and PGT128 bNAbs ( Fig 4B and 4C , +PGT121+PGT128 ) . The resulting +PGT121+PGT128 antigen showed significant enhancement of binding activity ( 12-fold increase ) to both bNAbs ( Fig 4D ) . These results argue that our antigen design approach based on the addition or removal of both distal and proximal agonistic or antagonist glycans , aimed at remodeling the overall topography of HIV Env antigens , can effectively improve antigenicity . As a yet further design challenge , we tested whether glycan engineering could selectively shift antigenicity between bNAbs–facilitating access to one particular bNAb while occluding another . Toward this aim , we designed antigens predicted to enhance PGT121 binding while inhibiting PGT128 binding or vice versa ( Fig 4B and 4C , +PGT121-PGT128 or -PGT121+PGT128 ) , despite the propinquity of the PGT121 and PGT128 epitopes . As an example for the +PGT121-PGT128 antigen , PGT128-specific antagonists such as N442 were added while PGT128-specific agonist glycan sites such as N140 were removed or kept absent . Because several glycan sites exist that impair binding of both bNAbs , decisions about which sites to include were dictated by the quantitative degree of a specific glycan contribution to individual bNAb binding models . For instance , N139 had a negative influence on binding for both PGT121 and PGT128 , but this effect was stronger for PGT121 than PGT128 ( Fig 4B , top panel ) ; accordingly , N139 was incorporated into the -PGT121+PGT128 antigen . The–PGT121+PGT128 antigen in fact demonstrated selective binding to PGT121 and PGT128 , to levels even higher than predicted; however , selective binding was not observed for the +PGT121-PGT128 antigen . While the algorithm successfully generated an antigen that was selectively recognized by PGT128 , we sought to understand why the antigen predicted to be PGT121-selective failed ( Fig 4D ) . Closer inspection of the glycan sites in +PGT121-PGT128 showed that a cluster of tightly packed glycans ( N339 , N355 , N362 , N363 , N392 , N396 N398 ) were chosen by the model , and we recognized that these could have adverse effects on PGT121 binding . To test the corollary hypothesis , then , that these new glycans may have created glycan clusters that sterically hindered Ab binding or unstabilized gp120 conformation ( Figure J in S1 File ) , we estimated their steric effects using glycan-glycan auto-correlations based on occupancy profiles . Relative to the total degree of steric hindrance observed for the MF535 . W0M . ENV . D11 gp120 sequence , which was approximately 1 ( Fig 4E ) , the overall degree of glycan steric effects across all 94 gp120s was relatively similar and homogenous across all sequences ( Fig 4E , pink region ) . This level of steric effects between glycans likely reflects a natural fitness landscape of the gp120 glycan density profiles that are naturally accommodated across clades and tiers of viruses . Similarly , glycan steric effects of the novel antigens , other than the +PGT121-PGT128 antigen , were close to those observed for the panel of gp120 proteins . On the other hand , the degree of steric hindrance observed for the +PGT121-PGT128 construct was significantly higher than all other antigens—nearly three times the degree of steric hindrance observed for the original MF535 . W0M . ENV . D11 sequence ( Fig 4E ) . Thus balancing steric effects , in addition to optimizing individual glycans , likely offers the optimal approach to modulate antigenicity . Based on this new insight , we designed the +PGT121-PGT128 antigen taking into consideration the mutual steric effects of glycans as an additional design principle . Reducing glycan steric hindrance while preserving the glycan that could concomitantly agonize PGT121 antagonize PGT128 , a new +PGT121-PGT128 2nd was generated . This novel antigen readily bound PGT121 and was not recognized by PGT128 ( Fig 4D ) , in a manner demonstrating good consistency between model predictions and experimental measurements ( r = 0 . 78 in PGT121 binding; r = 0 . 90 in PGT128 binding prediction , Figure K in S1 File ) , further validating our modeling approach . More broadly , these results suggest that an expanded analysis of the glycan occupancy profiles of a larger number of non-neutralizing and neutralizing Abs , against not only gp120 but also gp140 and the emerging native-like trimers , could prove beneficial for designing antigens yielding enhanced antigenicity to drive specific desired bNAbs .
With the growing appreciation for the participation of glycans in bNAb binding and the striking effects observed with the removal or addition of individual or groups of proximal glycans [39–45 , 48] , new antigen design approaches able to exploit these post-translational structures could offer a powerful avenue to improve development of therapeutically effective antigens . Our study aimed to develop a systems glycobiology approach to exploit glycosylation to enhance antigenicity , by integrating global site-specific glycan occupancy , peptide sequence , and Ab binding fingerprints to a panel of 94 distinct HIV envelope monomers , using computational modeling to predict engineered antigens optimized for desired binding characteristics . Our findings demonstrate an extensive heterogeneity in glycan occupancy profiles across 94 distinct HIV Env gp120s , dictated by unique sequons , which together define glycans that shape Ab binding . In addition , glycan mapping across these Env antigens raised a previously underappreciated role of both proximal and distal glycans together , distributed in 3D structures over the glycan surface , that govern Ab binding by potentially shaping the overall surface of the glycoprotein in manner that may alter epitope exposure and access . Moreover , this information was then used to build a computational antigen design algorithm , aimed at optimizing glycan occupancy profiles for desired antigenicity . Engineered antigens successfully enhanced binding to target bNAbs , and even were able to selectively skew antigenicity to discriminate between bNAbs that target overlapping regions of the viral envelope . These predictions were successfully validated by dedicated tests of a number of experimentally produced antigens with respect to binding properties across the set of targeted bNAbs . Of course , whether antigenicity represented by Ab binding will translate to enhanced immunogenicity[12 , 19] , and whether this approach more generally will apply to native Env trimers or other viral antigens , is uncertain . Nonetheless , our study provides proof of concept that a surface glycoprotein glycan shield can be engineered in a rationale , computation-based manner , to improve target Ab binding profiles . Analyses of glycan site occupancy on two Env trimers , the BG505 SOSIP . 664 [12] and Clade C CZA97 . 012 [13] trimers ( Figure L a-b in S1 File ) , has pointed to higher overall occupancy on trimeric forms of Env [13 , 37 , 72 , 73] compared to the various monomer sequences examined in this study . However , among these monomers occupancy varied from 63–97% ( Fig 1A , right panel ) , where some sequences exhibited occupancy profiles similar to previously analyzed native trimers , linked to the inclusion of specific high occupancy sequons that also exist within previously studied trimeric antigens . As occupancy is controlled co-translationally and post-translationally [74] , differences observed on trimers as well as monomers are likely not only related to misfolding or mis-processing of the proteins but also , critically , to sequence variation ( Figure D in S1 File ) , local secondary structure , as well as surface geometry around the sequons that may all contribute to determining whether Oligosaccharyltransferase ( OST ) can access a glycosylation site [75] . Moreover , that the highest degree of occupancy variation is observed at N-glycan sites critical for bNAb binding , further argues that the observed variation is not due to aberrations in glycosylation of monomers , but related to evolutionary selection of variability at key sites of viral vulnerability . However , beyond occupancy differences , that are likely less divergent between monomeric and trimeric protein structure , the composition of the glycan may be more highly influenced by quaternary structures due to enzyme accessibility issues . Thus , while only a limited number of trimers have been analyzed for glycan occupancy thus far , with restricted clade and neutralization tier coverage , future analysis of larger numbers of trimers may provide critical insights into the opportunities to glycoengineer more native-like molecules along the lines of our approach . Protein engineering strategies aimed at generating soluble native-like Env trimers , such as stablilzing the gp120:gp41 interaction [12–15] , fixing the epitopes in a closed conformation [16 , 17] , and optimizing combinations of different trimer variant sequences [76] have been proposed to either target bNAb precursors [21–23] or induce Abs with cross-reactive breadth [19 , 20] . We elected here to use a panel of 94 distinct gp120 monomers , due to the breadth of sequences that could be selectively analyzed , to exploit both sequence and glycan diversity in our glycan engineering efforts . Importantly , while nearly all sequons are reproducibly occupied on the SOSIP . BG505 immunogen similar to the previous report [73] ( Figure L c in S1 File ) , preliminary analyses of additional native-like trimers suggest increased variability in glycan occupancy ( Figure L a in S1 File ) , driven by diverging sequons . Despite these differences , the use of a library of monomeric antigens possessing sequence as well as glycan occupancy heterogeneity , gave us a unique opportunity to develop a robust model interrogating the role of individual glycans and sequences in tuning Ab binding profiles . This data set afforded the discovery of novel sequons that can be exploited to control levels of glycan occupancy in the future for HIV and other therapeutic proteins . Moreover , results from our study pointed to a role for distal glycans in shaping Ab binding profiles , emphasizing the need to include information outside of an Ab-binding footprint to develop antigenically enhanced antigens . Given that the glycan shield covers the majority of the Env surface , our glycan engineering approach may provide a novel strategy for vaccine development efforts via the modulation of overall antigen topography to selectively mask immunodominant , and potentially distracting , epitopes , while improving the targeted induction of bNAbs through the creation of targeted “glycan holes” that enable vulnerable site recognition [20 , 47] . Additionally , because inter-protomeric glycan:glycan interactions have been observed in the trimeric structure [37 , 77] , it may be additionally possible to optimize trimer stability or even adapt antigenicity fingerprinting to screen for improvements in germline-reverted B cell receptor recognition to prime immunity more effectively [44 , 78] . The MF535 . W0M . ENV . D11 gp120 sequence was selected as an initial point for our glycoengineering approach ( Fig 4B ) . While this wildtype gp120 contains N137 , N156 , and N301 , other critical glycans for PGT121 and PGT1218-binding ( N295 and N332 ) are missing [35 , 52 , 54 , 79] . Thus this sequence was optimal for engineering , given its negligible binding to both bNAbs ( Fig 2A , Figure E in S1 File ) . We designed our optimized gp120 proteins using a computational design algorithm in which the selected mutations were focused on the glycan site determinants for both PGT121 and PGT128 , including proximal and distal sites ( Fig 4B ) . While some N-glycans , such as the addition of N295 and N332 , likely have more profound effects on improving PGT121 and PGT128 binding , addition and removal of other glycans , that were both proximal and distal , also modulated overall as well as selective bNAb binding . Our computational design model involved a hybrid framework combining machine learning algorithms ( support vector machine , SVM ) with a Bayesian MCMC sampler . Advantages of integrating these two algorithms were multiple: the SVM implemented kernels capturing mutual dependencies among multiple variants ( glycans and protein residues ) ; the Bayesian MCMC sampler reduced computing time and mathematically ensured the identification of the optimal solutions in a high-dimensional feature space; the probabilistic machine learning model then dealt with uncertainties of measurements from high-throughput assays ( N-glycoproteome analysis , ELISA-based Ab binding ) , reducing model overfitting of noisy data; and , the flexible machine leaning enabled the incorporation of additional sequence , glycan structure , neutralization profiles , and affinity measurements , that ultimately led to the generation of a robust model . Future efforts aimed at designing next-generation glyco-engineered immunogens may benefit from a number of further advances including ( but not limited to ) : the incorporation of glycan occupancy data from a larger number of Env proteins ( including monomers , trimers , and native-like structures ) ; the inclusion of the glycan structure data itself ( e . g . , oligo-mannose , hybrid , complex ) ; the addition of binding profiles from a larger repertoire of Abs ( including both broadly neutralizing and non-neutralizing Abs ) , germline binding profiles , measures of binding affinity/avidity; the inclusion of virus:BCR evolutionary dynamics to help the algorithm learn to evolve envelope intermediates; and , the final consideration of occupancy profiles from envelopes on the surface of the virus or the cell , that may adopt distinct conformations . Given the enormous evolutionary landscape explored by HIV , and the fact that carbohydrates represent half the mass and most of the surface of the viral envelope , our study points to the importance of exploring both glycan occupancy and sequence diversity for optimal antigen design . In the future , our approach may be extended to address profiles associated with enhanced germline BCR binding or intermediate BCR ancestor interactions , ultimately aimed at designing sequential immunization strategies or “multi-valent vaccines” able to elicit multiple lineages of bNAbs from a single “super immunogen” . Ultimately , we believe that our work raises promising prospects for topographically designed optimized glycoproteins not only through protein engineering but also through the remodeling of the glycoprotein glycome , which in concert may maximally enhance immunogen antigenicity .
100 recombinant Gp120 protein monomers were purchased from Immune Technologies . Purity , protein quality , and testing were performed on recombinant antigens by Immune Technology Corp . Proteins were purified from 293 culture supernatants using nickel columns , and purity was assessed by SDS page by Immune Technology Inc . Additionally , we performed a second SDS-PAGE ( Figure B in S1 File ) upon receipt of these proteins to ensure that only high-quality , antigens were used for ELISAs and Mass Spectrometry . FPLC was performed on a subset of proteins to spot-check protein quality . New batches of proteins were acquired from Immune Technology if more than 1 band was observed on the SDS-page . Proteins that exhibited more than 1 band across 2 batches were eliminated from consideration , resulting in the final inclusion of only 94 recombinant gp120 monomers in the analysis ( S3 Table ) . PGT121 , A32 , F105 , 447-52D and VRC01 Abs were obtained from the NIH AIDS reagents program . 2G12 and B12 Abs were purchased from Polymun Scientific . PGT122 , 123 , 124 , 125 , 126 , 127 , 128 , 135 and 136 were generously provided by Dr Dennis Burton ( Scripps Research Institute ) . ELISA assays were performed by capturing gp120 monomers on D7324-coated ( anti–C-terminal gp120 sheep Ab , Aalto Bioreagents ) plates , to directionally and consistently position the recombinant proteins for antibody binding . Nunc Maxisorp 384-well plates were coated overnight at 4°C with 10 μg/ml of D7324 in 0 . 1M NaHCO3 ( pH 8 . 6 ) . Plates were washed 4 times with PBST ( PBS ±0 . 01% tween ) and blocked with PBSA ( PBS containing 5% BSA ) for 1 hour at room temperature . After 4 washes with PBST gp120 proteins were added at 80 ng/ml final concentration in PBSA ( optimal protein concentration for ELISA was evaluated by protein titration assays based on 6 selected proteins from different clades , Figure E in S1 File ) and incubated for 2 hours at room temperature . After washing 6 times with PBST , bNAbs were added at 10 μg/ml in assay diluent ( PBS containing 5% BSA and 20% sheep serum ) and incubated for 2 hours at room temperature . Following 6 washes in PBST , biotin-conjugated mouse anti-human IgG ( BD Biosciences ) was added to each well at 1:1000 dilution in PBSA and plates were incubated for 1 hour at room temperature . After 6 washes in PBST , high-sensitivity streptavidin-HRP ( Pierce ) was added to each well at 1:100 dilution in PBSA and incubated for 1 hour at room temperature . Plates were washed 6 times in PBST and developed by adding UltraTMB substrate ( Pierce ) to each well . Development was stopped by adding 2M sulfuric acid and plates read at OD450 with Tecan 1000 pro reader . Background OD values ( wells without gp120 ) were subtracted from test wells containing gp120 proteins . Two biological replicates and three technical replicates for each bNAb were performed to provide statistical estimation and evaluate assay reproducibility . ELISA data from biological replicates were normalized by median centering to avoid systematic variance , aiding in cross-experiments comparison . Reproducibility between replicate data sets was evaluated by examining the coefficients of variation ( CV ) . 94 HIV gp120 sequences were multiply aligned by Clustal Omega ( EMBL-EBI ) , based on the curated alignments HMM model from Los Alamos HIV database [80] . The consensus sequons ( N-X-S/T ) were identified in each aligned sequence to identify all potential glycosylation sites . The position number of individual sequons for each aligned sequence was mapped back to the HXB2 sequence . Nomenclature for the positions of the sequons not aligned to HXB2 was determined by the previously aligned position present within HXB2; for instance , N137 . 5 indicates the glycosylation site found at the fifth residue after Position 137 in HXB2 . 5 μg of each gp120 protein ( Immune Technology Corp . ) was denatured by incubating with 10 mM of dithiothreitol at 56°C for an hour and alkylated by 55 mM of iodoacetamide for 45 minutes in the dark prior to digestion with proteases ( Promega ) at 37°C overnight . Various combinations of proteases were used for different samples depending on their particular protein sequences to achieve optimal detection of potential glycosylation sites . The resulting glycopeptides were deglycosylated by incubating with PNGaseF ( ProZyme ) in the presence of H218O water ( Cambridge Isotope Laboratories ) overnight . The deglycosylated peptides were then dried and reconstituted in 0 . 1% formic acid . Deamidation of any Asn would result in a mass shift of +1 . We purposely controlled for mapping of N-linked sites by performing only the enzymatic deglycosylation in the presence of O18-water so as to generate a mass shift of +3 . Following deglycosylation , the sample was immediately removed from O18-water and replaced with non-isotope heavy buffers . We also performed the deglycosylation below pH7 since basic conditions promote deamidation [81] . Finally , when database searching , we allowed for Asn +1 and +3 for all Asn but only observed Asn +3 residues in consensus sequons further suggesting the lack of deamidation . In addition to Asn +1 ( deamidation ) and Asn +3 ( deglycosylation ) , we also allowed for oxidation of methionine and alkylation of cysteine residues . Peptides were separated on a 75 μm ( I . D . ) x 15 cm C18 capillary column ( packed in house , YMC GEL ODS-AQ120ǺS-5 , Waters ) and eluted into the nano-electrospray ion source of an Orbitrap Fusion Tribrid mass spectrometer ( Thermo Fisher Scientific ) with a 180-min linear gradient consisting of 0 . 5–100% solvent B over 150 min at a flow rate of 200 nL/min . The spray voltage was set to 2 . 2 kV and the temperature of the heated capillary was set to 280°C . Full MS scans were acquired from m/z 300 to 2000 at 120k resolution , and MS2 scans following collision-induced fragmentation were collected in the ion trap for the most intense ions in the Top-Speed mode within a 3-sec cycle using Fusion instrument software ( v1 . 0 , Thermo Fisher Scientific ) . A protein database was compiled from 94 gp120 protein sequences provided by Immune Technology Corporation . Raw spectra were searched against the gp120 protein database using SEQUEST ( Proteome Discoverer 1 . 4 , Thermo Fisher Scientific ) with full MS peptide tolerance of 20 ppm and MS2 peptide fragment tolerance of 0 . 5 Da , and filtered using ProteoIQ ( v2 . 7 , Premier Biosoft ) at the protein level to generate a 1% false discovery rate for protein assignments . Glycan occupancy Gai of a potential glycosylation sequon ai was determined as follows , Gai=∑POccupi∑POccupi+∑PUnoccupj ( 1 ) G represents a ratio of total spectral counts ( generated by ProteoIQ ) between the 18O-labeled peptides POccupi containing the glycosylated sequon ai and all detected peptides containing the sequon ai with glycosylated POccupi or unglycosylated PUnoccupi forms . The glycan occupancy threshold for ai was set as sum of POccupi and PUnoccupi > 10 to avoid instrumental noise at low spectral signal . Occupancy of peptides with multiple sequons was determined by analyzing the ratio of the fragments in MS/MS that in most cases lead to specific quantitative assignment of glycosylated sequons; in a limited set of cases where Asn residues were adjacent or 3 AAs apart and not able to be accurately separated for quantification by multiple unique MS/MS fragments , the sequons were considered to be equally occupied . To quantitatively assess direct relationships between site-specific glycosylation occupancy and Ab binding , a matrix of values for pairwise Pearson correlations between site occupancy and Ab ELISA binding profiles was constructed . A given correlation coefficient characterizes the strength of the relationship between occupancy levels at individual sites and Ab binding; the sign of the value informs whether glycan presence at that specific site has positive or negative influence on binding . bNAb binding to the HIV envelope usually involves multiple glycans , often located proximal to an epitope . However , distal glycans may also contribute through conformational changes to the envelope . Thus , to define global glycan site combinations concomitantly important for Ab binding , we developed an algorithm that integrated two computational methods . First , a support vector regression ( SVR ) algorithm that implements support vector machines ( SVMs ) [69] was employed as a supervised learning model to predict Ab binding activities based on glycan site occupancy values . Second , a Bayesian Markov Chain Monte Carlo ( MCMC ) with the biased random walk Metropolis–Hastings algorithm [70 , 71] was employed to approximate the high-dimensional posterior distribution within the data to identify the optimal combinatorial glycan determinants that best fit the model predictions for binding with the experimental data . The model . We identified 110 sequons from the alignment of the 94 gp120 sequences utilized for the Ab binding profiles . We hypothesized that each sequon could be classified as either a “determinant” ( critical for binding ) or a “non-determinant” ( not critical for binding ) for Ab binding . A determinant could either be an agonist or antagonist . The calculation to identify glycan determinants was based on a joint probability distribution P ( X ) from a multi-dimensional sequon space: Jointprobability=P ( X ) , X={x1 , x2 , . . . . . x110}xi={1determinant0otherwise ( 2 ) where the sequon index xi indicates a dimension , and contains two possible states [0 , 1] . The complete probability distribution containing 2110 states is overwhelming to exhaustive sampling . Therefore , we used an approximation to provide a statistical estimate of the glycan determinant distribution P ( X ) given a set of experimental observations . According to Bayes’ theorem , the conditional probabilities of the glycan determinants given an observed data set are: p ( Xopt|data ) =p ( data|Xopt ) p ( Xopt ) p ( data ) ( 3 ) where Xopt denotes the vector of the glycan determinants in the model and p ( Xopt | data ) represents the posterior of Xopt . p ( data | Xopt ) and p ( Xopt ) are likelihood and prior of Xopt , respectively . Since the ratio of posterior is only required following MCMC sampling , the equation simplifies: p ( Xopt|data ) ∝p ( data|Xopt ) p ( Xopt ) ( 4 ) Taking a log transformation then yields the sum of the log likelihood and the log prior: −ln ( posterior ( Xopt ) ) =−ln ( likelihood ( Xopt ) ) −ln ( prior ( Xopt ) ) ( 5 ) –ln ( likelihood ( Xopt ) ) was calculated by mean squared errors ( MSE ) between the prediction from SVR and the observed data . SVR was trained on a sub-set of glycan site occupancy profiles ( Xopt ) , and examined by means of 10-fold cross-validation randomly dividing the data into training and test sets with fitness estimated by calculating MSE . -ln ( prior ( Xopt ) ) was calculated by the sum of squared deviation between the selected and empirical glycan determinants , assuming that the distribution of the site identified between two states [0 , 1] follows a Gaussian distribution: −ln ( prior ( Xopt ) ) =∑j12σj2 ( xj−〈xj〉 ) 2 ( 6 ) σ2j and <xj> are variance and mean of glycan occupancy at the site xj . To sample the desired probability distribution P ( X ) we implemented a Metropolis–Hastings MCMC walk , in which the algorithm iteratively computed the posterior at current and next position . With this posterior assumed proportional to P ( X ) , the decision of jumping to the next position was then obtained: Xs+1={X'ifUnif ( 0 , 1 ) ≤α ( Xs , X' ) Xsotherwiseα ( Xs , X' ) =min ( 1 , posterior ( X' ) posterior ( Xs ) ) ( 7 ) Xs indicates the current position and X’ the candidate for the next position; Xs+1 the next position . The distribution α of the next position depends only on the current position value , by definition of the Markov chain . The algorithm accordingly generated a sequence of sample values in which the distribution of values closely approximated the original probability distribution P ( X ) . The calculation procedure started from a randomly selected set of glycan determinants and performed 50 , 000 iterations of MCMC walk for each Ab , in order to converge on a desired distribution ( Figure M in S1 File ) ; the first 5 , 000 iterations were discarded as a burn-in period which identified very different distributions . The output of this calculation provided marginal probabilities for each sequon identified as a determinant ( Figure N in S1 File ) . We performed 10 cycles of MCMC walk for each Ab to avoid the potential problem of the search getting trapped in local optima . To estimate statistical significance among the sequons , a background model was calculated in which the original glycan occupancy matrix was permuted by shuffling the order of gp120 proteins to preserve the heterogeneity of glycan occupancy yet capture noise effects . This background model was performed for each Ab using 100 permutations to estimate null distribution for individual sequons , with the null hypothesis posed that the glycan at the site has no impact on the Ab binding . Only the sites whose q-value after multiple-testing adjustment showed statistical significance ( q<0 . 01 ) ( Figure F in S1 File ) , rejecting the null hypothesis , were identified as glycan site “determinants” . We used a customized Matlab script for Metropolis–Hastings MCMC algorithm , implemented Matlab LIBSVM package [82] for SVR with SVR parameters , kernel type = linear , cost = 0 . 75 , and the epsilon = 0 . 1 . The algorithm was run on Linux computing clusters . To evaluate whether model balance between under- or over-fitting , a learning curve technique [83] was employed to calculate training error and cross-validation error as a function of training set size . Etrain ( θ ) =12N∑i=1N ( Jθ ( x ( i ) ) −y ( i ) ) 2Ecv ( θ ) =12Ncv∑i=1Ncv ( Jθ ( xcv ( i ) ) −ycv ( i ) ) 2 ( 8 ) where N indicates the training set size and Ncv the testing set size in cross-validation . Jθ ( x ) denotes the regression function of SVR and y represents the actual Ab binding fingerprint . A learning curve was used to evaluate the model fitting given the predictors ( Figure O in S1 File ) . An under-fitting model was recognized if only one glycan site occupancy profile was found as a predictor , in which large training errors quickly emerged for a small training set ( Figure O in S1 File , the left panel ) . An over-fitting model was obtained if all glycan occupancy profiles were found as predictors , in which the cross-validation errors decreased shortly but then grew as the training set size increased further ( Figure O in S1 File , the middle panel ) . A model using the glycan site determinants identified from the Bayesian machine learning algorithm as the predictors showed excellent fitting behavior , in which both the training and cross-validation errors converged at the level of the expected errors from experimental noise ( Figure O in S1 File , the right panel ) . To determine how the glycan determinants modulate Ab recognition , the directional weight ( DW ) was introduced and calculated: DW=p ( xi ) ×dd={1ifcorr ( Aby , Gxi ) >0‑1otherwise ( 9 ) p ( xi ) denotes the probability of a sequon being a “determinant” which was obtained from the MCMC-SVR model and d represents a direction [1 , –1] determined by the covariance matrix between individual glycan occupancy pattern versus the Ab fingerprint ( Fig 2C ) . Bayesian MCMC-SVR was then extended to identify protein residues critical for Ab binding . The peptide sequences of 94 gp120 proteins were pre-processed to remove less informative positions . Based on multiple sequence alignment , 534 aligned positions were identified across the 94 proteins . Positions at which the residues were identical in all proteins or where more than 10% of the proteins had unknown residues , denoted as gaps , were removed , leaving only 303 common positions . For importing these data into the model , the finalized sequence data was encoded as a 6 , 363–dimensional vector ( 303 positions × 21 amino acid types including the gap ) , where every single residue of the 303 positions in each sequence was converted from a categorized amino acid into a binary vector . Ultimately , three models—glycan alone , sequence alone , and both glycan and sequence—were constructed for each Ab to assess the model prediction performance as well as identify the critical glycan site and protein residues that may tune Ab binding . Using the Bayesian MCMC-SVR algorithm that identified the glycan determinants positively or negatively impacting Ab binding , the data were further analyzed to rationally design antigens with selective Ab antigenic-profiles . The same training set ( glycan occupancy , Ab fingerprints and envelope sequences ) and model algorithm were utilized , but now modified to search for global optimum solutions for best or worst binding ( the objective function -ln ( posterior ( Xopt ) ) is modified ) . Given a set of previously identified Ab-specific glycan determinant sites denoted as X , the model in this section attempted to determine the likelihood of the presence or absence of each critical site associated with the optimized Ab binding: Jointprobabilityoftheoptimizedbinding=P ( X ) X={x1 , x2 , . . . . . xm}xi={1Presence0Absence ( 10 ) where each glycan determinant xi contained two states [0 , 1] . Thus , the model aimed to approximate the joint probability distribution P ( X ) by implementing the same Bayesian MCMC sampling as earlier ( Eq 5 ) . –ln ( posterior ( Xopt ) ) represented the predicted antigenicity from the trained SVR model given a set of combinatorial sites Xopt . For sampling within the multi-dimensional probability distribution P ( X ) , we performed 200 , 000 iterations to reach convergence from an initial gp120 sequence , and estimated the global sequon occupancy profile for an antigen that yields optimized Ab binding . Each iteration simulated the evolutionary process of mutagenesis , in which the state of a selected site xi was altered between 0 and 1 analogous to random knockout/insertion of the sequon; an evolutionary driving force is assumed to mainly depend on predicted antigenicity . Accordingly , in each iteration determination was made of whether the sequon should be added or removed by computing the ratio of posteriors for the current and previous iteration according to Eq 7; that is , we computed the ratio of predicted antigenicity between current versus previous iterations . The final output of the model was the marginal probabilities of two states at every glycan determinant . High probability of state 1 suggested that the glycan determinant most likely facilitated Ab binding , so that the sequon should be included in the optimal antigen design . Conversely , high probability of state 0 suggested that occupancy at that sequon would impair Ab recognition , suggesting that this site should be excluded . Ultimately , the model was able to identify optimized profiles of glycan determinants on the antigens with ideal antigenicity . When generating a de novo peptide sequence from a gp120 sequence template to accommodate an ideal glycan determinant profile , addition and deletion of sequons was accomplished according to these rules: ( 1 ) if a sequon needed to be eliminated , a Asn-to-Gln substitution at the sequon was introduced; ( 2 ) if a new sequon needed to be introduced , a N-L-T sequence was substituted into the original sequence ( although if multiple sequons were modeled that were in close proximity [e . g . , N187 . 6 and N187 . 7] only the first site would be added to avoid steric hindrance ) . To incorporate the steric hindrance effects of proximal glycans , pairs that negatively impacted one another were identified . Our hypothesis was that if two glycans sterically constrained each other , the glycan occupancy patterns should be mutually exclusive; i . e . , they should have strong negative correlation . Therefore , a covariance matrix was constructed containing all pairs of glycan-glycan auto-correlations . To determine statistical significance , a nominal P-value of each pair was calculated , assuming that the distribution of all pairs of correlation coefficients followed a Gaussian process , and an adjusted p-value was corrected to account for multiple hypothesis testing ( False discovery rate adjusted p-value < 0 . 1 ) . Glycan pairs with significant mutual exclusion were then identified ( Figure J in S1 File ) . Next , to estimate the total degree of steric hindrance potentially induced on the epitope of the antigen , the mutually exclusive glycan pairs were taken into account . Each glycan pair possessed a steric hindrance weight based on the glycan-glycan correlation coefficient , and total weights of the included glycan pairs were summed up to represent total degree of steric hindrance . Mean and s . d . of the steric hindrance were estimated across all 94 gp120 proteins which represented the basal level of the steric hindrance occurring on naïve gp120 proteins . To incorporate the steric hindrance effects on antigen design , the antigens , after de novo design , were then evaluated for total degree of steric hindrance . Steric hindrance levels greater than the basal level ( > mean + 1 s . d . ) were considered to be larger than acceptable . Glycan sites that showed the least effect on bNAb binding from steric hindrance were removed one at a time until an acceptable level of steric hindrance was achieved . The process of removing steric glycan sites was repeated until total degree of steric hindrance was lower than mean+1 s . d . The sequence encoding wild-type MF535 . W0M . ENV . D11 gp120 and PGT121+/-/PGT128+/- optimized glycoproteins , including a signal peptide-‘MPMGSLQPLATLYLLGMLVASVLA’ at N-terminus and an avitag-‘GLNDIFEAQKIEWHE’ followed by a histag-'HHHHHH’ at the C-terminus , were synthesized and inserted into pcDNA3 . 1 ( Thermo Fisher ) . Plasmid DNA was purified and verified by sequencing . Plasmids encoding these proteins were transfected into 293-F cells ( Life Technologies , cat . no . R790-07 ) and proteins were isolated from expression supernatants 6 days after transfection . Briefly , gp120 proteins were purified by metal affinity chromatography using Ni-NTA resin ( Qiagen ) . Fractions containing gp120 were combined and oligomers , trimers and monomers were separated by gel filtration chromatography using a Hi-Load 16/60 Superdex 200pg column ( GE Healthcare ) . Protein purity was confirmed by SDS-PAGE gel electrophoresis and Western blotting using a mouse monoclonal Ab Chessie 13–39 . 1 ( NIH , AIDS Reagents Program ) ( Figure H in S1 File ) . MATLAB source codes for the Bayesian MCMC-SVR algorithm to identify glycan/sequence determinants , and the de novo antigen optimization design program , can be accessed from Supplementary Information . | Carbohydrates on the HIV Env glycoprotein , previously often considered as a “shield” permitting immune evasion , can themselves represent targets for broadly neutralizing antibody ( bNAb ) recognition . Efforts to define the impact of individual glycans on bNAb recognition have clearly illustrated the critical nature of individual or groups of glycans on bNAb binding . However , glycans represent half the mass of the HIV envelope glycoprotein , representing a lattice of interacting sugars that shape the topographical landscape that alters antibody accessiblity to the underlying protein . However , whether alterations in individual glycans alter the broader interactions among glycans , proximal and distal , has not been heretofore rigorously examined , nor how this lattice may be actively exploited to improve antigenicity . To address this challenge , we describe here a systems glycobiology approach to reverse engineer the complex relationship between bNAb binding and glycan landscape effects on Env proteins spanning across various clades and tiers . Glycan occupancy was interrogated across every potential N-glycan site in 94 recombinant gp120 recombinant antigens . Sequences , glycan occupancy , as well as bNAb binding profiles were integrated across each of the 94-atngeins to generate a machine learning computational model enabling the identification of the glycan site determinants involved in binding to any given bNAb . Moreover , this model was used to generate a panel of novel gp120 variants with augmented selective bNAb binding profiles , further validating the contributions of glycans in Env antigen design . Whether glycan-optimization will additionally influence immunogenicity , particularly on emerging stabilized trimers , is unknown , but this study provides a proof of concept for selectively and agnostically exploiting both proximal and distal viral protein glycosylation in a principled manner to improve target Ab binding profiles . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"machine",
"learning",
"algorithms",
"medicine",
"and",
"health",
"sciences",
"immune",
"physiology",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"applied",
"mathematics",
"immunology",
"microbiology",
"retroviruses",
"viruses",
"immunodeficiency",
"viruses",
"algorithms",
"simulation",
"and",
"modeling",
"mathematics",
"rna",
"viruses",
"artificial",
"intelligence",
"glycosylation",
"glycoproteins",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"immune",
"system",
"proteins",
"computer",
"and",
"information",
"sciences",
"sequence",
"alignment",
"bioinformatics",
"proteins",
"medical",
"microbiology",
"hiv",
"antigens",
"microbial",
"pathogens",
"recombinant",
"proteins",
"biochemistry",
"machine",
"learning",
"post-translational",
"modification",
"physiology",
"viral",
"pathogens",
"database",
"and",
"informatics",
"methods",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"lentivirus",
"glycobiology",
"organisms"
] | 2018 | Exploiting glycan topography for computational design of Env glycoprotein antigenicity |
West Nile virions incorporate 180 envelope ( E ) proteins that orchestrate the process of virus entry and are the primary target of neutralizing antibodies . The E proteins of newly synthesized West Nile virus ( WNV ) are organized into trimeric spikes composed of pre-membrane ( prM ) and E protein heterodimers . During egress , immature virions undergo a protease-mediated cleavage of prM that results in a reorganization of E protein into the pseudo-icosahedral arrangement characteristic of mature virions . While cleavage of prM is a required step in the virus life cycle , complete maturation is not required for infectivity and infectious virions may be heterogeneous with respect to the extent of prM cleavage . In this study , we demonstrate that virion maturation impacts the sensitivity of WNV to antibody-mediated neutralization . Complete maturation results in a significant reduction in sensitivity to neutralization by antibodies specific for poorly accessible epitopes that comprise a major component of the human antibody response following WNV infection or vaccination . This reduction in neutralization sensitivity reflects a decrease in the accessibility of epitopes on virions to levels that fall below a threshold required for neutralization . Thus , in addition to a role in facilitating viral entry , changes in E protein arrangement associated with maturation modulate neutralization sensitivity and introduce an additional layer of complexity into humoral immunity against WNV .
Flaviviruses are a group of positive-stranded RNA viruses that are of global significance due to their widespread distribution and their ability to cause a variety of diseases in humans [1] . West Nile virus ( WNV ) is a mosquito-borne member of this genus and is the etiologic agent of West Nile encephalitis . WNV is endemic in parts of Africa , Australia , Europe , Asia , and the Middle East and has been responsible for periodic outbreaks of encephalitis in humans and horses . The introduction of WNV into North America in 1999 and its rapid spread across the United States into Canada , Mexico , and the Caribbean identifies this virus as an emerging pathogen of clinical and economic significance for the Western Hemisphere ( reviewed in [2] ) . While seroprevalence studies indicate that most WNV infections of humans are subclinical , clinically apparent infections range from a febrile illness ( West Nile fever ) to more severe and potentially fatal neurologic disease [3] . Currently , no WNV vaccine has been approved for use in humans and treatment is supportive . Flaviviruses are small ( ∼50 nm diameter ) spherical virions composed of three structural proteins ( envelope ( E ) , premembrane ( prM ) , and capsid ( C ) ) , a lipid envelope , and an ∼11 kilobase monocistronic RNA of positive-sense polarity [1] . Crystal structures of the E protein of several related flaviviruses ( WNV , dengue virus ( DENV ) , tick-borne encephalitis virus ( TBE ) ) reveal an organization of three domains connected by flexible hinges ( reviewed in [4] ) . Domain III ( DIII ) is an immunoglobulin-like fold that is thought to participate in interactions between virions and cellular factors associated with virus entry . Domain II ( DII ) is a long , finger-like domain that contains a stretch of 13 conserved , hydrophobic residues that form an internal fusion loop . DIII and DII are linked together by a β-barrel structure that comprises domain I ( DI ) . The structure of prM , which forms heterodimers with the E protein during virion biogenesis , is presently unknown . Flaviviruses assemble at the endoplasmic reticulum ( ER ) and bud into the lumen as immature virus particles [5] . Cryoelectron microscopic reconstructions of immature virions reveal an icosahedral arrangement of 60 trimeric spikes composed of prM∶E heterodimers in which the prM protein is positioned to cover the fusion peptide located at the distal end of each E protein of the trimer [6] , [7] . In this position , prM may prevent conformational changes that would inactivate the E protein during virion egress through mildly acidic compartments of the secretory pathway [8] , [9] . During transit through the trans-Golgi network ( TGN ) , prM is cleaved by a cellular furin-like protease resulting in the formation of a small virion-associated M peptide and the release of the amino-terminal “pr” portion of the protein [10] . This required cleavage step promotes a rearrangement of E proteins on the surface of the virion and the formation of a mature virus particle . Mature flavivirus virions are relatively smooth and composed of 90 anti-parallel E protein dimers arranged with pseudo-icosahedral symmetry [11] . Antibodies are a critical component of host defenses against flavivirus infection and mediate protection via effector functions and by direct neutralization of virus ( reviewed in [12] ) . The primary target for neutralizing antibodies is the E protein , although antibodies specific for prM have been identified [13]–[15] . More than twelve distinct epitopes have been identified on the surface of the E protein that elicit antibodies characterized by varying degrees of neutralization potency in vitro and efficacy in vivo [16]–[22] . Neutralization of flavivirus infection is a multiple “hit” phenomenon in which virus inactivation occurs once the number of antibodies bound to a virion exceeds a required threshold [23] , [24] . Previous studies with an extremely potent neutralizing mAb specific for a highly accessible epitope on an upper lateral surface of WNV DIII ( DIII-lr ) suggest this threshold is approximately 30 mAbs [23] . The pseudo-icosahedral arrangement of E proteins on the virion displays the E protein in three distinct chemical environments defined by proximity to the two- , three- , or five-fold axes of symmetry [25] . Epitopes in each of these environments may be differentially accessible for antibody binding due to steric constraints imposed by adjacent E proteins on the virus particle [18] , [26]–[28] . As a result , the number of sites available for binding may differ among structurally distinct epitopes on the virion . Antibodies that bind highly exposed determinants may exceed the stoichiometric threshold for neutralization by binding a small fraction of accessible epitopes on the virion ( low occupancy ) . In contrast , epitopes predicted to be poorly exposed may require nearly complete occupancy to achieve threshold requirements for neutralization [23] . Furthermore , some epitopes on the virion may not be accessible to antibody engagement with a stoichiometry that exceeds the threshold required for neutralization . Thus , antibodies that recognize such epitopes may neutralize poorly , or not at all , even at concentrations that permit saturation because too few antibodies can simultaneously dock on the virion . Paradoxically , many antibodies that recognize poorly accessible epitopes on the mature virion still show neutralizing activity in vitro and in vivo [18] , [26] . How antibodies engage poorly accessible epitopes on virions with a stoichiometry that permits neutralization is difficult to reconcile using existing static models of virion structure and envelope organization . In this study , we investigate how changes in flavivirus structure associated with virion maturation impact the neutralizing activity of antibodies to WNV .
The neutralization potential of an antibody is governed by the number of sites on the virion available for binding ( epitope accessibility ) and the strength of binding ( affinity ) [23] , [29] , [30] . Although molecular modeling studies of the mature virion suggest that many of the known epitopes on the E protein are poorly accessible , antibodies to these determinants still neutralize infection to varying degrees [18] , [26] . To investigate mechanisms that govern the potency of antibodies that target poorly accessible epitopes , high-resolution neutralization profiles were generated for a group of antibodies that recognize structurally distinct epitopes on the WNV E protein [17] , [18] ( Fig . S1 ) . Neutralization potency was estimated for each antibody using a recently described and validated neutralization assay employing WNV reporter virus particles ( RVPs ) [23] , [31] . RVPs are produced by complementation of a sub-genomic replicon by expression of the structural proteins of the virus in trans , allowing infection to be measured as a function of reporter gene expression [32] . Monoclonal antibodies ( mAbs ) that bind a highly accessible determinant on the upper lateral ridge of DIII ( DIII-lr ) potently neutralize WNV infection in vitro and exhibit significant protective capacity in vivo [16]–[18] , [23] , [31] . Dose-response curves obtained with the DIII-lr-specific mAb E16 were sigmoidal , relatively steep , and revealed complete neutralization of WNV RVPs at low concentrations of antibody ( Fig . 1a; EC50 0 . 043 nM ) [23] . In contrast , none of the previously described E-protein-specific monoclonal antibodies recognizing determinants outside the DIII-lr were capable of completely blocking infection [18] , [23] ( Fig . 1b–d ) . By comparison with E16 , dose-response profiles obtained for most DI- and DII-specific mAbs were relatively flat and revealed reduced neutralization potency , as illustrated by the DII fusion loop ( DII-fl ) -specific mAb E64 [Fig . 1b; EC50 32 . 2 nM ( n = 2 ) ] . A second pattern was observed for mAbs E121 [Fig . 1c; EC50 0 . 69 nM +/− 0 . 39 ( n = 6 ) ] and E53 [Fig . 1d; EC50 1 . 0 nM +/− 1 . 4 ( n = 5 ) ] that recognize determinants on the lateral ridge of DI ( DI-lr ) and the DII-fl , respectively [18] ( Fig . S1 ) . Dose-response profiles for these antibodies were sigmoidal , but in contrast to results obtained with the DIII-lr specific mAb E16 ( Fig . 1a ) , did not completely neutralize even at high concentrations of antibody . Instead , the level of neutralization plateaued to reveal a fraction of virions resistant to neutralization at any concentration of antibody tested . The existence of a fraction of virions resistant to neutralization even at saturating concentrations of antibody ( ∼32% +/− 9% and ∼19% +/− 7% for E53 and E121 , respectively ) implies heterogeneity in antibody accessibility among individual virions in the population , with some virions unable to accommodate antibody binding with a stoichiometry sufficient for neutralization [23] . The arrangement of E proteins on the virion determines the accessibility of epitopes and thereby influences the neutralization potency of antibodies . Therefore , the dramatic changes in conformation and organization of E proteins on the virion during the flavivirus life cycle have the potential to impact antibody binding and function [6] , [7] , [25] . On immature virions , E proteins exist in trimeric spikes as heterodimers with the prM protein . The role of prM in this context is to regulate when E proteins acquire the capacity to mediate membrane fusion , presumably by controlling the oligomerization state of the E protein on the virion . Indeed , genetic studies indicate that furin-mediated cleavage of prM is required for the release of infectious virions [33] . However , relatively little is known about the relationship between the extent of prM cleavage and the acquisition of infectivity on individual virus particles . Virions may acquire the potential to infect cells following prM-cleavage-mediated activation of a relatively small fraction of the E proteins on the virion [8]–[10] , [34] , [35] , as observed for the analogous E2 protein of alphaviruses [36] . Virions containing at least some uncleaved prM protein are present in bulk virus populations and are infectious [35] , [37] . Because epitope accessibility may differ among individual viruses at intermediate or incomplete stages of maturation , we investigated the impact of virion maturation on the neutralization sensitivity of populations of virions containing different levels of uncleaved prM protein . WNV RVPs that are produced using our standard complementation conditions ( std-RVPs ) contain detectable amounts of uncleaved prM protein ( Fig . 2a ) . This observation was true after production in several cell lines ( BHK-21 , 293T , Vero ) ( Fig . 2a , data not shown ) , and agrees with previous studies with infectious virus [9] , [32] , [38]–[40] or subviral particles [38] . Over-expression of human furin protease in transfected BHK-21 cells producing RVPs increased the efficiency of prM cleavage to levels that no longer allowed detection of unprocessed prM by Western blotting ( furin-RVPs ) [37] , [41] . In contrast , treatment of BHK-21 producer cells with the weak base ammonium chloride reduced but did not absolutely block prM cleavage as has been described for other flaviviruses ( NH4Cl-RVPs ) [34] . Together , these modifications to our complementation approach allow for the production of RVPs that contain virtually no ( furin-RVPs ) , low levels ( std-RVPs ) , or high levels ( NH4Cl-RVPs ) of uncleaved prM . RVPs produced using all three transfection strategies are infectious , as shown by their capacity to infect Raji-DCSIGNR cells , albeit to differing degrees ( Fig . 2b ) . To determine if the cleavage state of prM present on infectious virions impacts sensitivity to neutralization by anti-E antibodies , we compared the ability of mAbs to block infection of RVPs produced using either standard transfection conditions ( Fig . 3; std-RVPs , blue circles ) or cells that over-express human furin protease ( Fig . 3; furin-RVPs , green circles ) . Three antibodies directed against the DIII-lr neutralized infection of both populations of RVPs with relatively similar potency ( Fig . 3a–c; 0 . 9- , 1 . 5- , and 4 . 5-fold reduction in neutralization sensitivity of furin-RVPs for mAbs E16 ( p = 0 . 67 ) , E24 ( p = 0 . 06 ) , and E49 ( p = 0 . 01 ) , respectively ) . In contrast , mAbs recognizing determinants on the DI-lr ( Fig . 3d , E121 ) and DII-fl ( Fig . 3e , E60 and Fig . 3f , E53 ) displayed a markedly reduced capacity to neutralize infection of completely mature virions ( furin-RVPs ) . When assayed using furin-RVPs , the potency of E60 was reduced approximately 95-fold relative to std-RVPs ( 4 . 3 +/− 2 nM ( n = 8 ) vs . 408 +/− 113 nM ( n = 8 ) for std- and furin-RVPs respectively ( p<0 . 004 ) ) whereas mAbs E53 ( n = 5 ) and E121 ( n = 6 ) did not neutralize a significant fraction of furin-RVPs . The reduced ability of mAbs E53 , E60 , and E121 to neutralize mature virions was not due to an inability to bind the virus particle . Antibody-dependent enhancement of infection ( ADE ) describes the dramatic increase in infection of Fc-γ or complement receptor-bearing cells in the presence of sub-neutralizing concentrations of antibody or immune sera . Each of these mAbs strongly enhanced infection of std- and furin-RVPs in cells expressing activating Fc-γ receptors [42] ( Fig . 4a , 4d , 4g , respectively ) . Consistent with this , both std-RVPs and furin-RVPs bound E53 , E60 , and E121 mAbs when either captured directly to plastic ( data not shown ) , or used in a sandwich ELISA in which virus was captured using a humanized form of mAb E16 ( a DIII-lr antibody that did not exhibit maturation-dependent differences in binding or neutralization ) . Using the latter approach , we measured the strength of binding of E53 , E60 , and E121 to WNV RVPs produced in the presence ( furin-RVPs ) or absence of furin ( std-RVPs ) ( Fig . 4b , 4e , 4h , respectively ) . All three mAbs bound with relatively high affinity to both std-RVPs and furin-RVPs , with no significant differences in the strength of antibody binding to furin-RVPs relative to std-RVPs noted ( ( 1 . 1 fold difference; p = 0 . 83 ) , ( 1 . 2 fold difference; p = 0 . 67 ) , and ( 1 . 3 fold difference; p = 0 . 38 ) for E53 , E60 , and E121 , respectively ) . The ability of the DI- and DII-specific mAbs to bind virions with relatively high affinity regardless of maturation state appears to conflict with their modest neutralizing capacity . Integration of affinity data with the neutralization dose-response curves , however , reveals that E53 , E60 , and E121 inhibit WNV infection only when a large fraction ( 99% ) of accessible epitopes on virions are bound by antibody ( Fig . 4c , 4f , 4i , respectively ) . The presence of a fraction of virions in std-RVP populations that remain infectious even at concentrations of antibody that allow complete occupancy suggests that a subset of virions do not display these epitopes at a stoichiometry sufficient to exceed the threshold required for neutralization [23] . In agreement , the occupancy requirements of these DI and DII mAbs suggest they bind an epitope that is not accessible on the average virion in std-RVP populations at a level that greatly exceeds the threshold requirements for neutralization [23] . Thus , even a small change in accessibility can significantly affect the potential for virion neutralization . If the changes in the arrangement of E proteins on virions that occur during maturation of WNV results in a reduction in the accessibility of epitopes in DI and DII , we would predict that populations of less mature virions containing greater amounts of non-cleaved prM should be more sensitive to neutralization by these mAbs . To test this hypothesis , WNV RVPs were produced in HEK-293T cells in the presence or absence of exogenous human furin plasmid , or in the presence of ammonium chloride , which inhibits prM cleavage . HEK-293T cells were used to produce the RVPs for these experiments because they support the release of higher titers of infectious RVPs relative to BHK-21 cells , allowing production of NH4Cl-RVPs at titers sufficient for neutralization studies under conditions of antibody excess . Neutralization profiles for all three populations of RVPs were generated using the indicated mAbs ( Fig . 5 ) . In agreement with experiments using BHK-derived RVPs , the potency of mAbs that recognize the highly accessible DIII-lr did not appreciably change when the proportion of mature and immature virus particles in the population was manipulated ( Fig . 5a ) . In contrast , the DI- ( E121; Fig . 5b ) and DII-specific mAbs ( E53; data not shown ) were less effective at neutralizing furin-RVPs , but displayed an increased capacity to block infection of virions with increased levels of prM ( NH4Cl-RVPs ) . Neutralization profiles by some DIII-specific mAbs exhibited a prominent resistant fraction , similar to our results with E121 . E22 binds an epitope on a lower lateral surface of DIII that is poorly exposed on mature virions [23] . By comparison to studies with std-RVPs , dose-response studies with E22 revealed a reduced capacity to neutralize furin-RVPs , whereas studies with NH4Cl-RVPs revealed greater neutralizing activity ( Fig . 5c ) . These results indicate that changes in neutralization sensitivity associated with maturation can occur for epitopes throughout the E protein and are not restricted to those in proximity to regions involved in prM-E protein interactions . Recent studies indicate that the human humoral immune response to flavivirus infection is narrower than anticipated , with antibody specificity focused on determinants in the fusion loop in DII . B-cell repertoire analysis of three WNV-infected humans revealed that only 8% of WNV-specific B-cell clones produced antibodies specific to DIII , whereas almost half produced antibody that bound determinants in DII [43] . In addition , functional studies of the polyclonal response of WNV-infected horses and humans indicate that the neutralization activity of sera is not dependent upon antibodies directed against the DIII-lr epitope [44] , [45] . Thus , during the natural course of flavivirus infection , many of the neutralizing antibodies present in the polyclonal response may be directed against determinants that are modulated by the maturation state of the virion . To investigate whether the potency of a polyclonal immune response is sensitive to the maturation state of WNV , we obtained serum from participants of two different phase I clinical trials of candidate WNV vaccines . First , the neutralizing antibody response in serum of twelve individuals immunized with a DNA vaccine encoding WNV prM-E was characterized [46] , [47] . This vaccine encodes the prM-E proteins of WNV , and is thought to promote the release of subviral particles in vivo . RVPs that incorporate a single amino acid substitution in the E protein ( T332K ) that abrogates antibody binding to all strongly neutralizing mAbs that map to the DIII-lr epitope [16] , [17] , [23] were used to establish that the neutralizing response to this vaccine was associated with non-DIII-lr antibodies . Neutralization titers obtained with T332K-RVPs were not significantly reduced in any individual when compared to std-RVPs ( Fig . S2a ) , consistent with prior studies in WNV infected humans and horses [44] , [45] . In contrast , in parallel studies performed with furin-RVPs , we observed a significant reduction in the capacity of sera from half of the recipients to neutralize furin-RVPs as compared to std-RVPs ( Fig . 6a and S2b ) . In fact , some individuals mount a response with very little capacity to neutralize mature virus ( furin-RVPs ) ( compare Fig . 6a and 6b ) . Because the E proteins of subviral particles are arranged with a different geometry ( T = 1 ) relative to the pseudo-icosahedral ( T = 3 ) structure of infectious virions , it is possible that the repertoire of antibodies elicited by this form of vaccination may be more sensitive to the maturation state of WNV than one elicited by an infectious virion . Therefore , we also characterized the polyclonal response of six recipients of a live-attenuated WNV vaccine . In agreement with the results described above , the neutralizing antibody response of roughly half the recipients of the live-attenuated candidate vaccine also displayed a significantly reduced capacity to neutralize furin-RVPs relative to their potency against std-RVPs ( Fig . S2c ) . Two of these subjects failed to neutralize furin-RVPs at all , despite a capacity to neutralize std-RVPs ( compare Fig . 6c and 6d ) . In contrast , sera from two of these recipients neutralized furin-RVPs at modestly higher titers than observed with std-RVPs . As the concentrations and specificities of the individual antibodies that comprise a polyclonal response are probably dynamic and could differ following vaccination with different antigens or using different immunization schedules , additional studies will be required to understand the factors that determine the sensitivity of the humoral response to heterogenous populations of flaviviruses . Together , these studies suggest that the maturation state of WNV represents an additional layer of antigenic complexity , with the polyclonal response of some vaccine recipients directed against epitopes with a reduced capacity to neutralize mature virions .
Neutralization of WNV is a “multiple-hit” phenomenon that requires the simultaneous engagement of the virus particle by as many as 30 antibody molecules [23] , [24] . In this regard , the neutralization potential of an antibody is determined by the strength of binding and the abundance of its epitope on the virion ( reviewed in [29] ) . Antibodies that bind the highly accessible DIII-lr epitope can exceed the threshold required for neutralization by binding a fraction of available epitopes displayed on the virion [23] . However , many of the epitopes recognized by antibodies produced following natural infection are not predicted to be accessible on the fully mature virion due to steric constraints imposed by the pseudo-icosahedral arrangement and packing of the E protein [18] , [26] . Antibodies that recognize poorly exposed epitopes must bind a larger fraction of sites to exceed the stoichiometric requirements for neutralization [23] . Consistent with this , studies with mAbs E53 , E60 , and E121 which bind two distinct , poorly exposed epitopes on the E protein ( DII-fl and DI-lr ) indicate that neutralization of WNV infection occurs only when all ( >99% ) accessible determinants on the average virion are engaged by antibody . For antibodies that bind epitopes requiring virtually complete occupancy to achieve neutralization , even modest changes in epitope accessibility can significantly affect the outcome of antibody binding . While the changes in E protein organization that define flavivirus maturation provide a mechanism for regulating the fusion activity of the class II glycoproteins of the virion [48] , a consequence of maturation may be to mask epitopes recognized by the humoral response . Increasing the efficiency of prM cleavage ( virion maturation ) significantly reduces neutralization potency of many antibodies specific for epitopes predicted to be poorly exposed on the mature virion . This is apparent by comparing the neutralization profiles of the DI-specific mAb E121 with RVPs containing virtually no ( furin-RVPs ) , low levels ( std-RVPs ) , or high levels ( NH4Cl-RVPs ) of uncleaved prM . In contrast , potently neutralizing mAbs specific for a highly accessible epitope on the DIII-lr ( e . g . , E16 ) neutralize WNV RVPs regardless of the amount of prM present . Antibodies specific for poorly accessible epitopes do not completely lose the capacity to bind mature virions , nor do they bind with significantly reduced affinity . Instead , our studies suggest that maturation reduces the number of antibodies that may simultaneously bind the virion to levels that do not exceed a required threshold for neutralization even at full occupancy . Because the arrangement of E protein on “partially mature” virions that retain uncleaved prM protein has not been resolved by structural studies , it remains uncertain how prM increases the accessibility of epitopes that are poorly accessible on the mature virion . While not observed in our studies , the presence of prM on virions will likely reduce the binding of some epitopes , as has been observed by ELISA using DENV and TBE [8] , [35] . However , biochemical analyses of bulk virion populations may be limited , as these methods average the contribution of individual virus particles without regard to their infectious potential . While none of the antibodies analyzed in this study bind prM directly , one potential mechanism for an increased stoichiometry of binding to prM-containing viruses is an increase in the accessibility of epitopes on the E proteins when arranged as trimers associated with prM , relative to their accessibility in the pseudo-icosahedral mature virion . However , several of the mAbs with altered neutralization sensitivity recognize the fusion loop , which may be masked by direct interactions with prM [7] . Alternatively , changes in E protein epitope accessibility associated with maturation may reflect dynamic aspects of virion structure that are not evident from existing structural studies . Cryo-EM reconstructions provide an average structural state of E proteins on the virion . These static models of E protein arrangement cannot account for antibody binding to transiently exposed determinants that result from dynamic and/or lateral movement among E proteins on the virion . Partially mature virions that contain uncleaved prM may be more dynamic than mature virus with respect to the number or stability of alternate structural conformations that E proteins can attain . Depending upon the proportion of immature , partially mature , and mature virions in a population of WNV , the same antibody may have little , modest , or significant neutralizing activity . From a technical perspective , maturation state-dependent neutralization has significant implications for the reproducibility of neutralization tests among laboratories , and may be important for determining whether antibodies in a given serum sample are judged as protective . More importantly , these studies suggest that the protective capacity of an antibody may depend in part upon the maturation state of virus delivered through the bite of a mosquito or released from infected human tissues in vivo . While the functional contribution of antibodies of different specificities in sera from infected or vaccinated humans is not yet understood , humans do produce large numbers of antibodies specific for determinants outside of the DIII-lr [43] , [44] . These may exhibit a significantly reduced potential to neutralize mature WNV . Indeed , half the recipients of two candidate WNV DNA vaccines evaluated in this study generated a polyclonal response that was notably less effective at neutralizing mature virus . Whether response to immunization with other classes of WNV vaccines , different immunization schedules , or natural infection will result in responses more capable of neutralizing all forms of the virus is of significant interest . Together these studies suggest that a consequence of maturation of WNV is a reduction in sensitivity to neutralization by antibodies recognizing specificities that are an important component of the humoral response to infection and vaccination . The influence of maturation on the neutralization sensitivity of WNV identifies an unappreciated functional consequence for the heterogeneity of prM cleavage in populations of flaviviruses [8]–[10] , [34] , [35] , [49] , and introduces an additional layer of complexity into analyses of humoral immunity against WNV .
BHK-21 WNIIrep-G/Z , 293T , Vero , Raji-DCSIGNR and K562 cells were maintained as described previously [23] , [31] . Neutralization studies were performed on sera obtained from two WNV vaccine trials sponsored by the NIH . First , sera were obtained from twelve participants in a single-site , Phase I , open-label study to examine the safety , tolerability , and immune response to an investigational recombinant DNA WNV vaccine encoding prM and E , described in detail elsewhere [46] . Second , sera from six participants of a Phase I double-blinded , placebo-controlled study to evaluate the safety , infectivity , attenuation and immunogenicity of a live-attenuated WNV/DENV4 vaccine were obtained for analysis ( A . Durbin , unpublished ) . These studies and subsequent analyses were performed in compliance with the guidelines of The U . S . Department of Health and Human Services ( DHHS ) , and the protocols were approved by the respective Institutional Review Boards . Neutralization studies with WNV-immune sera were performed on Raji-DCSIGNR cells with RVPs produced in BHK-21 cells as described below . RVPs were produced using methods and constructs described previously [23] , [31] , [41] . For RVPs produced in BHK cells , BHK-21 WNIIrep-G/Z cells were transfected with a plasmid encoding WNV C-prM-E and an empty pcDNA vector using a 1∶3 ratio by mass ( std-RVPs ) . Std-RVP stocks derived from HEK-293T cells were produced by transfection with plasmids encoding either the WNIIrep-G/Z or WNIIrep-REN replicon , C-prM-E , and pcDNA using a 0 . 5∶1∶2 . 5 ratio . Maturation of RVPs produced in either cell type was enhanced ( furin-RVPs ) by including a plasmid encoding human furin protease in place of the pcDNA vector . To produce immature RVPs , media from transfected cells was exchanged with media containing 20 mM ammonium chloride ( in phosphate buffered saline ) at 12 hours post-transfection , and again two hours later . Transfections were performed in T75 flasks using 40 µg of DNA and Lipofectamine 2000 ( Invitrogen , Carlsbad , CA ) and harvested at 48 hours post-transfection . The efficiency of prM cleavage of populations of RVPs produced in either BHK-21 or HEK-293T cells was determined by Western blotting as described below . The infectious titer of RVP preparations was determined by infecting Raji-DCSIGNR cells with serial four-fold dilutions of RVPs in triplicate . Cells were lysed 36–40 hours post-infection and assayed for luciferase activity according to the manufacturer's instructions ( Promega , Madison , WI ) . To relate the infectivity of each population of RVPs relative to the number of particles in the supernatant , the RNA content of each stock was measured as described previously [38] . Neutralization studies were performed using Raji-DCSIGNR cells as described previously [18] , [23] . WNV RVP stocks were diluted and incubated with mAb for 60–120 minutes at room temperature . Antibody-RVP complexes were then added to pre-plated cells in triplicate . Percent infection was measured by flow cytometry at 48 hours after RVP addition . The EC50 of each antibody was predicted by non-linear regression analysis using a variable slope . Statistical comparisons of the neutralization potency of different mAbs were performed using the T-test ( GraphPad Prism 4 , GraphPad Software Inc . , San Diego CA ) . Antibody-dependent enhancement of infection was measured using K562 cells that express the activating Fc-γ receptor CD32a as described [23] . RVPs were harvested at 48 hours post-transfection , filtered using a 0 . 22 µM filter , and concentrated and partially purified by ultracentrifugation through a 20% sucrose cushion . RVPs were resuspended in TNE buffer ( 100 mM Tris , 2 M NaCl , 100 mM EDTA , pH adjusted to 7 . 4 ) and analyzed using SDS-PAGE and Western blotting . E protein was detected using the mAb E16 ( 1 µg/ml ) , whereas prM was detected using a commercial antibody specific for residues 8–27 of the WNV M protein ( 1 µg/ml ) ( Imgenex , San Diego , CA ) . Antibody affinity was measured using an indirect ELISA and WNV RVPs produced in HEK-293T cells . High-protein-binding plates were coated with a 1 µg/ml humanized mAb E16 overnight at 4 °C using alkaline conditions . Plates were blocked in blocking buffer ( BB: 1X PBS , 0 . 05% Tween-20 , and 1% BSA ) and WNV RVPs diluted in BB were then bound at room temperature for 2 hours with gentle shaking . Virus particle-coated plates were then incubated in the presence of serial dilutions of anti-WNV antibodies under conditions of antibody excess . Bound antibody was detected using a horseradish peroxidase-conjugated goat anti-murine kappa chain antibody . Antibody affinity was estimated using non-linear regression with a hyperbolic equation that describes the binding of a ligand to a receptor under conditions that follow the law of mass action ( one-site binding equation; Bound = ( ( Bmax ) ( X ) / ( KD+X ) ) . Statistical comparisons were made using the T-test ( GraphPad Prism 4 ) . The occupancy requirements for neutralization by each antibody were estimated by plotting data from mAb dose-response curves ( y-axis ) against the percentage of accessible epitopes bound by antibody at each concentration of antibody as described previously [23] . The percentage of accessible epitopes bound by antibody was computed at each point on the dose-response curve using the affinity data obtained above by solving the equation: percent bound = [Ab]/ ( [Ab]+KD ) . | West Nile virus ( WNV ) virions incorporate 180 envelope ( E ) proteins that are the primary target of neutralizing antibodies . As newly formed WNV virions are released from infected cells , the E proteins undergo a significant organizational change associated with maturation into an infectious virus . However , this process is not always efficient , as populations of infectious WNV include virions that did not complete the maturation process and may be heterogeneous with respect to the arrangement of E proteins on the virion . In this study , we found that neutralization by antibodies specific for epitopes commonly recognized in vivo is strongly impacted by the maturation state of WNV . Our studies suggest that maturation of WNV reduces the accessibility of some , but not all , epitopes on the virion for antibody binding . Virions that retain some immature character can be neutralized by monoclonal antibodies that fail to block infection of populations of WNV composed solely of mature virions . Similar results were found using polyclonal human serum obtained from volunteers of two clinical trials of candidate WNV vaccines . These studies identify unappreciated aspects of the antigenic complexity of WNV and highlight the importance of understanding the heterogenous forms of WNV that may be introduced into or replicating within the host . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"virology/vaccines",
"infectious",
"diseases/viral",
"infections",
"immunology/immune",
"response",
"virology/host",
"antiviral",
"responses"
] | 2008 | Maturation of West Nile Virus Modulates Sensitivity to Antibody-Mediated Neutralization |
In plants , pollinator adaptation is considered to be a major driving force for floral diversification and speciation . However , the genetic basis of pollinator adaptation is poorly understood . The orchid genus Ophrys mimics its pollinators' mating signals and is pollinated by male insects during mating attempts . In many species of this genus , chemical mimicry of the pollinators' pheromones , especially of alkenes with different double-bond positions , plays a key role for specific pollinator attraction . Thus , different alkenes produced in different species are probably a consequence of pollinator adaptation . In this study , we identify genes that are likely involved in alkene biosynthesis , encoding stearoyl-acyl carrier protein ( ACP ) desaturases ( SAD ) , in three closely related Ophrys species , O . garganica , O . sphegodes , and O . exaltata . Combining floral odor and gene expression analyses , two SAD homologs ( SAD1/2 ) showed significant association with the production of ( Z ) -9- and ( Z ) -12-alkenes that were abundant in O . garganica and O . sphegodes , supporting previous biochemical data . In contrast , two other newly identified homologs ( SAD5/6 ) were significantly associated with ( Z ) -7-alkenes that were highly abundant only in O . exaltata . Both molecular evolutionary analyses and pollinator preference tests suggest that the alkenes associated with SAD1/2 and SAD5/6 are under pollinator-mediated divergent selection among species . The expression patterns of these genes in F1 hybrids indicate that species-specific expression differences in SAD1/2 are likely due to cis-regulation , while changes in SAD5/6 are likely due to trans-regulation . Taken together , we report a genetic mechanism for pollinator-mediated divergent selection that drives adaptive changes in floral alkene biosynthesis involved in reproductive isolation among Ophrys species .
Understanding the genetic basis of adaptation is of great interest to evolutionary biologists . For over a century , it has been debated whether adaptations are likely caused by a large number of mutations of small phenotypic effect or by very few genetic changes of large effect [1]–[3] . To address this question , it is necessary to identify the genetic basis of adaptive traits and their ecological significance in any given study system [4] . Pollinator-mediated selection on floral traits has been considered to be a major driving force of floral diversification and speciation in plants [4]–[9] . Closely related species featuring distinct floral traits , such as floral color , odor , or spur lengths , are widely thought to be a consequence of pollinator adaptation [4] , [6] , [8]–[11] . Furthermore , pollinator adaptation often conveys reproductive isolation [12] , [13] , and thus may directly contribute to the origin of novel species . Therefore , floral traits associated with pollinator adaptation are of special interest for the understanding plant speciation and evolution . Ophrys orchids mimic their pollinators' mating signals and are pollinated by male insects during mating attempts with the flower . This pollination by so-called sexual deception is very specific , and each orchid species only attracts one or very few insect species [14] , [15] . Specific pollinator attraction has been reported to be the main reproductive barrier in Ophrys [15]–[17] . The key to specific pollinator attraction is the chemical mimicry of the insect female's sex pheromone [11] , [18]–[22] , usually a blend of cuticular hydrocarbons , namely alkanes and alkenes . Among these , alkenes with different double-bond position are particularly important for the specificity of pollinator attraction [11] , [20] , [22] . Thus , genes specifying alkene double-bond positions may be directly associated with pollinator adaptation . In plants , alkene double-bonds are likely determined by desaturases such as stearoyl-ACP ( acyl carrier protein ) desaturases ( SADs ) [23]–[25] . At the onset of alkene biosynthesis , desaturases can insert a cis-double-bond into a saturated fatty acid ( FA ) intermediate , such as 16:0-ACP ( C:D denotes a fatty acyl group of C carbons length with D double-bonds ) , to produce an unsaturated FA such as 16:1ω-7-ACP or 16:1ω-9-ACP ( double-bond at position ω-7 or ω-9 , counting from the aliphatic end ) . Unsaturated FA intermediates are elongated from the ACP end [26] , leading to the production of ( Z ) -7- or 9-alkenes from ω-7 or ω-9 FA intermediates , respectively [23] , [24] ( Since all alkenes in this study are presumed to be in the cis ( Z ) configuration , only double-bond position will be indicated for alkenes in the following text ) . Therefore , changes in SAD enzyme activity or SAD gene expression may result in alkene double-bond position differences among species [25] . In this study , we focused on three closely related sympatric species of the genus Ophrys: O . garganica , O . sphegodes , and O . exaltata , among which the phylogenetic relationship could not yet be resolved by neutral sequence markers [27] . These three co-flowering species largely overlap in their geographical distributions and are similar in floral morphology [28] , [29] , however , they are reproductively isolated from each other due to the attraction of different pollinators [17] . Although the color of floral petals also varies among species , the key factor for differential pollinator attraction is floral scent [30] . Flowers of both O . garganica and O . sphegodes produce a high proportion of 9- and 12-alkenes ( with differences in carbon chain length ) , whereas O . exaltata produces high amounts of 7-alkenes . These 7-alkenes have previously been shown to be important for attraction of the bee , Colletes cunicularius , the pollinator of O . exaltata [19] , [31] . We have previously shown that in O . sphegodes , 9- and 12-alkene production is associated with the SAD2 gene , which encodes a functional desaturase [24] . However , the genetic basis of 7-alkene biosynthesis in O . exaltata and the driving forces for evolutionary divergence in alkene production among species are unknown . Here we investigate gene expression and evolutionary relationships among SAD gene family members in different natural orchid populations and species . Moreover , we discuss the biosynthetic origin of 7-alkenes as well as the role of pollinator-mediated selection in changing alkene composition among species . Specifically , we address the following questions: ( a ) Which desaturase is responsible for 7-alkene biosynthesis ? ( b ) What is the role of pollinator-mediated selection in the evolution of alkene production ? ( c ) How is the biosynthesis of different alkenes by different Ophrys species regulated genetically ?
We identified six SAD homologs , which we named SAD1–SAD6 . Three of these ( SAD1–3 ) have been described previously [24]; SAD4–6 were identified by homology from high-throughput transcriptome sequencing of O . sphegodes flowers . Phylogenetic analysis of plant SAD genes indicated that all six Ophrys homologs evolved via gene duplication , forming three distinct lineages , SAD1/2 , SAD3 , and SAD4/5/6 . The split of SAD1/2 from SAD4/5/6 was more recent than the split of these groups from SAD3 ( Figure 1 ) . To test for the signature of selection , a codon-based analysis comparing the rates of synonymous and non-synonymous mutations was performed . It revealed a significant signal indicative of positive ( ω9 and ω10 ) and relaxed purifying selection ( ω11 , all p<0 . 001 , Figure 1 ) after the split of SAD1 and SAD2 , concordant with previous findings [24] . However , no indication of positive selection was found for any other SAD locus or clade . Purifying selection significantly stronger than the background rate was found on the SAD3 branch , as well as prior to the split of SAD1/2 , and prior to the split of SAD4/5/6 . Among the six SAD homologs investigated for tissue- and floral developmental stage-specific expression , five ( SAD1–5 ) were found to be expressed in the 11 tested greenhouse-grown individuals . Four homologs ( all except SAD3 ) showed flower-specific expression ( Figure S1 ) . The expression levels of SAD2 and SAD5 were significantly associated with the presence of alkenes: SAD2 expression was significantly ( p<0 . 001 ) associated with both 12- and 9-alkenes , whereas SAD5 expression was significantly ( p<0 . 001 ) associated with 7-alkenes in O . exaltata across different tissues and floral developmental stages ( Figure 2 ) . All SAD homologs except SAD3 and SAD4 showed species-specific patterns of allelic variation among the three studied orchid species ( Figure S2 ) . Two allele groups were found for each SAD1 ( SAD1-A/B ) , SAD5 ( SAD5-A/B ) and SAD6 ( SAD6-A/B ) , whereas four allele groups were found for SAD2 ( SAD2-A/B/C/D ) . Among these SAD allele groups , biochemical activity assays suggest that SAD1-B and SAD2-C alleles do not encode functional desaturases , whereas one SAD2-A allele has been shown to be functional [24] . However , two further allele groups are unlikely to be functional: one group ( SAD2-D ) had a repetitive sequence insertion at the start of the coding sequence , and one ( SAD6-B ) contained significantly more stop and frame-shift mutations than expected by chance ( Table S1 ) . Combining putative coding sequence functionality and biochemical activity data , we classified all alleles into three categories: ( a ) putatively functional and expressed , ( b ) putatively nonfunctional and expressed , and ( c ) non-expressed alleles ( Figure S3 ) . For SAD1/2/5/6 , the distributions of these allele categories are significantly different between O . exaltata and the other two species ( Figure 3 ) . For SAD1/2 , functional expressed alleles were significantly more common in O . garganica and O . sphegodes than in O . exaltata . By contrast , SAD5/6 showed the opposite pattern , functional expressed alleles being more common in O . exaltata . To estimate FST values for all expressed SAD homologs in all three species , an in silico resampling approach was employed ( see Methods ) , treating all individuals as diploids based on flow cytometry data [17] . O . garganica was not included in this analysis due to the smaller sample size , and FST could not be calculated for SAD5 because it was only observed in O . exaltata . SAD2 and SAD6 showed significantly higher FST , ( 0 . 44±0 . 02 and 0 . 32±0 . 006 respectively , mean ± standard error , p<0 . 01 ) than SAD1 , SAD3 and SAD4 ( 0 . 22±0 . 005 , 0 . 25±0 . 006 and 0 . 23±0 . 01 respectively , mean ± standard error ) . Five SAD copies ( all except SAD3 ) showed divergent species-specific gene expression ( Figure S3 ) . Among the alleles of SAD1 , SAD1-A was highly expressed in O . sphegodes and O . garganica but not in O . exaltata ( Figure 4 ) , whereas SAD1-B was highly expressed in O . sphegodes and O . exaltata ( Figure S3 ) . Among the SAD2 allele groups , SAD2-A/B were highly expressed in both O . sphegodes and O . garganica , whereas the expression of SAD2-C/D was low in all study species ( Figure S3 ) . The expression of SAD4 was high in O . garganica and O . exaltata , but low in O . sphegodes . All alleles of SAD5 and SAD6 were highly expressed mostly in O . exaltata . Natural F1 hybrids among O . sphegodes and O . exaltata ( identified from AFLP data ) showed a similar scent [17] and SAD expression pattern to O . sphegodes ( Figure S3 ) . For SAD1 and SAD2 , alleles ( SAD1-A/B , SAD2-A/B ) that were most likely inherited from O . sphegodes were found to be highly expressed in these F1 hybrids , whereas none of the SAD5/6 alleles was expressed ( Figure S3 ) . Statistical analysis of data from natural populations showed a strong correlation between the expression of specific SAD allele groups and alkene production . The expression level of SAD1-A was found to be significantly ( p<0 . 001 ) positively correlated with 9-C29 , 12-C27 and 12-C29 alkenes; SAD2-A was significantly ( p<0 . 001 ) positively correlated with 9-C27 , 9-C29 , 12-C25 , 12-C27 , and 12-C29 alkenes ( Figure 2 , Figure S4 ) ; SAD5-A/B were significantly ( p<0 . 001 ) correlated with all 7-alkenes ( Figure 2 , Figure S4 ) . To understand the driving force for allelic evolution of SAD homologs in Ophrys , we tested the effects of alkenes with different double-bond position – associated with different SAD homologs – on pollinator behavior . We quantified pollinator responses to control and manipulated flowers and scent extracts of O . exaltata and O . sphegodes . Addition of 9- and 12-alkenes ( associated with SAD2-A ) to O . exaltata labella reduced the attractiveness to its pollinator ( the masked bee Colletes cunicularius ) by about 40% for approach and contact ( N = 18 , p = 0 . 023 and p = 0 . 015 respectively , Wilcoxon signed rank test , Figure 5A ) . Adding 7-alkenes ( associated with SAD5-A/B ) to floral scent extracts of O . sphegodes reduced the attractiveness to its pollinator ( the mining bee Andrena nigroaenea ) by about 30% and 60% for approach and contact respectively ( N = 17 , p = 0 . 028 and p = 0 . 045 , Wilcoxon signed rank test , Figure 5B ) .
Our data indicate that alkene biosynthesis is associated with the expression of certain SAD homologs in Ophrys . SAD2 catalyzes the introduction of a double-bond at position 9 of 18:0-ACP and position 4 of 16:0-ACP [24] , producing 18:1ω-9-ACP and 16:1ω-12-ACP , which should eventually lead to the production of 9- and 12-alkenes [23] , [24] . The significant correlation of SAD2 expression with the amounts of certain 9- and 12-alkenes in different plant tissues , floral developmental stages , and natural populations lends further support to this hypothesis . Although biochemical assays suggested that SAD1 ( allele group SAD1-B ) is not catalytically active [24] , this may not be true for other SAD1 alleles ( SAD1-A ) found in both O . sphegodes and O . garganica , SAD1-A and SAD1-B differing by 14% at the amino acid sequence level . Furthermore , SAD1-A expression was significantly correlated with some 9- and 12-alkenes ( Figure S4 ) . This indicates that , like SAD2 , SAD1 may also contribute to 9- and/or 12-alkene biosynthesis in natural populations . The significant correlation of SAD5 expression with the amount of 7-alkenes ( Figure 2 ) suggests that SAD5 is involved in 7-alkene biosynthesis . In addition , the high sequence identity ( >95% at the amino acid level ) of SAD5 to SAD6-A , which was highly expressed in one O . exaltata population ( SPF ) , indicates that both may have the same enzymatic function . We hypothesize that SAD5/6 introduces a double-bond at position 11 into 18:0-ACP , producing 18:1ω-7-ACP , or at position 9 of 16:0-ACP , producing 16:1ω-7-ACP . Further biochemical studies are required to test this hypothesis . These unsaturated intermediates could then be elongated to produce 7-alkenes . Therefore , changes in the expression of SAD5/6 could directly lead to different amounts of 7-alkenes in different Ophrys species . Our data suggest that SAD homologs evolve under pollinator-mediated selection , considering that genetic drift is a less likely explanation given large effective population sizes in our study species [11] . The expression of SAD1/2 was high in both O . garganica and O . sphegodes , but was very low in most O . exaltata individuals ( Figure 4 , Figure S3 ) . In those very few individuals of O . exaltata that did highly express SAD2 , it either was rendered nonfunctional by a repetitive sequence insertion ( SAD2-D; population MDL ) , or had amino acid substitutions located on the surface of the protein ( SAD2-C allele group ) , which have been suggested to reduce the activity of SAD2 [24] . This indicates that disruptive selection might act on SAD2 ( in terms of gene expression or overall enzymatic activity ) to maintain ( in O . sphegodes and O . garganica ) or reduce ( in O . exaltata ) the production of 9- and/or 12-alkenes in Ophrys floral odor . Evidence of positive selection on SAD1/2 detected by dN/dS ratio tests is consistent with this hypothesis ( Figure 1 and ref . [24] ) . Indeed , our pollinator-attraction assay suggests disruptive selection on 9- and/or 12-alkenes among species . This behavioral test demonstrated that adding 9- and 12-alkenes onto the floral labella of O . exaltata reduced its attractiveness to the pollinator by about 40% ( Figure 5A ) . This indicates that while 9/12-alkenes act as the main attractants of O . sphegodes to its pollinator A . nigroaenea [18] , these compounds actually reduce the attractiveness of the odor bouquet of O . exaltata to its pollinator C . cunicularius . Reduced responses of pollinator males to heterospecific odor blends may have evolved under sexual selection for maximum speed and accuracy of finding conspecific females [32] . Therefore , pollinator-imposed disruptive selection acts to change 9/12-alkene composition among these two Ophrys species by changing the expression or enzymatic activity of SAD1/2 . However , for SAD5/6 , the opposite pattern was observed . SAD5/6 were highly expressed in O . exaltata , but hardly expressed in O . sphegodes and O . garganica . A significantly higher frequency of frame-shifts or premature stop codons was found in SAD6 of O . sphegodes and O . garganica ( Table S1 ) , indicating that SAD6 alleles in these two species may be released from purifying selection such that loss-of-function mutations can accumulate . Neither positive selection nor purifying selection was detected on SAD5/6 using codon-based methods ( Figure 1 ) , suggesting that pollinator-mediated selection on 7-alkenes in O . sphegodes and O . garganica primarily acts on the expression level of SAD5/6 . Indeed , selection against 7-alkenes in O . sphegodes was confirmed by behavioral tests with its pollinator , since addition of 7-alkenes to the floral scent of O . sphegodes resulted in a significant reduction in pollinator attraction ( Figure 5B ) . Therefore , while 7-alkenes attract O . exaltata's pollinator [19] , [31] , they reduce the attractiveness of O . sphegodes to its pollinator . Hence , pollinator-mediated disruptive selection may also drive the evolution of 7-alkene quantity in these two Ophrys species by changing the expression of SAD5/6 . In contrast to these genes that are associated with alkene production , SAD3 and SAD4 , which were not significantly correlated with alkene occurrence , showed no significant sequence divergence among species ( Figure 3 , Figure S2 ) , as would be expected for genes that are not targets of selection . Overall , our data suggest that pollinator adaptation in Ophrys is achieved via reciprocal regulation or activity changes of desaturases involved in 7- and 9/12-alkene biosynthesis in response to disruptive selection by different pollinator preferences . The changes in 7- and 9/12-alkene production that are linked to differences in SAD gene expression may be explained by the action of cis- or trans-acting elements . The expression of SAD1/2 , which is associated with 9/12-alkene production , differed among O . exaltata ( weak expression ) and O . sphegodes ( strong expression ) ( Figure S3 ) . However , two putative F1 hybrids only expressed the alleles expected to be inherited from O . sphegodes , but not O . exaltata ( Table S2; Figure S3 ) . This indicates that down-regulation of SAD1/2 expression in O . exaltata might be due to changes in a cis-regulatory element ( such as a promoter or enhancer ) . In contrast , although differences in expression of SAD5/6 , which are associated with 7-alkene production , were found between O . sphegodes and O . exaltata , the putative F1 hybrids did not express either allele expected from the parental species ( Table S2; Figure S3 ) . This suppression of expression of SAD5/6 in F1 hybrids indicates that – while additional cis-regulatory changes cannot be ruled out – a trans-acting factor is likely involved in the different SAD5/6 gene expression among species . This suggests the presence of a ( dominant ) suppressor of SAD5/6 expression in O . sphegodes ( e . g . , a transcriptional repressor or a miRNA reducing SAD5/6 mRNA levels ) that is absent or inactive in O . exaltata . However , it is also possible that the dominant expression of SAD5/6 genes in F1 hybrids is due to epigenetic changes upon hybridization [33] , [34] . In conclusion , our data based on multiple independent lines of evidence suggest that pollinator adaptation in the three studied Ophrys species is largely due to changes in SAD1/2 and SAD5/6 , in terms of gene expression and potentially also in terms of the function of their gene products , and that both cis- and trans-regulation of gene expression contribute to this process . Our data indicate that pollinator adaptation in plants with a specialized pollination system may be due to few changes in the genome , with a large phenotypic effect . Furthermore , because reproductive isolation among closely related Ophrys species is mainly a consequence of specific pollinator attraction , such adaptation to different pollinators can directly prevent gene flow and ultimately lead to speciation .
Population samples were collected in southern Italy ( Table S3 ) , at the same locations as described in Xu et al . [17] , with three additional O . exaltata individuals from San Pietro in Fine ( SPF ) in southern Italy ( N41°25′38″ , E13°58′04″ ) . Two F1 hybrids between O . exaltata and O . sphegodes were previously identified based on AFLP markers [17] . For each plant individual , one labellum of an unpollinated flower was used for floral odor extraction as described previously [17] , and then immediately flash frozen in liquid nitrogen , and stored at −80°C until RNA extraction . For developmental stage- and tissue-specific analysis of hydrocarbons and gene expression , five O . exaltata and six O . sphegodes individuals grown in a greenhouse at the Botanical Garden of the University of Zürich were used and processed as described previously [24] . GC and GC/MS analysis , identification and quantification of compounds were performed as described previously [11] , with modifications [17] . Discrimination of 11- and 12-alkenes was not possible with the GC parameters used , however , 12-alkenes were earlier determined to be the predominant isomers in O . sphegodes [35] , [36] . The absolute amounts of alkenes and alkanes with a carbon chain length from 21 to 29 were calculated based on an internal standard . For tissue/stage-specific samples , the relative amount of alkenes was calculated since the use of comparable amounts of tissue could not be ensured . Total RNA was extracted using Trizol reagent ( Invitrogen , Carlsbad , USA ) following the manufacturer's protocol , and RNA quality and quantity were assessed by agarose gel electrophoresis and spectrophotometry on a NanoDrop ND-1000 ( Witec AG , Littau , Switzerland ) . First strand cDNA was synthesized as described in [24] . To obtain the full-length coding sequence of candidate genes , 5′ RACE was performed as described in [37] , with minor modifications [24] , and 3′ RACE as in [38] . Gene-specific primers used for RACE are listed in Table S4 . Advantage GC 2 DNA Polymerase ( Clontech Laboratories Inc , Mountain View , USA ) was used for RACE PCR with a touchdown program: 96°C 15 s; 3 cycles of [94°C 20 s , 68°C 3 min 30 s]; 7 cycles of [94°C 20 s , 67°C ( 1°C decrease per cycle ) 30 s , 68°C 3 min 30 s]; 30 cycles of [94°C 25 s , 55°C 30 s , 68°C 3 min 30 s]; final extension at 68°C for 10 min . The amplified fragments were cloned into pDRIVE vector ( Qiagen , Hilden , Germany ) , following the provided protocol . Desaturase homologs were amplified for all cDNA samples , using gene-specific primers containing attB adapter sequences ( Table S4 ) . RT-PCR was performed in 15 µl reaction volume containing cDNA template equivalent to 15 ng RNA as follows: 95°C 3 min; 33 cycles of [95°C 30 s , 58–60°C 30 s ( see Table S4 for annealing temperatures ) , 72°C 1 min 30 s]; final extension at 72°C for 10 min , using REDTaq ReadyMix ( Sigma-Aldrich , St . Louis , USA ) mix supplemented with 0 . 6 units Pfu polymerase ( Promega AG , Dübendorf , Switzerland ) . Three µl PCR product were loaded on an agarose gel to confirm amplification . Amplified PCR products from each population of each species were pooled and then purified with Wizard SV Gel and PCR Clean-Up kit ( Promega AG , Dübendorf , Switzerland ) , and recombined into Gateway cloning vector pDONR221 ( Invitrogen , Carlsbad , USA ) using the manufacturers' protocols . Competent E . coli One Shot TOP10 cells ( Invitrogen , Carlsbad , USA ) were used for transformation . In order to recover all possible alleles , the number of clones picked and screened by PCR was at least three times the number of possible alleles in diploids , for each cloning library . Clones were PCR amplified , purified and sequenced as previously described [24] . All sequences were deposited in GenBank ( accession numbers are listed in Table S5 ) . Forward and reverse sequences of each clone were assembled and manually edited in SeqMan v7 . 1 . 0 ( Lasergene DNAstar , Wisconsin , USA ) . For each SAD homolog , the assembled sequences of each clone were aligned using Clustal W [39] . Sequences with less than two nucleotide differences were considered to be the same allele with PCR or sequencing errors , and were merged into one consensus sequence . The consensus sequences and all singleton sequences , which differed by more than two nucleotides , were used for assignment to allele groups . To do so , given the low sequence divergence , a dendrogram was constructed for each SAD homolog in MEGA 4 . 0 [40] , using a pairwise distance and the UPGMA method , with pairwise deletion of gaps , and a homogeneous substitution pattern among lineages and sites . Bootstrap analysis was conducted using 1000 pseudo-replicates . Allele groups were assigned based on UPGMA tree topology with bootstrap support ( Figure S2 ) . The relationship among alleles was also inferred by Bayesian analysis ( Figure S5 ) in MrBayes ( v3 . 2 . 1; for details see below ) [41] , which was congruent and largely confirmed the clusters from UPGMA analysis . The Bayesian analysis showed SAD2-C alleles to be nested in SAD2-B , and SAD6-A in SAD6-B , but was in agreement with our primary allele group assignment . SAD gene expression was assessed by semi-quantitative RT-PCR with allele-specific primers ( Table S4 ) . PCR was performed in 10 µl reaction volume with cDNA from 12 ng total RNA as a template . Each PCR was performed as: 95°C 3 min; 29 cycles of [95°C 30 s , 58–60°C 30 s ( see Table S4 for different primer annealing temperatures ) , 72°C 1 min 30 s]; final extension at 72°C for 10 min using REDTaq ReadyMix ( Sigma-Aldrich , St . Louis , USA ) . For all RT-PCRs , the putative Ophrys housekeeping gene G3PDH [24] was used as control . Five µl of each PCR product were loaded on 0 . 8% agarose gel , recorded and quantified using ImageJ ( 1 . 42q ) [42] as described in [24] . Bioassays for pollinator preferences were performed between beginning of March to middle April 2011 at Capoiale ( Southern Italy ) and Charrat ( Wallis , Switzerland ) for O . exaltata and O . sphegodes , respectively . For both species , pollinator preferences on control and manipulated scent ( for O . exaltata , 9/12-alkenes added; for O . sphegodes , 7-alkenes added ) were tested ( Table S6 and Table S7 ) . The preference of O . exaltata's pollinator , C . cunicularius , was assessed with whole inflorescences ( bearing 2–3 flowers ) assayed individually . Each inflorescence was placed on bushes where C . cunicularius males were abundant; pollinator responses were recorded for 10 minutes . Afterwards , a 9/12-alkene mixture mimicking natural blends occurring in O . sphegodes was added onto each floral labellum of the same inflorescence ( see Table S6 and Table S7 ) and the pollinator responses monitored for a further 10 min . For each subsequent test , the plants were placed at a different position to avoid habituation effects often found after multiple subsequent testing at one location . In total , 18 inflorescences that had at least 2 flowers were tested . The preference of the pollinator of O . sphegodes , A . nigroaenea , was assessed in choice experiments on floral scent with black plastic beads . This different testing procedure was chosen because no natural plants were available at this testing location . The floral scent of floral labella was first extracted with 500 µl hexane [17] . For each choice experiment , 100 µl of floral extract was tested against 100 µl of floral extract of the same flower plus 7-alkene mixture ( see Table S6 and Table S7 ) . The dummy was placed on bushes where A . nigroaenea males were abundant . Pollinator responses were recorded for six minutes . In total , 17 replicates of this experiment were performed . For all pollinator behavioral tests , pollinator responses were classified into: approach ( a zig-zagging or undulating approach towards the tested flowers or beads ) and contact ( either a short punching contact or landing on the tested flowers or beads ) [19] . All monocot SAD sequences were taken from the plant SAD homolog data set of [24] . Sequences were re-aligned based on amino acid sequence using Muscle 3 . 8 . 31 [43] . Phylogenetic analysis was performed in MrBayes 3 . 1 . 2 [41] ( burn-in 13 out of 40 million generations ) using the GTR+I+G model , which was estimated to be the best model by MrModeltest ( 2 . 3; AIC criterion ) [44] . All sequence data were partitioned by codon positions , and MrBayes analysis performed using one cold and three heated chains , trees sampled every 1000 generations , and combined into a 50% majority rule consensus tree . The signature of selection on selected branches was tested using PAML 4 . 4 [45] , as previously described [24] . The significance of different amounts of floral odor and gene expression among species was assessed by ANOVA after normality testing of the data distribution by the Shapiro test [46] . Differences in allele distribution among different species were assessed by χ2-testing . To estimate the pattern of divergence for each SAD homolog , we first genotyped each Ophrys individual by allele-specific RT-PCR; then we randomly sampled two sequences from the allele groups based on species and population information . This procedure was repeated 100 times for in silico re-sampling . FST was calculated using Arlsumstat , a modified version of Arlequin ( v . 3 . 5 . 1 . 3 ) [47] . The association between floral scent and gene expression in natural populations was assessed using a generalized linear model ( GLM ) and a linear mixed-effect model ( LME ) with population as random factor . These models were simplified by stepwise removal of factors using the stepAIC method [48] . For the tissue/stage-specific dataset , the relative amount of each floral scent compound was used ( arcsine square-root transformed ) as described by Schlüter et al . [24] , since the size of floral labella varies in different developmental stages . For the population data set , the absolute amount of each floral scent compound was used . The significance of the presence of nonfunctional alleles in different allele groups was tested using Fisher's exact test . All statistical analyses were performed in R 2 . 11 . 0 [49] . | In plants , the extraordinary floral diversity has been suggested to be a consequence of divergent adaptation . However , the genetic basis of this process is poorly understood . In this study , we take advantage of the high specificity of plant-pollinator interactions in the sexually deceptive orchid genus Ophrys . We leverage the available , ample evidence showing that floral odors , especially alkenes , are the key factor for specific pollinator attraction in certain species of these orchids . Further , we investigate the genetic basis of pollinator adaptation . By applying an inter-disciplinary approach , including chemical ecology , gene expression analysis , population genetics , and pollinator-behavioral tests , we show that genetic changes in different copies of a biosynthetic gene are associated with the production of different floral scents and with pollinator adaptation in these orchid species . Moreover , we found that both cis- and trans-regulatory factors are likely involved in controlling gene expression of these biosynthetic gene copies . These findings support the hypothesis that adaptation is mediated by very few genetic changes with large phenotypic effects , rather than requiring a large number of co-adapted genes . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"plant",
"science",
"plant",
"evolution",
"plant",
"biology",
"gene",
"expression",
"genetics",
"plant",
"genetics",
"population",
"genetics",
"biology",
"evolutionary",
"biology",
"plant",
"ecology",
"evolutionary",
"genetics",
"genetics",
"and",
"genomics",
"plant-environment",
"interactions"
] | 2012 | The Genetic Basis of Pollinator Adaptation in a Sexually Deceptive Orchid |
Although the detection rate is decreasing , the proportion of new cases with WHO grade 2 disability ( G2D ) is increasing , creating concern among policy makers and the Brazilian government . This study aimed to identify spatial clustering of leprosy and classify high-risk areas in a major leprosy cluster using the SatScan method . Data were obtained including all leprosy cases diagnosed between January 2006 and December 2013 . In addition to the clinical variable , information was also gathered regarding the G2D of the patient at diagnosis and after treatment . The Scan Spatial statistic test , developed by Kulldorff e Nagarwalla , was used to identify spatial clustering and to measure the local risk ( Relative Risk—RR ) of leprosy . Maps considering these risks and their confidence intervals were constructed . A total of 434 cases were identified , including 188 ( 43 . 31% ) borderline leprosy and 101 ( 23 . 28% ) lepromatous leprosy cases . There was a predominance of males , with ages ranging from 15 to 59 years , and 51 patients ( 11 . 75% ) presented G2D . Two significant spatial clusters and three significant spatial-temporal clusters were also observed . The main spatial cluster ( p = 0 . 000 ) contained 90 census tracts , a population of approximately 58 , 438 inhabitants , detection rate of 22 . 6 cases per 100 , 000 people and RR of approximately 3 . 41 ( 95%CI = 2 . 721–4 . 267 ) . Regarding the spatial-temporal clusters , two clusters were observed , with RR ranging between 24 . 35 ( 95%CI = 11 . 133–52 . 984 ) and 15 . 24 ( 95%CI = 10 . 114–22 . 919 ) . These findings could contribute to improvements in policies and programming , aiming for the eradication of leprosy in Brazil . The Spatial Scan statistic test was found to be an interesting resource for health managers and healthcare professionals to map the vulnerability of areas in terms of leprosy transmission risk and areas of underreporting .
Leprosy is a chronic infectious-contagious disease caused by Mycobacterium Leprae , an obligate intracellular bacillus that affects the skin and the peripheral nervous system [1] . Leprosy is characterized as a slowly advancing disease and , due to the characteristics of the bacillus , presents high infectiousness and low pathogenicity [1] , being a potentially disabling and stigmatizing disease . For diagnosis and definition of the treatment regime with polychemotherapy ( PCT ) , two operational classifications are used that are based on the number of cutaneous lesions , according to the following criterion: Paucibacillary Cases ( PB ) with up to five skin lesions; Multibacillary Cases ( MB ) with more than five skin lesions [2] . The disease is part of the group of neglected diseases and is relevant for public health due to its magnitude and range , causing disabilities in low-income and economically active populations [3] . Among the infectious diseases , it causes the greatest number of permanent disabilities . The high burden of the disease signals the maintenance of the epidemiological chain of transmission , representing one of the most important epidemiological indicators [3] . Although approximately 30 years have passed since the introduction of multidrug therapy ( dapsone , rifampicin and clofazimine ) , prevalence and incidence rates remain considerable in the country . This demonstrates that other factors play a decisive role in its causal network , such as the biology of the etiological agent , the genetic or immunological characteristics of the host , and social and economic factors , such as precarious living conditions , migration , malnutrition and poverty , among others [4 , 5] . In 2014 , 31 , 000 cases were reported in Brazil , with a detection rate of 15 . 31 cases per 100 , 000 inhabitants and a prevalence rate of 1 . 56 cases per 10 , 000 inhabitants , putting the country second in the global ranking , only behind India [6] . Also in 2014 , Brazil presented a detection rate of cases with grade 2 disability of one case per 100 , 000 inhabitants [7] . It should be highlighted that Brazil is the only country in the Americas that has been unable to eliminate leprosy ( prevalence < 1 case per 10 , 000 inhabitants ) . Since 2000 , a movement has been ongoing in Brazil to reduce the burden of leprosy through strategies that are intended to expand the actions to the entire Health Care Network . These include early diagnosis and qualification of patient care , promoting the decentralizion of diagnosis , treatment and prevention actions to Primary Health Care ( PHC ) , the reorganization of services , disclosure regarding the characteristics , signs and symptoms of the disease and universal access [8] . Another strong point is the identification of the most problematic areas of the disease which , once identified , can be the target or focus of healthcare or intersectorial actions , considering the relationship with social determinants [5] . In a literature review that used the descriptors ‘spatial analysis’ AND ‘leprosy’ , a large number of articles related to the theme were retrieved , considering different branches or approaches to the clinical aspects of the disease [9] , the evolution of the treatment ( cure or abandonment ) [10] , disabilities and late diagnosis [11] . In general , the majority of studies aimed to provide a more exploratory description of leprosy in space , without considering the risk certain communities are exposed to in relation to others . Depending on the risk , emergency measures need to be adopted immediately to solve it or avoid the dissemination of the disease . Inferences regarding risk are an important tool for management , because they permit targets to be outlined and priority levels to be defined . Risk is traditionally focused on the quantification of the probability of negative consequences of one or more factors identified as harmful to health [12] . In view of the importance of advancing health policies in Brazil to eliminate leprosy and provide more solid research approaches to measure the risk or vulnerability in some communities , the aims established were to outline a case profile according to the operational classification of the disease and to identify areas of greater and lesser risk for the occurrence of leprosy .
A descriptive and ecological study was performed [13] . Ribeirão Preto is a city in the interior of the state of São Paulo ( Fig 1 ) , located at 47°48’24”W longitude and 21°10’42”S latitude , 314 Km from the state capital São Paulo and 697 Km from Brasília . The city has an area of approximately 650 Km2 and a high demographic density of 995 . 3 inhabitants per Km2 . The estimated population in 2010 corresponded to 647 , 862 inhabitants , 99 . 7% of whom lived in urban areas14 . Concerning the social and economic indicators , the city ranks in Group 2 of the São Paulo Social Responsibility Index ( IPRS ) , that is , with high levels of wealth but unsatisfactory social indicators [15] . The Municipal Human Development Index ( IDHM ) corresponds to 0 . 80 , the Poverty Index to 11 . 75% and the Gini Index to 0 . 45 [14 , 15] . Concerning the Health Care Network , the city is divided into five Health Districts ( DS ) , North , South , East , West and Central , with a total of 49 Primary Health Care services , including five District Primary Health Care Services ( UBDS ) , 18 Family Health Services ( USF ) and 26 Primary Health Care Services ( UBS ) , with a total coverage of 22 . 27% of the population of Ribeirão Preto [16] . The hospital network includes 15 institutions , including one University Hospital , nine hospitals affiliated with the public network of the Brazilian National Health System ( SUS ) and five non-affiliated hospitals [16] . Regarding leprosy , the entire service is concentrated in three clinics: Centro de Referência em Especialidade Central , Centro de Referência José Roberto Campi and CSE Sumarezinho [17] . The study population consisted of cases of leprosy diagnosed by the health services between January 1st 2006 and December 31st 2013 . In the phase of exploratory analysis of the data , initially , the descriptive analysis of the data was performed using the Statistica version 12 . 0 software , calculating central tendency measures for the continuous variables and absolute and relative frequencies for the categorical variables . The continuous variable age was categorized . To analyze the profile of the leprosy cases according to the clinical forms , the dependent variable ( operational classification of PB or MB leprosy ) was crossed with the independent variables ( age , gender , ethnic origin , education , assessment of physical disability at diagnosis , assessment of physical disability at cure , number of contacts examined ) , applying the chi-square association test with Yates’ correction or Fisher’s exact test . The probability of type I error was set at 5% . To detect the risk for spatial and spatial-temporal clusters of the leprosy cases , the cases were geocoded , using the TerraView version 4 . 2 . 2 software , standardizing and equalizing the addresses of the resident cases in the urban zone of the city with the StreetBase digital address map in UTM projection—Zone 23S/WGS1984 available in the Shapefile extension , purchased from the Imagem Soluções de Inteligência Geográfica company . In this phase , cases with blank or incomplete addresses were ignored , as were cases in rural areas and cases in which the address was the municipal prison . Geocoding was obtained through the linear interpolation of the full address , including the zipcode , with a point in a corresponding address segment , which permitted patterns of event points to be set up . In addition , for the registers not located in the cartographic database , the Google Earth open access tool was used , in which the Geographic Coordinates of the addresses ( latitude and longitude ) were found . It should be highlighted that the census sectors were used as the ecological analysis units , with the advantage of being the most disaggregated level of population and socioeconomic groups , collected systematically , periodically and according to a national standard [18] . The cartographic database of the census sectors in Ribeirão Preto was obtained from the website of the Brazilian Institute of Geography and Statistics ( IBGE ) , which consists of 1004 census sectors , of which only 988 were considered due to representing to the urban zone of the city and presenting a resident population [19] . Next , the scanning spatial analysis technique was used , which was developed by Kulldorff and Nagarwalla [20] to detect clusters in space and in space and time . The search for clusters is made by placing a circle with the radius of the variable around each centroid and calculating the number of events within the circle . If the coefficient observed for the region delimited by the circle , called region z , is higher than expected , the circle is called a cluster . This procedure is repeated until all centroids have been tested [21] . Based on this situation , the hypotheses formally elaborated to detect clusters were H0: there are no clusters in the region of Ribeirão Preto and H1: region z is a cluster . To identify essentially spatial clusters , the SaTScan 9 . 4 software was used and , as the events studies ( leprosy cases ) were counts and rare in relation to the population , Poisson’s discrete model was used . Thus , the following conditions were adopted: no geographical overlapping of the clusters , maximum cluster size equal to 50% of the population exposed , circular cluster and 999 replications . It should be highlighted that , in this phase , the information on the year of occurrence of the event was not used . As well as permitting the spatial analysis , the scanning statistics also permitted the incorporation of the temporal factor , in which the identification of clusters of events [22] simultaneously in space and time is of interest . Thus , the SaTScan 9 . 4 software was also used to detect spatial-temporal clusters , under the same conditions as defined above for the spatial clusters , however , considering the maximum size of the temporal cluster as equal to 50% of the study period , the precise time , as day , month , year , and the time period between 2006 and 2013 . In addition , the spatial and spatial-temporal scanning techniques were processed , controlling for the occurrence of cases by population size of the census sectors , by their age distribution and according to gender , as well as attempts to detect high and low relative risk ( RR ) . The relative risk allows information from distinct areas to be compared , standardizing it and removing the effect of the different populations , therefore showing how intensely a certain phenomenon occurs in relation to all other study regions [12 , 13] . A p-value <0 . 05 was adopted for statistically significant clusters . Clusters that considered only one census sector and presented zero cases of leprosy were ignored . Furthermore , thematic maps were constructed from the scanning analyses , containing the RR of the clusters obtained using the ArcGIS 10 . 1 software . Approval for the study was obtained from the Research Ethics Committee of the Ribeirão Preto College of Nursing , with Evaluation No . ( CAAE ) 44637215 . 0 . 0000 . 5393 . Signing of a consent form was not necessary as secondary data were used and the participants were not identified .
In total , 434 cases of leprosy were identified , with a predominance of males ( n = 264; 60 . 83% ) between 15 and 59 years of age ( n = 297; 68 . 43% ) . Concerning education , 244 ( 56 . 22% ) subjects had complete or incomplete elementary education and 10 ( 2 . 30% ) had not attended school . Regarding ethnicity , 237 ( 54 . 61% ) subjects referred to themselves as white and 118 ( 27 . 19% ) mixed race . Considering the clinical form of the disease , it was observed that 188 ( 43 . 31% ) presented the dimorphic form and 101 ( 23 . 28% ) the lepromatous form . In Table 1 , analyzing the crossing between the PB or MB operational classification and the independent study variables , a statistically significant association ( p = 0 . 013 ) was found for gender , with higher proportions of MB cases in males ( n = 197; 45 . 39% ) . The physical disability at diagnosis variable showed a statistically significant association ( p = 0 . 020 ) , with greater proportions of the MB cases with disabilities 1 and 2 ( n = 166; 44 . 04% ) , when compared to the PB cases . At the end of the treatment , when the disabilities were assessed , a statistically significant association was again observed ( p = 0 . 000 ) , with greater proportions of disabilities 1 and 2 in the MB cases ( n = 70; 24 . 82% ) .
The aim of the study was to outline the case profile according to the operational classification of leprosy and to identify areas of greater and lesser risk for the occurrence of the disease . A statistically significant association was observed for the gender variable , with a higher occurrence of MB forms among men; among the MB forms , more disabilities were also identified at the moments of diagnosis and cure . Regarding age , it was observed that the subjects most affected by the MB form were between the ages of 15 and 59 years , a phenomenon that has been observed in other contexts in Brazil [23] . This age range refers to the economically active age ranges of the population , in which the health services should concentrate on preventive measures , with the diagnosis , treatment and identification of the appearance and development of lesions , disabilities and reactive conditions . Early intervention avoids or minimizes the high social cost the disease provokes , due to the withdrawal of this population from productive activities . The significant predominance of male cases in the MB form can be related to the physiopathological mechanisms of the disease . However , this aspect has not been explored nor well clarified in the scientific literature . Other studies have shown that males are more vulnerable , presenting more severe and disabling clinical forms , at the moment of the diagnosis as well as at the time of completing the medication regime [24–26] . Considering females , in the last two decades , there has been a decline in the number of women with the more severe clinical forms , which strengthens the hypothesis raised . However , the literature shows that women tend to visit health services earlier and attend more regularly [27] , which should also be considered in the causal line of the disease . Concerning the educational level of the subjects , despite the lack of a statistically significant association , the highest proportions of MB cases presented elementary education , in accordance with findings of other studies [28] . A case-control study by Kerr-Pontes [28] suggested that low education acts as a risk factor for the transmission of leprosy . In a study conducted in a city in the state of Tocantins—Brazil , it was suggested that low education is associated with the development of physical disabilities [23] . The number of disabilities at the moment of diagnosis and at the end of the treatment among MB cases was also highlighted in the study , as an aspect that should be considered in the planning and organization of the health services . Studies indicate that MB patients present 5 . 7 times greater chance of presenting disabilities , mainly at the end of treatment , when compared to PB patients [29] . The result raises the hypothesis that MB cases are diagnosed later , with a higher grade of neural commitment , which can favor the development of disabilities , closely linked to the time factor [30] . The disabilities reflect the late diagnosis and the quality of care delivered in the context of the health services . If the disability staging of the patient evolves , the technologies the health agents use are , most probably , not suitable or in accordance with the needs of the patient . In Brazil [2] , it is recommended that the health services , especially Primary Health Care , assess and determine the grade of disability of the leprosy patient at the moment of diagnosis , at least once per year during the treatment and at the end of treatment , identifying and preventing physical deformities as early as possible . The grade of physical disability can be attributed not only to late diagnosis , but also to neuropathies , to treatment irregularities , mainly related to the administration of MDT , and to self-care advice , such as the use of eyewashes , moisturizers , maintenance of domestic appliances and clothing [23] . The present study also evidenced the predominance of the dimorphic and lepromatous clinical forms , a phenomenon also demonstrated in other research scenarios [31–35] . The transmissible clinical forms , dimorphic and lepromatous , due to their high bacillary load , can be highly disabling and stigmatizing , being the main reservoir of the disease [2 , 8] . Hence , the large proportion of new cases of these clinical forms indicates errors in the active detection of cases and in the search for communicants [31 , 36] . Through the scanning statistics , risk areas can be stratified and census sectors identified that are inclined toward the establishment of statistically significant clusters for leprosy , in space as well as in space-time . This method permits the spatial distribution of cases to be understood , testing whether the pattern observed is random , regularly distributed or clustered . It also permits the existence of possible environmental factors and the extent of the infection risk to be identified [37] . The high-risk clusters , in the spatial as well as the spatial-temporal analysis , included the West , North and Central Health Districts , which comparatively showed great similarities in the socioeconomic profile and in the occupation profile of their populations . The North Health District had the lowest social indicators among the five health districts studied , with the highest percentage of people earning less than one minimum wage , the lowest Gross School Attendance Rate ( TBFE ) , the lowest Municipal Human Development Index in Education ( MHDI-E ) and the largest number of subnormal clusters ( communities ) in the city [38] . In the West District , the health network was more complex , having the largest number of SUS health services , with the second highest percentage of exclusive users . A high percentage of residential housing was identified , with a considerable number of residents per household and a predominance of households with incomes between one and five minimum wages . The Central Health District , in turn , consisted of the oldest neighborhoods of the city , with a traditional service network , without coverage of the Family Health Strategy ( FHS ) , with the main referral center for the treatment of leprosy cases situated there [38] . The district presented a significant percentage of people over 60 years of age and a high burden of chronic health conditions among its inhabitants . Concerning the low-risk areas identified in the study , it was noticed that the East and South Health Districts corresponded to the protection areas in the spatial and spatial-temporal analyses . The East and South Districts were both characterized by good TBFE and MHDI-E rates and a low rate of population without income . It should be highlighted that the East District possessed the highest percentage of families earning 50 or more minimum wages and the lowest percentage of exclusive SUS users . The low-risk areas in the city should be highlighted and analyzed with caution in terms of the spatial behavior of leprosy , as the representation of these clusters can evidence errors in the detection of causes by the health services in the “protection” areas for leprosy . This can result in underreporting and/or late diagnosis , serving as an alert for the need to intensify active search actions in order to detect a larger number of cases earlier in these regions [39] . The risk clusters estimated by the scanning statistic revealed the focal and unequal behavior of leprosy among the regions of the city , showing that neighborhoods in the North , West and Central Districts presented populations at high risk of catching the disease . The relationship between the disease and these regions can be associated with the fact that they include the neighborhoods with the greatest social inequalities [35 , 38 , 40] , with precarious housing , many residents per household , low income and low education , which are factors that favor the dissemination of leprosy [37] . In this context , studies demonstrate that populations exposed to social vulnerabilities , such as unhealthy housing conditions , basic sanitation deficits , low per capita income , irregular occupation of places unsuitable for habitation and many people sharing the same house , represent the main factors for leprosy [18 , 31 , 34 , 36 , 41 , 42 , 43] , which , in terms of spatial distribution , presents the focal behavior the disease , as evidenced in earlier studies [1 , 35 , 44 , 45 , 46 , 47 , 48] . It should be highlighted that places with many people living in the same house are the main source of maintenance of the leprosy transmission chain and form the domestic contacts . People living near individuals with leprosy and their social contacts have a greater risk of infection [49] . Frequent clusters are not only associated with lower socioeconomic levels and populational clusters , but are also related to shortages in health services [31] . Concerning the constitution of the care network for leprosy patients , the city follows a centralized and verticalized model , in which municipal referral centers serve as the main access for the entry and follow-up of cases . The municipal care flowchart recommends that , when primary health care services identify a suspected case of the disease , they should forward it to the municipal referral centers to confirm the diagnostic hypothesis and monitoring of the case . Although the current strategies for the control and elimination of the disease have provided positive results , mainly in the last three decades , they are still considered insufficient to eliminate the disease . The condition identified as essential to achieve the target of elimination of the disease , as proposed by the WHO , is the increased supply of health services through the decentralization of control actions in the cities and the inclusion of leprosy treatment in the Primary Health Care services , mainly in areas of social inequality [9] . The strengthening of PHC , with improved access to the health services , the earlier and more effective detection of leprosy cases , the active search for communicants by Community Health Agents’ ( ACS ) and the free distribution of MDT are essential strategic actions to eliminate the disease [45] . However , other measures attributed to PHC should be highlighted . The community’s understanding of leprosy ( its transmission mechanisms , clinical manifestations and treatment ) should be widely discussed and valued , in order to empower the population and reduce the stigma the disease causes , which is still one of the main factors that impede the elimination of leprosy [50] . The advance towards the elimination of leprosy should be based on contact surveillance , on the prevention of multidrug resistance , on the prevention and rehabilitation of physical disabilities and on the reduction of the stigma associated with the disease [5] . The results of this study suggest that the distribution of leprosy is restricted to spaces where a number of factors coincide for its production , including environmental , individual , socioeconomic and health service organization factors . In the last two decades , spatial analysis has been frequently used as a leprosy control tool in Brazil and in countries with a high prevalence of the disease [44 , 51] . According to the WHO , this is an effective management tool for disease elimination programs , with its use being recommended in all endemic countries [52] . The identification of spatial clusters through an ecological study is a strategic tool to enhance the understanding about the distribution of a disease and a health resource allocation tool [37] . Being a neglected endemic disease and a severe public health problem , knowledge about the spatial distribution of leprosy and identification of risk areas for the disease are fundamental for its control , being important measures to improve the surveillance actions in certain locations [45] . This study only considered the secondary data available , therefore there might have been underreporting of cases in the areas with lower spatial risk . These findings contribute to the identification of risk areas for leprosy , presenting elements to consider in the organization and strengthening of the health services in these places in terms of the active search for cases . Considering the early diagnosis and the empowerment of this population , awareness about leprosy needs to be increased and people encouraged to work in partnership with the health services to deal with the problem . Furthermore , social institutions within these areas , such as churches , kindergartens and schools , should be mobilized to work jointly with the health services , as they are located in risk areas . The study permits a reflection about the health services in these clusters and the mechanisms or technologies these services use to provide access for leprosy patients . Due to the number of grade 2 disabilities found , advances are needed , not only in early diagnosis , but also in the management of cases throughout the treatment , preventing patients from becoming worse after the diagnosis . Therefore , another possibility is the introduction of self-care workshops , preparing patients and families to cope with the disease , with the taboos , prejudices and stigma , and demonstrating how to avoid disabling accidents in the home and work environments , permitting their physical and social rehabilitation . In view of the above , the study contributes to the advance of knowledge in this area and to advances towards the eradication of leprosy in Brazil . | Brazil has still not achieved the goal of leprosy elimination established by the World Health Organization . The diagnosis and treatment of leprosy are available and the country is striving to fully integrate leprosy services into the existing general health services . Access to information , diagnosis and treatment with multidrug therapy ( MDT ) remain key elements in the strategy to eliminate the disease as a public health problem , defined as reaching a prevalence of less than 1 leprosy case per 10 , 000 inhabitants . Thus , this study aimed to identify spatial clustering of leprosy and to classify high-risk areas in a major leprosy cluster . A total of 434 cases were identified , with 188 ( 43 . 31% ) being of borderline leprosy and 101 ( 23 . 28% ) lepromatous leprosy . There was a predominance of males , with ages ranging from 15 to 59 years , and 51 patients ( 11 . 75% ) presented G2D . Two significant spatial clusters and three significant spatial-temporal clusters were also observed . These results can assist health services and policy makers to improve the health conditions of the Brazilian population , advancing towards the goal of elimination in Brazil . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"and",
"health",
"sciences",
"disabilities",
"tropical",
"diseases",
"geographical",
"locations",
"census",
"geoinformatics",
"health",
"care",
"bacterial",
"diseases",
"research",
"design",
"neglected",
"tropical",
"diseases",
"spatial",
"analysis",
"research",
"and",
"analysis",
"methods",
"public",
"and",
"occupational",
"health",
"infectious",
"diseases",
"computer",
"and",
"information",
"sciences",
"geography",
"south",
"america",
"health",
"systems",
"strengthening",
"brazil",
"people",
"and",
"places",
"socioeconomic",
"aspects",
"of",
"health",
"survey",
"research",
"earth",
"sciences",
"leprosy",
"health",
"care",
"policy"
] | 2017 | Spatial clustering and local risk of leprosy in São Paulo, Brazil |
Rates of evolution differ widely among proteins , but the causes and consequences of such differences remain under debate . With the advent of high-throughput functional genomics , it is now possible to rigorously assess the genomic correlates of protein evolutionary rate . However , dissecting the correlations among evolutionary rate and these genomic features remains a major challenge . Here , we use an integrated probabilistic modeling approach to study genomic correlates of protein evolutionary rate in Saccharomyces cerevisiae . We measure and rank degrees of association between ( i ) an approximate measure of protein evolutionary rate with high genome coverage , and ( ii ) a diverse list of protein properties ( sequence , structural , functional , network , and phenotypic ) . We observe , among many statistically significant correlations , that slowly evolving proteins tend to be regulated by more transcription factors , deficient in predicted structural disorder , involved in characteristic biological functions ( such as translation ) , biased in amino acid composition , and are generally more abundant , more essential , and enriched for interaction partners . Many of these results are in agreement with recent studies . In addition , we assess information contribution of different subsets of these protein properties in the task of predicting slowly evolving proteins . We employ a logistic regression model on binned data that is able to account for intercorrelation , non-linearity , and heterogeneity within features . Our model considers features both individually and in natural ensembles ( “meta-features” ) in order to assess joint information contribution and degree of contribution independence . Meta-features based on protein abundance and amino acid composition make strong , partially independent contributions to the task of predicting slowly evolving proteins; other meta-features make additional minor contributions . The combination of all meta-features yields predictions comparable to those based on paired species comparisons , and approaching the predictive limit of optimal lineage-insensitive features . Our integrated assessment framework can be readily extended to other correlational analyses at the genome scale .
Different proteins evolve at drastically different rates [1] . Some proteins are highly conserved across distantly diverged species , such as the ribosomal and histone proteins in eukaryotes [2] . Other proteins evolve much more quickly , often to the point where they occur in one species but cannot be identified in other closely related species , possibly due to deletion or major sequence divergence [3] . What are the main driving forces of such differences in protein evolutionary rate ? What percentage of this variation can be attributed to simple protein properties that we can quantitatively measure in a genome-wide fashion ? The answers to such questions are critical to achieving a systematic understanding of molecular evolution . With the advent of reliable high-throughput functional genomic measurements , particularly in the model organism Saccharomyces cerevisiae ( baker's yeast ) , it is now possible to rigorously assess the functional genomic correlates of protein evolutionary rate . Many studies have focused on calculating the correlation between protein evolutionary rate and a single protein feature that can be determined for a large fraction of yeast proteins , followed by statistical hypothesis testing of the observed correlation . This method has been successful in identifying a number of key correlates of protein evolutionary rate , such as protein abundance [4] , essentiality [5] , and number of interactors [6] . For further review of individual correlates , see [1] , [7] , [8] . Assessing the relative strengths , synergistic effects , and redundancy among such correlations requires more sophisticated statistical methods . Multivariate techniques have already been applied in a number of studies aimed at simultaneously dissecting multiple correlates of evolutionary rate [9]–[16] . Partial correlation and principle component regression , two popular techniques in this area , have been shown to produce discrepant results when applied to similar data [13] , [15] , [17] . Arguments have been made against both techniques regarding their sensitivity to noise among protein features and a tendency to over- or under-estimate the number of independent determinants of evolutionary rate . Analyses of evolutionary rate correlation have been historically limited by less-than-complete coverage of the genome—often far less . Consider the calculation of evolutionary rate itself . A commonly used reference dataset was produced by Wall et al . , where a set of evolutionary rate calculations were meticulously performed in yeast [12] . In generating these data , they placed demands on gene orthology and phylogenetic relationships that substantially reduced genome coverage ( to roughly 3 , 000 genes , relative to the roughly 6 , 000 open reading frames in S . cerevisiae ) . Moreover , a reduction in genome coverage may be accompanied by the introduction of specific biases . For example , stringent demands on gene orthology automatically bias a dataset toward more slowly evolving proteins . Coverage will tend to be further limited—and the dataset further biased—as more genomes and more protein features are added to an analysis . In this study , we used an integrated probabilistic modeling approach to assess genomic correlates of protein evolutionary rate for 5 , 537 proteins in the yeast genome ( 94 . 5% coverage relative to the 5 , 861 total yeast ORFs ) . We assembled a list of diverse protein sequence , physicochemical , and functional genomic features with high coverage of the proteins in S . cerevisiae and assessed their correlations with an approximate , high coverage measure of protein evolutionary rate . To manage potential outliers , noise , and non-linear relationships , we employed robust measures of correlation , such as rank correlation and mutual information . By considering many protein attributes simultaneously , it was possible to rank them according to their degrees of association with evolutionary rate . Our high-coverage framework allows us to re-assess known genomic correlates of evolutionary rate , while simultaneously identifying new , statistically significant correlates . In addition , we employed a logistic regression framework on binned data to assess the information contribution of sets of features in the task of predicting slowly evolving proteins . Our framework is flexible and robust , and is able to account for intercorrelation , non-linearity , and heterogeneity within features . Using this framework , we were able to group overlapping and interrelated features into natural ensembles ( “meta-features” ) , and quantitatively assess their combined predictive power . Next , natural ensembles were evaluated in progressively larger groups to measure the independent significance of their contributions . Finally , we show that our optimal predictions of S . cerevisiae protein evolutionary rate are comparable to those based on paired species comparisons , and approaching those based on the lineage-independent component of evolutionary rate .
We employed an approximate , high coverage method for calculating yeast protein evolutionary rate based on multiple paired species comparisons . Figure 1 illustrates this procedure , which we outline here briefly ( see the Methods section for further details ) . We selected five closely related yeasts for evolutionary comparison . 5 , 537 proteins in S . cerevisiae possessed an annotated ortholog in at least one of these species . Evolutionary rates were calculated for pairs of orthologous sequences following previously established procedures ( e . g . , codon alignment followed by dN/dS calculation ) . These rates were then ranked and normalized within a given paired species comparison . The evolutionary rate of a given protein is the average of its ranked , normalized rates across all paired species comparisons in which an ortholog was present . We operate under the initial assumption that the ranked evolutionary rate of a protein is constant over time , and should therefore be approximately equal when estimated using different yeast species pairs . Averaging over multiple paired species comparisons involving S . cerevisiae serves to enhance the signal-to-noise ratio for evolutionary rate along the S . cerevisiae lineage . As our measure of evolutionary rate is approximate , we rely on relative rankings and binning to limit the influence of error . Our rankings were tested in comparison with the well established Wall et al . dataset of yeast dN/dS values [12] . Where the datasets overlap , the correlation between our ranks of evolutionary rate and those inferred from Wall et al . is high at 0 . 938 ( i . e . , 88 . 0% of the variation in the Wall et al . rankings can be explained by our approximate method ) . The advantage of our method is a substantial increase in genome coverage: Wall et al . assigned evolutionary rates to 3 , 038 proteins ( 51 . 8% genome coverage ) , while we assign ranks to 5 , 537 proteins ( 94 . 5% genome coverage ) . We collected a list of 42 high coverage protein sequence , structure , and functional genomic attributes that potentially correlate with evolutionary rate ( Table 1 ) . We ranked these features according to their absolute rank correlation coefficients with evolutionary rate . The top twenty correlates are listed in Table 2 . Three categorical variables ( GO slim biological process , molecular function , and cellular compartment ) were excluded from this analysis as correlation coefficients cannot be computed for categorical variables . The most dominant genomic correlates of evolutionary rate are those associated with protein abundance ( e . g . , codon bias and absolute mRNA expression , both correlating negatively ) and a subset of amino acid composition ( serine and asparagine content correlating positively , and glycine , alanine , and valine content correlating negatively ) . Other significant correlates include: native disorder , GC content , number of interactors , degree of gene duplication , essentiality , and—reported here for the first time—number of transcriptional regulators . In addition , to deal with categorical variables and potential non-linear relationships between genomic features and evolutionary rate , we converted continuous variables into discrete variables through binning and then ranked genomic features according to their mutual information with evolutionary rate . The top twenty correlates under this scheme are listed in Table 3 . The resulting order is similar to that produced by the rank correlation analysis , except that broad functional assignment ( GO slim molecular function , biological process , and cellular compartment ) joins protein abundance and amino acid composition as a dominant genomic correlate of evolutionary rate . Statistical significance was determined from the distribution of correlation measures resulting from 100 randomizations of the feature data annotations . All correlations discussed here and listed in Tables 2 and 3 are highly statistically significant ( rank correlation z-scores>6 , mutual information z-scores>40; all p-values≪0 . 001 ) . We selected the slowest evolving 20% of the proteins and asked which features best distinguish them from the remainder of the genome using a fold enrichment analysis ( Figure 2 ) . In agreement with previous studies [4]–[6] , [13] , [18]–[22] , we found that slowly evolving proteins tend to be more abundant , be essential , have many gene duplications , and have more interaction partners . In addition , slowly evolving proteins are overrepresented in certain biological processes ( such as translation ) and depleted in others ( such as protein modification ) . Similar arguments can be made for molecular function: slowly evolving proteins are common among structural molecules , but rare among transcriptional regulators . Slowly evolving proteins also tend to have low predicted native disorder [23] , and have characteristic amino acid compositions [24] . We again observe a new , significant correlation between evolutionary rate and transcriptional regulation: slowly evolving proteins tend to have more transcriptional regulators . The correlation between protein evolutionary rate and number of predicted transmembrane helices is low , as are the correlations with predicted secondary structure features ( not depicted ) . This may be in keeping with recent findings at the interface of protein structure and evolution: while structural characteristics impose clear constraints at the residue level , these constraints do not always scale to the level of whole proteins in a straightforward manner [14] , [25] . The methods described so far have helped us to overcome some of the difficulties inherent to analyzing the relationships between protein feature data and evolutionary rate . We have been able to rank the importance of the various features in a robust statistical framework with full genome coverage , while compensating for non-linear relationships , mixed data types , and to some extent noise . We next sought to address two additional issues—joint information contribution and contribution independence—without losing the gains that our approach had already made . In order to accomplish this , we applied a logistic regression model to study the information contribution of features ( and sets of features ) in the task of predicting slowly evolving proteins . Logistic regression has been used in the past for predicting protein-protein interactions [26] , [27] . It is capable of integrating discrete and continuous data to model non-linear relationships ( through binning ) , and is robust against redundancy among features . This last advantage makes logistic regression particularly powerful for simultaneously modeling groups of features , a prerequisite for our next objectives . We constructed a positive dataset consisting of the slowest evolving 20% of the proteins , and a negative dataset of the same size consisting of a random sampling of the remaining 80% of the proteins . The positive and negative datasets were then divided into five partitions . Using five-fold cross validation , we trained a logistic regression classifier using four of the five partitions , and then evaluated our model using the remaining partition as a test set . Results take the form of correct classification rates—i . e . , when evaluating the model using the test data , the percentage of the proteins that were correctly assigned to their respective classes ( slowly evolving versus not slowly evolving ) . Since all datasets were balanced prior to training and testing , a random classifier would produce correct predictions 50% of the time . This is a lower bound to which the feature-based classifications can be compared . Mathematical details of the logistic regression procedure can be found in the Methods section . The top panel of Figure 3 reports the correct classification rates for a sampling of single protein features . As with the previous methods , we are able to rank features according to the strength of their relationships with evolutionary rate . Note that while some features are closely related ( under the umbrella of “amino acid composition , ” for example ) , they may differ in their degrees of predictive power . We now turn to the major advantage of the logistic regression approach: the ability to consider sets of features simultaneously . In some cases , the protein features that we consider serve as proxies to some well-defined ( but difficult-to-measure ) property of a protein . If there are multiple proxies for a general protein property , then we expect them to be highly redundant . Other features are important in their own right , but are more tractable when considered together ( e . g . , amino acid composition ) . Features like these may also possess hidden interdependencies that we would like to model . We address these issues by grouping related features into natural ensembles , which we call meta-features . The logistic regression classifier can be trained and tested based on a meta-feature in order to assess the joint information contribution of its constituent features . Working with meta-features has several advantages: ( i ) it compensates for redundancy and interrelations among features , ( ii ) it averages out noise present in individual features , and ( iii ) it summarizes the many individual features into a handful of highly relevant general protein properties . Note that in all analyses based on subsets of features ( including meta-feature analyses ) , we rely on prior expert knowledge to form the subsets , as there are too many possible subsets to exhaustively enumerate . We pooled our 42 individual features into seven natural meta-features and then evaluated the information contribution of each using the logistic regression model ( see Figure 3 , bottom panel , light green bars ) . The meta-feature groupings are detailed in Table 1 . The phenotype meta-feature is the poorest predictor of slowly evolving proteins , producing correct classifications only 58 . 5% of the time; this is likely due in part to the difficulty and noise associated with measuring phenotypic information . The network meta-feature provides a reasonably improved 63 . 1% correct classification rate; it too is likely to suffer from experimental noise . Genomic properties , structural properties , and functional annotations yield progressively improved rates of 64 . 0% , 68 . 0% , and 68 . 9% , respectively . Abundance , previously implicated as the single dominant factor in determining a protein's evolutionary rate [13] , produces the best correct classification rate of any single meta-feature , 78 . 8% . Amino acid composition falls in second place with 74 . 8% correct classifications . Note how the individual features of the top panel compare to their related meta-features below . The meta-feature scores are always better than those of their constituent features . Some meta-features , such as abundance , have only one dominant dimension . In these cases , the component features make similar and largely overlapping contributions to predictive power , and the integration produces a minor increase in correct classification rate mainly due to noise reduction . Other meta-features , such as structure , have multiple intrinsic dimensions . In these cases , the component features make partially independent contributions to predictive power , and the integration produces a larger increase in correct classification rate . Meta-features can be further grouped in order to test the independence of their contributions ( see Figure 3 , bottom panel , dark bars ) . When amino acid composition and abundance are grouped , we achieve a slight gain in predictive power relative to the individual meta-features ( 80 . 6% correct classifications ) . One explanation is that the meta-features are highly correlated , and combining them boosts predictive power through noise reduction . An alternative explanation is that each meta-feature makes a partially independent contribution to evolutionary rate prediction . As the meta-features are already noise-reduced from the combination of individual features , we conclude that independent contributions are at least partially responsible ( see Figure S1 for further support ) . A much larger gain is made when the other five meta-features are combined ( 77 . 5% , up 8 . 6% compared to function alone ) . The combination of all seven meta-features produces further improvement ( 82 . 0% ) , suggesting that abundance and amino acid composition are the dominant predictors , and that other meta-features make small , individual contributions . Our best feature-based predictions of slowly evolving proteins reach 82 . 0% correct classification rate , which is slightly beyond the midpoint of random ( 50% ) and perfect ( 100% ) classification . Here , we evaluate the significance of this performance in comparison with predictions based on other methods of estimating evolutionary rate . Paired species comparison is a traditional method for estimating evolutionary rate that requires minimal genomic information . We used ranked evolutionary rates derived from a single paired species comparison ( S . cerevisiae versus one of the other five yeasts ) to predict the slowest evolving 20% of the proteins among the average rankings of the four remaining paired species comparisons ( see Figure 1 and the Methods section for details of the general ranking procedure used in all analyses ) . This procedure was repeated five times , once for each isolated paired species comparison . On average , paired species comparison correctly identified slowly evolving proteins 83 . 0% of the time , which is strikingly similar to our optimal feature-based predictions . This can be interpreted either as a testament to the power of our feature-based predictions , or as a warning regarding the limitations of paired species comparison for evolutionary rate estimation . The suboptimal performance of paired species comparison in the task of predicting slowly evolving proteins points to the existence of considerable rate heterogeneity among yeasts . In general , protein evolutionary rate can be decomposed into two components: a conserved component that is common to all yeast species , and a lineage-specific component that is unique to a particular yeast species ( reflecting common and lineage-specific selection pressures , respectively ) . The magnitude of the conserved component of protein evolutionary rate is an important quantity , as it defines the upper limit for evolutionary rate prediction using only broad , lineage-insensitive genomic features . The intuition here is simple: genomic features that do not vary across lineages cannot distinguish the fine details of lineage-specific evolutionary rate variation . We directly estimated the conserved component of evolutionary rate that is common to all yeasts by averaging over all paired species comparisons that do not involve S . cerevisiae . The predictive power of this common component in the classification of slowly evolving S . cerevisiae proteins is reasonably high , producing 92 . 8% correct classifications ( Figure 3 ) ; this is also the upper limit for correct classification based on lineage-insensitive ( meta- ) features . Our optimal feature-based predictions are able to explain three quarters of this upper limit . The value of 92 . 8% is the predictive upper limit only when the integration is restricted to lineage-insensitive genomic features . How conserved are the genomic features that we consider here ? Gross structural properties and broad functional assignments are likely to be conserved for homologous proteins [28] , [29] . This makes biological sense: although subtle details may change in recent evolution , an all-alpha helix enzyme in the cytosol of one yeast is unlikely to become an all-beta sheet transcription factor in the nucleus of a second yeast . As a result , such features cannot predict lineage-specific evolutionary rate variation , and their predictive power is therefore bounded by the upper limit . For amino acid composition , we assessed conservation using the orthology mappings from our evolutionary rate calculation: for each yeast protein , we calculated the average ranked amino acid composition across its orthologs in the other yeast species , and correlated this average with the ranked amino acid composition in S . cerevisiae . The average correlation coefficient is high at 0 . 917 , suggesting that amino acid composition is generally well conserved among yeasts , yet still subject to some degree of lineage-specific variation . As for abundance , experimental expression data for other yeast species are limited , condition-specific , and susceptible to noise . Here , we use codon bias as a proxy for expression level , and compare S . cerevisiae-specific values to values averaged over the other yeast species . Here the correlation coefficient is high at 0 . 886 , again indicative of general conservation with elements of lineage-specific variation . The genomic features most likely to be variable among yeasts are network-based features , since transcriptional regulation and protein-protein interaction are known to vary between yeasts [30] , [31] . The predictive limit of lineage-insensitive features is always bounded by 92 . 8%; as a feature's lineage specificity increases , its predictive limit can in principle approach 100% .
This study focused on identifying genomic features which contribute to the task of predicting evolutionary rate . While the purpose and relevance of many prediction tasks is immediately clear—for example , predicting gene essentiality in order to avoid the difficulty and expense of experimental determination [32]—one may question the need for predicting evolutionary rate . Simple methods for evaluating evolutionary rate based on species comparisons exist ( e . g . , the dN/dS ratio ) and can be evaluated with relative ease at the genomic scale . In the absence of such comparisons , we would have few means by which to test the validity of our predictions , given the timescale over which natural evolution operates . Why then do we wish to predict evolutionary rate ? The answer is that we are not interested so much in the predictions themselves , but rather the features which provide them . Biologists have long been interested in understanding the forces that drive evolution at various scales of life . However , our knowledge of the causal forces which underlie evolution at the molecular scale remains limited . By ranking the degree of correlation between various protein features and evolutionary rate , we hope to highlight those features which best dictate the selective constraint on a given protein . From the careful dissection of these individual correlations , one stands to gain a deeper understanding of their underlying biological significance . For example , the observed correlation between protein abundance and evolutionary rate led to biological insights regarding the evolution of translational robustness [33] . In a similar spirit , the correlation between evolutionary rate and number of transcriptional regulators that we discovered here leads to biological insights regarding the evolution of transcriptional regulation: target hubs in the transcriptional regulatory network are evolutionarily more constrained than non-hubs . Protein abundance , biological function , and amino acid composition consistently appeared in our analyses as dominant correlates of evolutionary rate . As we have mentioned before , the significance ( if not dominance ) of abundance is generally accepted; the significance of function has also been previously described [34] . However , the significance of amino acid composition in determining evolutionary rate has been a subject of some debate [24] , [35] . The information contribution analysis indicated that the predictive power of amino acid composition is high ( relative to other meta-features ) in the task of classifying slowly evolving proteins . We are able to partially explain the correlation between amino acid composition and evolutionary rate by revealing a hidden correlation with protein expression ( see Table 4 ) . For example , the top three amino acids that are negatively correlated with evolutionary rate , glycine , alanine , and valine , are also the most enriched in highly expressed proteins ( perhaps reflecting a preference for metabolically inexpensive building blocks ) . On the other hand , the correct classification rate improved when we combined abundance and amino acid composition , suggesting that amino acid composition makes at least a partially independent contribution . This additional contribution can be partially attributed to differences in amino acid mutability , as defined by Jones et al . [36] ( see Table 4 ) . For example , the top two amino acids that are positively correlated with evolutionary rate , serine and asparagine , are among the top three in terms of mutability . It is interesting to note that something as simple as amino acid composition can be highly predictive for both protein abundance and evolutionary rate . The previous section highlighted the importance of understanding within-feature correlation , specifically that between protein abundance and amino acid composition , in the search for determinants of evolutionary rate . Figure 4 explores the network of within-feature correlations for the twenty numerical features that best correlate with evolutionary rate , as listed in Table 2 ( glutamine content , the weakest of these correlates , has no strong within-feature correlations and is not depicted ) . We notice that related features can occur in tightly correlated clusters ( for example , the cluster of Codon Adaption Index , codon bias , protein expression , and absolute mRNA expression ) . This observation reinforces the value of considering such features together as meta-features , as we have done here . In general , the network exhibits clique-like behavior , characterized by dense connections among the feature nodes . This is not entirely surprising , as many features are known to be related to one another , and they all share a common correlation with evolutionary rate . However , this rampant intercorrelation is a significant hindrance to the task of isolating specific features as evolutionary determinants using traditional multivariate statistical techniques . By taking an integrated probabilistic approach ( considering all features together ) , we have been able to largely circumvent this issue . Figure 4 further reveals that the vast majority of correlations between pairs of features and evolutionary rate are transitive: if features A and B both positively ( or both negatively ) correlate with feature C , then feature A usually correlates positively with feature B . For example , increased GC content and increased codon bias are both associated with decreased evolutionary rates . At the same time , GC content and codon bias are positively correlated with one another . These transitive correlations are easy to understand . Interestingly , we also observed non-transitive correlations , for example between evolutionary rate , number of transcriptional regulators , and marginal essentiality . Both number of transcriptional regulators and marginal essentiality are negative correlates of evolutionary rate ( rs = −0 . 142 and −0 . 146 , respectively; p≪0 . 001 in both cases ) . However , as previously noted [37] , number of transcriptional regulators and marginal essentiality correlate in a negative manner with one another ( rs = −0 . 104 , p≪0 . 001 ) . The observed non-transitive correlations are statistically significant ( p≪0 . 001 ) , although we note that the correlations are rather weak and account for 1 . 1% to 2 . 1% of the variance . This seemingly counter-intuitive observation can be explained in the following way . Slowly evolving proteins can be divided into two largely non-overlapping groups: ( i ) those that are important under all conditions , meaning that they are essential , but not necessarily highly regulated , and ( ii ) those that are important only under specific conditions , which may experience sophisticated regulation , but are not necessarily annotated as essential . Proteins in the first group drive the negative correlation between essentiality and evolutionary rate , while proteins in the second group drive the negative correlation between number of regulators and evolutionary rate . This explains the observed non-transitive correlations among evolutionary rate , number of transcriptional regulators , and marginal essentiality . The patterns of correlation among features in Figure 4 provide further insights into our observed correlation between evolutionary rate and number of transcriptional regulators . Number of regulators is correlated with many other genomic features , most significantly with codon bias and Codon Adaption Index . Highly regulated proteins , though not necessarily essential or even expressed under laboratory conditions , may be strongly selected in the real world for their roles in stress response . Take , for example , HSP26—a player in the yeast response to heat shock , and the most highly regulated protein in our dataset . This protein is neither expressed nor essential under laboratory conditions . However , its coding sequence contains high codon bias , consistent with selection for efficient translation under stress . We therefore expect that translational selection and selection for the protein's stress-induced function have constrained its evolution in the wild . The integrated probabilistic approach we have taken in this study has both drawbacks and advantages . Like other correlational approaches , our approach is not able to distinguish correlation from causation , nor is it able to isolate cause from effect . We do not explicitly model the noise within the feature data as some other methods do [15] , which will tend to underestimate the predictive power of single features . On the other hand , the effect of noise is minimized by the binning of single features and introduction of meta-features . Our approach is flexible and robust , and is able to distinguish between dominant correlations and marginal ones . We are able to consider any features we choose , including those that are categorical ( rather than continuous ) or correlated with evolution in a non-linear manner . Furthermore , our approach compensates for redundancy among features , which , as with noise , we expect to be significant . Most importantly , our analyses feature high coverage of the yeast genome , thus making our results highly general . Accomplishing this requires the introduction of several approximations ( a relaxed definition of evolutionary rate , collecting feature data from a single species , and modeling missing data ) , though none of these are found to have a major effect on accuracy . To our surprise , we found that integrating a diverse collection of single-genome features was roughly equivalent to paired species comparison for identifying slowly evolving proteins , but still worse than what lineage-insensitive features can in principle predict . Our conclusion from this finding is that the dominant , independent correlates of evolutionary rate are likely known , even though other significant and interesting correlates may remain to be found ( see [38] for one recent example ) . Further dissection of individual correlations between protein features and evolutionary rate will be needed in order to gain a deeper understanding of their biological significance . As we have demonstrated in the cases of amino acid composition , protein abundance , essentiality , and number of transcriptional regulators , there is also great insight to be had by exploring the relationships between protein features .
We based our measure of protein evolutionary rate on comparisons between Saccharomyces cerevisiae and five related yeast species: S . paradoxus , S . mikatae , S . bayanus , S . castellii , and S . kluyveri . Of the 5 , 861 open reading frames ( ORFs ) in the S . cerevisiae genome , 324 had no annotated orthologs [39] among these species , and were therefore discarded . The remaining 5 , 537 ORFs ( 94 . 5% genome coverage ) each have at least one ortholog [39] in at least one of the five related yeasts; this group forms the basis of our evolutionary rates dataset . We first performed local alignment [40] between each ORF and its annotated orthologs across the five species . If an ORF had multiple orthologs in a given species , only the most significant alignment with the highest score was saved . These protein alignments ( having 95% ORF coverage , on average ) were used to generate corresponding DNA codon alignments , which were then piped into PAML [41] to calculate dN/dS [42] . All dN/dS values resulting from a given paired species comparison ( i . e . , S . cerevisiae versus one other yeast ) were then treated as follows: ( i ) dN/dS was first adjusted according to the method of [43] to compensate for selection at synonymous sites; ( ii ) adjusted dN/dS values were next sorted and converted to ranks; and ( 3 ) ranks were normalized relative to the total number of alignments considered in the paired species comparison . Finally , a single evolutionary rate was generated for a given ORF by averaging over its normalized ranks from all paired species comparisons in which an ortholog was present and dN/dS was successfully calculated . The values were then re-ranked and divided into five equally populated bins corresponding to low , medium low , medium , medium high , and high evolutionary rate . This procedure is summarized in Figure 1 . We provide the average ranks and bins of yeast protein evolutionary rate in Table S1 . Sequence data for S . kluyveri were obtained from [44]; all other sequence data were obtained from [45] . Basic protein information about each ORF was downloaded from the Saccharomyces Genome Database [45] . Protein GO annotations were downloaded from the Gene Ontology project website [46] . Protein-protein interaction data were downloaded from BioGRID [47] . Transcriptional regulatory data were obtained as described in ( Wang , Zhang , and Xia , submitted ) . Protein native disorder was predicted from sequence using DISOPRED [48] . Transmembrane helix content was predicted from sequence using TMHMM [49] . All other feature data were assembled following the procedures outlined previously [50] . Note that the majority of our features are derived or predicted from sequence alone , and therefore have high coverage of the yeast genome . At the same time , some features that we considered contain missing data . In the mutual information and subsequent analyses , missing data are treated as a separate feature bin . For example , the mRNA expression feature now has six categorical values: high , medium high , medium , medium low , low , and missing . These “completed” features are then correlated with evolutionary rate . Here , we assume that missing data bins such as “unknown biological process” or “missing mRNA expression” can be correlated with evolutionary rate just as we would correlate regular feature bins , such as “constituent of the ribosome , ” or “high mRNA expression level . ” This approach involves fewer assumptions about the nature of missing data than alternative strategies , such as listwise deletion , mean substitution , and imputation . Given N pairs of quantities ( xi , yi ) , i = 1 , … , N , the Spearman rank correlation coefficient rs is computed in the following way:where Ri is the rank of xi among the other x's , Si is the rank of yi among the other y's . The mutual information I between two discrete random variables X and Y is computed in the following way: For a given protein , we want to predict the class label yi ( 1 if the protein evolves slowly , and 0 otherwise ) by integrating genomic features F . There are m categorical features , F1 , … , Fm , where each feature Fj can take on rj different values , fj1 , fj2 , … , . The training set , { ( F ( i ) , y ( i ) ) ; i = 1 , … , n} , contains n samples . Logistic regression can be expressed as the following weighted voting scheme:Where I is the indicator function—I ( X ) is 1 when statement X is true , and 0 otherwise . wjk are weights associated with each piece of evidence . p ( y = 1|F ) is the probability that the protein evolves slowly given the features . The protein is predicted to evolve slowly if and only if p ( y = 1|F ) is larger than 0 . 5 . All weights are chosen to optimize the following log-likelihood function for the training set , i . e . the log-probability of observing the data given the weights:The right-hand side of the above equation measures the agreement between the actual class labels y and the predictions p ( y|F ) . | Proteins encoded within a given genome are known to evolve at drastically different rates . Through recent large-scale studies , researchers have measured a wide variety of properties for all proteins in yeast . We are interested to know how these properties relate to one another and to what extent they explain evolutionary rate variation . Protein properties are a heterogeneous mix , a factor which complicates research in this area . For example , some properties ( e . g . , protein abundance ) are numerical , while others ( e . g . , protein function ) are descriptive; protein properties may also suffer from noise and hidden redundancies . We have addressed these issues within a flexible and robust statistical framework . We first ranked a large list of protein properties by the strength of their relationships with evolutionary rate; this confirms many known evolutionary relationships and also highlights several new ones . Similar protein properties were then grouped and applied to predict slowly evolving proteins . Some of these groups were as effective as paired species comparison in making correct predictions , although in both cases a great deal of evolutionary rate variation remained to be explained . Our work has helped to refine the set of protein properties that researchers should consider as they investigate the mechanisms underlying protein evolution . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"computational",
"biology/comparative",
"sequence",
"analysis",
"computational",
"biology/evolutionary",
"modeling",
"molecular",
"biology/bioinformatics",
"computational",
"biology/macromolecular",
"sequence",
"analysis",
"evolutionary",
"biology/bioinformatics",
"genetics",
"and",
"genomics/bioinformatics"
] | 2009 | Integrated Assessment of Genomic Correlates of Protein Evolutionary Rate |
As increasingly more genomes are sequenced , the vast majority of proteins may only be annotated computationally , given experimental investigation is extremely costly . This highlights the need for computational methods to determine protein functions quickly and reliably . We believe dividing a protein family into subtypes which share specific functions uncommon to the whole family reduces the function annotation problem’s complexity . Hence , this work’s purpose is to detect isofunctional subfamilies inside a family of unknown function , while identifying differentiating residues . Similarity between protein pairs according to various properties is interpreted as functional similarity evidence . Data are integrated using genetic programming and provided to a spectral clustering algorithm , which creates clusters of similar proteins . The proposed framework was applied to well-known protein families and to a family of unknown function , then compared to ASMC . Results showed our fully automated technique obtained better clusters than ASMC for two families , besides equivalent results for other two , including one whose clusters were manually defined . Clusters produced by our framework showed great correspondence with the known subfamilies , besides being more contrasting than those produced by ASMC . Additionally , for the families whose specificity determining positions are known , such residues were among those our technique considered most important to differentiate a given group . When run with the crotonase and enolase SFLD superfamilies , the results showed great agreement with this gold-standard . Best results consistently involved multiple data types , thus confirming our hypothesis that similarities according to different knowledge domains may be used as functional similarity evidence . Our main contributions are the proposed strategy for selecting and integrating data types , along with the ability to work with noisy and incomplete data; domain knowledge usage for detecting subfamilies in a family with different specificities , thus reducing the complexity of the experimental function characterization problem; and the identification of residues responsible for specificity .
Despite the best research efforts , a substantial and ever-increasing amount of predicted proteins still lack functional annotation [1] . Indeed , the unprecedented increase in the number of new protein sequences being produced by genomics and proteomics projects , as well as the copious amounts of structures for proteins of unknown functions being solved by structural genomics , directly highlight the need for computational methods to determine , quickly and accurately , the molecular and cellular functions of such proteins , given that experimental investigation is difficult , costly , and time-consuming [2 , 3] . As the number of sequenced genomes rapidly increases , the vast majority of gene products may only be annotated computationally [4] . However , no high-throughput approaches currently exist capable of revealing the function of every hypothetical gene in the already sequenced genomes . This goal can only be reached per the efforts of several experimental , structural , and computational biologists [5] . The work presented herein is a computational effort aiming to take a step toward that goal . The commonest protein function annotation approach is homology-based annotation transfer , which assumes proteins sufficiently alike in sequence and structure perform similar functions [3] . Such methods have various limitations due to this assumption [6] . On account of protein function plasticity and of the intrinsic imprecision in related databases , various aspects of function cannot be accurately transferred between similar sequences indiscriminately [7] . In fact , homology-based annotation transfer methods are considered one of the main sources of annotation errors due to an excessively liberal application of function inheritance [3] , which fails when similar proteins cannot be identified or when they , too , lack reliable annotations [8 , 9] . Moreover , such methods also fail for proteins that have the same function despite being different in sequence and structure ( i . e . , convergent evolution ) [10] , and also for those which are sequentially and/or structurally similar yet functionally diverged during evolution [9] . Automatic protein function annotation methods depend on a correlation between functional and sequential or structural similarity measures [11] , the simplest of which explores global sequence similarity . Other measures commonly employed in the literature are local sequence motifs , global and local structural similarities , and 3D templates . Since such similarity measures focus on different protein features , one may expect they yield better functional annotations when combined [11] . In fact , literature shows using a single data type ( e . g . , sequence similarity ) is insufficient to precisely annotate protein functions due to the immense amount of factors involved in determining a function , and to the consequent complexity of the automatic annotation problem [4 , 7–10 , 12–20] . A combined approach is usually more powerful than its individual components [3] , so blending various data types is crucial in order to transfer annotations more reliably [21] . This attests to the great importance and need for automatic function analysis methods capable of integrating various data types . Increasing the difficulty , one ought to consider attributing a function to a protein family is further complicated by the fact that many families are composed of proteins with multiple folds and/or functions . In such cases , determining possible subfamilies may lead to important information about a related protein’s function and structure , as well as about the functional diversification acquired by the family during evolution [7] . Therefore , a family of homologous proteins may be divided into subtypes which share specific functions uncommon to the family as a whole [22] . We believe determining such subfamilies to be a first step toward reducing the protein function annotation problem’s complexity . Hence , this work’s purpose is the detection of isofunctional subfamilies in a protein family of unknown function , along with the identification of residues responsible for subfamily differentiation . Various methods have been proposed to identify amino acid conservation patterns which distinguish subgroups in a protein family [22–29] . In general , such methods have the considerable disadvantage that subfamilies must be known a priori . Apart from the scarcity of experimental information about subfamilies , this requirement is prohibitive when working with protein families of unknown function . To the best of our knowledge , a single method in the literature is similar to ours in that it first attempts to identify subfamilies in a Pfam [30] family through clustering and , then , to detect specificity determining residues which characterize them: Active Sites Modeling and Clustering ( ASMC ) [31] , which clusters proteins according exclusively to active site composition . Simply put , given a Pfam family , ASMC first performs homology-based structural modeling of its members with a reference structure , later superposing such models to the structure in order to identify residues aligned to its active site . As a result , it builds a multiple sequence alignment ( MSA ) that represents the active site composition for each protein in the family . This MSA is then subjected to a hierarchical clustering , generating a tree whose nodes are protein groups and whose levels represent successive subdivisions of the family: the root of the tree has all proteins in the same group , whereas the leaves represent singleton clusters . Afterward , the authors manually cut this tree in order to obtain clusters which they find most interesting . The reported number of clusters in the family is , thus , manually defined . Finally , the authors perform a statistical significance analysis to determine the active site positions which were most important to differentiate among groups . Considering the various challenges to automatic function annotation may be extended to the problem of detecting subfamilies , in order to overcome the previously mentioned major obstacles faced by homology-based methods , we adopt an approach that integrates various data types . For this purpose , the similarity between protein pairs according to different knowledge domains is interpreted as evidence , albeit weak , of functional similarity . We integrate such data types using genetic programming and , afterward , provide it as input to a spectral clustering algorithm Our main goal is to propose a strategy for selecting and combining pieces of functional similarity evidence between protein pairs , and to analyze the manner in which integrating information from different knowledge domains is capable of directing a clustering process to detect , in a protein family of unknown function , isofunctional subfamilies , along with the residues that differentiate them . This goal was successfully achieved . The proposed framework’s capability of using diverse data types , even if incomplete or uncertain , is of remarkable importance for application scenarios such as this , since data from biological experiments are naturally imprecise , mainly due to the dynamic nature of the phenomena investigated as well as to experiment interpretation errors [4] , and certain types of information are relatively scarce . Additionally , protein function is determined by various factors , and the complementarity of the different data sources allows for the algorithm to work with as much information as possible . Our main contributions are the proposed strategy for selecting and integrating various data types , along with the ability to work with noisy and incomplete data; the possibility of using domain knowledge for detecting isofunctional subfamilies in a protein family with different specificities or even of unknown function , thus reducing the complexity of the experimental function characterization problem; and the identification of residues responsible for specificity .
Once a Pfam family of interest is defined , a filtering process is applied to obtain the protein set to be studied . First , we collect the family’s full sequence alignment from Pfam and extract the UniProt identifiers , together with the subsequences which contain the domain that characterizes the family . The protein set is later filtered by subsequence size , eliminating those whose lengths differ more than a standard deviation from the family average , as done by ASMC [31] . Afterward , we collect the structures associated to the family from PDB , separate the chains , and select the reference structures to be used as templates for structurally modeling the family sequences using Modeller [32] , and also to search for pockets that are possible active sites using Fpocket [33] . We prioritize structures obtained by X-ray crystallography , with high resolution and which contain ligands . Next , the protein set is further filtered according to similarity with the reference structures , eliminating those with less than 30% identity to all structures , which is the minimum level accepted by Modeller for creating a structural model . For each sequence , we choose the model with the smallest energy , as done in [1] and [31] . By the end of this filtering process , the database contains the UniProt identifiers , amino acid subsequences containing the family domain , and structural models for all remaining proteins . The steps taken to collect and apply the various data types interpreted as evidence of functional similarity are described next . Any data that may be used to compare proteins pairs can be added to the process . Given the data integration method ( i . e . , genetic programming ) is capable of learning which data types contribute to achieving good clusterings and filter out those of little use , even data types unlikely to be related to functional similarity were included . Each column in the database corresponds to one of the aforementioned similarity matrices , each of which is normalized to [0 , 1] , or to [-1 , 1] in case negative values exist . The data types for which smaller values indicate greater similarity , as is the case for those involving differences or distances , have their intervals reversed . Hence , all matrices may be interpreted in the same way: the higher the value , the larger the similarity between that protein pair according to the corresponding data type . In order to combine such primary similarity matrices into a single matrix to be provided as input to the clustering algorithm , we use genetic programming ( GP ) . GP is a natural computing technique that automatically solves problems without the user having to know or specify the form of the solution . Basically , in each generation , a population of individuals , each of which represents a combination of data types in this work , is stochastically transformed into a presumably better population [47] . The execution ends when a maximum number of generations is reached or when some other stopping criterion is met . Such transformations are accomplished by genetic operators of crossover , reproduction , and mutation , which recombine parts of individuals from one population to create individuals for the next [48] . Crossover works by randomly selecting parts of two individuals and switching them . For mutation , a random part of a single individual is replaced by new code , whereas in reproduction , an individual is selected and copied into the next generation [47] . Genetic operators are usually mutually exclusive , and their probability of application is called the operator rate . Individuals are selected to undergo such operations according to their fitness value . Thus , fitter individuals are more likely to be selected to “breed” , producing new individuals for the next generation . This work’s GP system was implemented using the lil-gp library [48] in C . Starting from random data combinations , it learns , over generations , which ones yield better protein clusters . Each primary matrix is depicted by a variable , so individuals represent equations that combine different matrices through addition . For each individual in the population , the GP system calculates the final similarity matrix by applying its equation to each protein pair , and subsequently runs the spectral clustering algorithm with the calculated matrix , returning the quality of the yielded clustering as the individual’s fitness value . By evolving a population of equations that combine the various data sources , apart from the actual clusters generated , results will allow to check which types of information are most useful to discriminate among groups in a protein family . Eq 1 shows an example of individual which calculates the similarity sij between each protein pair ( i , j ) by adding the number of InterPro annotations they have in common , their conserved neighborhood score in STRING , and three times the TM-score of their structural alignments . Clustering is a data mining technique which consists in dividing a set of objects into natural clusters , each of which represents a significant subpopulation , so that objects in the same group are very similar to each other , while different from those in other clusters [49] . In this work , we consider partitional clustering methods , in which none of K clusters are empty and each object belongs to a single cluster [50] . Among the various algorithms presented in the literature , we opted for employing spectral clustering , since it is capable of solving very complex problems such as the case that , when plotted , the objects from each cluster are positioned in intertwined spirals , which cannot be separated by something simple as a line or a curve . Such an algorithm was necessary for our application scenario since families of homologous proteins can rarely be separated into subfamilies easily . Spectral clustering uses the eigenvectors and eigenvalues of the similarity matrix to reduce the number of dimensions before performing clustering in the reduced space . First , a similarity graph is built from the inputted similarity matrix . Next , its Laplacian matrix is calculated , along with its eigenvectors and eigenvalues . The eigenvectors corresponding to the K smallest eigenvalues are taken , each as a dimension in the new data set representation . This change in representation from the original space to a K-dimensional space accentuates the cluster properties in the data , so that clusters may be trivially detected in the new representation [51] , so much so that a simple clustering algorithm like K-Means may be used . In this work , given the similarity matrix calculated by a GP system individual , we define the adjacency matrix of the totally connected similarity graph , and calculate its normalized asymmetric Laplacian matrix , its eigenvalues and eigenvectors , taking the K first eigenvectors . This new N×K matrix , where N is the number of proteins in the family and K is the desired number of clusters , is then provided as input to the K-Means clustering algorithm . Given our goal of detecting isofunctional subfamilies in a protein family , and considering each cluster is described by an active site composition-based profile , a quality measure which numerically reflects the differences among cluster profiles is required . We consider pointwise mutual information ( PMI ) [52] , which measures the amount of information the occurrence of a specific value x contributes to making the correct classification of an object relative to cluster y [53] . The PMI is a measure of how much the event co-occurrence probability ( p ( x , y ) ) differs from expected based on the individual event probabilities and on the independence assumption ( p ( x ) p ( y ) ) , and is calculated by Eq 2 [54] . If there exists a genuine association between the values , then p ( x , y ) ≫ p ( x ) p ( y ) and , consequently , PMI ( x , y ) ≫ 0 . If no interesting relationship exists , then p ( x , y ) ≈ p ( x ) p ( y ) and PMI ( x , y ) ≈ 0 . Finally , if x and y are in complementary distributions , then p ( x , y ) ≪ p ( x ) p ( y ) , hence PMI ( x , y ) ≪ 0 . PMI is “pointwise” because it is calculated for two values x and y , whereas mutual information ( MI ) is calculated for two variables X and Y , and corresponds to the expected PMI over all possible values , i . e . , MI ( X , Y ) = ∑x∑y p ( x , y ) PMI ( x , y ) [54] . MI measures the information dependence or overlap between two random variables , reaching maximum value when the variables are perfectly correlated [54 , 55] . We consider a cluster to be interesting when it has ( almost ) exclusive residues for specific active site positions . Hence , we compare each cluster to the union of the others . For each position pi , residue rk’s importance for cluster cj is measured by PMIpi ( cj , rk ) , whereas its importance in the union of the remaining clusters ( c j ¯ ) is calculated by P M I p i ( c j ¯ , r k ) . This yields MIpi ( cj , rk ) , calculated by Eq 3 , in which the ppi ( cj , rk ) and p p i ( c j ¯ , r k ) probabilities are estimated by residue rk’s frequency in cluster cj and in the other clusters at position pi . Because PMIpi ( cj , rk ) and P M I p i ( c j ¯ , r k ) values have opposite signs , and considering that we only deem important to a cluster those residues more frequent in it than in the other clusters , only residues for which PMIpi ( cj , rk ) > 0 are considered . In case the residue also occurs in other clusters , then P M I p i ( c j ¯ , r k ) < 0 , and their addition will reduce rk’s importance for cluster cj . Finally , if PMIpi ( cj , rk ) ≤ 0 , we consider MIpi ( cj , rk ) = 0 . For a given cluster , there might be multiple residues in a specific active site position . Hence , considering fk as residue rk’s frequency in cluster cj , we have that MIpi ( cj ) = ∑k fk MIpi ( cj , rk ) . Gaps are not considered in this calculation . Finally , the quality measure for the clustering as a whole is the overall average , calculated by Eq 4 , where P is the total number of positions , and C is the number of clusters . The GP system uses this as fitness function . Thus , it attempts to maximize the mutual information between the active site residues and the clusters , which is equivalent to searching for clusters that present characteristic active site compositions . When a ground-truth exists such as the SFLD family classification , external validation measures may be used to calculate a clustering’s agreement with it . Pairwise measures consider the cluster labels and ground-truth classifications over all pairs of objects . For an object pair with the same ground-truth classification , the objects may be attributed to a same ( true positive—TP ) or different ( false negative—FN ) clusters . Analogously , a pair with different ground-truth classifications , may be assigned to a same ( false positive—FP ) or different ( true negative—TN ) clusters . The precision ( P ) for a given clustering is the percentage of object pairs that are in a same cluster and actually have the same ground-truth classification ( P = T P T P + F P ) , while the recall ( R ) is the fraction of pairs with the same ground-truth classification that were assigned to a same cluster ( R = T P T P + F N ) . The F1 score tries to balance the precision and recall values by computing their harmonic mean , and is calculated as F 1 = 2 × P × R P + R . The Rand index measures the fraction of true positives and negatives over all object pairs , and is defined as R a n d = T P + T N T P + F P + F N + T N . It is symmetric in terms of true positives and negatives , and measures the fraction of pairs where the clustering and the ground-truth classification agree . The Rand index has a value between 0 and 1 , with 0 indicating complete disagreement and 1 indicating the clusters are exactly the same as the ground-truth classification . The Jaccard coefficient measures the fraction of true positives when ignoring the true negatives . It is defined as J a c c a r d = T P T P + F P + F N . Since it ignores true negatives , it is asymmetric in terms of the true positives and negatives . Thus , it emphasizes the similarity in terms of the object pairs that belong together in both the clustering and the ground-truth , but discounts the pairs that do not belong together [49] . The larger the values for these measures , the better the agreement of the clustering with the ground-truth classification . Additionally , the variation of information measures the amount of information not shared between the clustering and the ground-truth , and is calculated as VI = H ( S ) + H ( S′ ) − 2I ( S , S′ ) , where H is the entropy of a data partition , and I is the mutual information between two partitions of the same data . Lastly , the edit distance is defined as the minimum number of split or merge operations required to transform the clustering into the ground-truth classification , where a split or merge affecting multiple objects is considered one operation . The edit distance between the ground-truth classification and a clustering , with class k and cluster k’ , respectively , is calculated as Edit = 2 ( ∑rk , k′ ) − K − K′ , where rk , k’ equals 1 if class k and cluster k’ have items in common , and zero otherwise . K is the number of classes , while K’ is the number of clusters [64] . The smaller the values for the variation of information and the edit distance , the more similar the clustering is to the ground-truth classification .
Nucleotidyl cyclases are a family of cytosolic or membrane-attached domains that catalyze the transformation of a nucleotide triphosphate into a cyclic nucleotide monophosphate [25] . These proteins have fundamental roles in a wide range of cellular processes , and two functional subfamilies exist , namely adenylate cyclases , which act on ATP to form cAMP , and guanylate cyclases , which catalyze the conversion of GTP to cGMP [59] . Mutations of only two residues ( Glu-Lys and Cys-Asp ) are sufficient to completely alter the specificity from GTP to ATP [25] . After removing , from the original set , 75 proteins that became obsolete in UniProt , 461 remained in this family , of which 186 are labeled as adenylate cyclases and 275 , as guanylate cyclases , according to the labels employed in [31] . Thus , the GP system was run to divide this family into two clusters . However , with the same parameter values used in [31] , ASMC produced , for the updated protein set , a hierarchical clustering whose first level divided the family into three clusters , and whose second level divided it into six . Hence , in order to compare results , the GP system was also run with three and six clusters . Table 5 presents the data combinations produced by the GP system which yielded the best results for the nucleotidyl cyclases in five runs for each considered number of clusters . Since the MI is based on active site composition , it was expected that the related similarity matrices would be involved in the best results . Other data types which stood out were the global and local sequence alignments scores ( seqAliG and seqAliL ) , the structural alignment identities ( strAliId ) , and the differences in aliphatic residue content ( difAliphRes ) . One may observe a large amount of data types was required by the GP system to partition the family into two clusters . This Pfam family , defined by the presence of a conserved domain of unknown function , was studied in [1] because it contains the Kce protein , which was of interest to the authors for they had previously discovered an initial association between it and a formerly orphan enzyme activity . Such activity is involved in the lysine fermentation pathway and catalyzes the condensation of β-keto-5-amino-hexanoate ( KAH ) and acetyl-CoA to produce aminobutyryl-CoA and acetoacetate . Since the DUF849 proteins do not all come from organisms capable of fermenting lysine , this suggests there are various biochemical reactions catalyzed by different family members [1] . Hence , the authors considered DUF849 a good case study for discovering new activities in a protein family of unknown function , and named the family “BKACE” , which stands for β-keto acid cleavage enzyme . The main result presented in [1] is a division of the set of 725 proteins into seven groups , obtained by manually altering ASMC’s hierarchical clustering . This manipulation allows to consider different hierarchy levels for each tree branch , thus distorting the algorithm’s output to build clusters the authors consider most interesting . Group logos are presented in Fig 3 . According to the authors , G3 presents five subgroups , and there is high correlation among the distribution of the proteins in the seven clusters and the nature of the compounds they transform , as presented in Table 7 . Enzymatic activity distribution in these groups , however , is not as clear as depicted , since there are proteins which showed activity for substrates related to other clusters . Considering the substrates for which activities were tested in [1] , S1 Table shows the number of times an activity was detected for each group , considering two repetitions for each test . The distribution , among the manually produced clusters , of the number of enzymes considered active for each substrate , shows the complexity of clustering this family into isofunctional subfamilies due to the promiscuity it presents . The manually defined groups have MI = 14 . 05 . For comparison , the GP system was run to divide the DUF849 family into seven clusters . The best result has MI = 36 . 51 and uses equation 2ASid + csmDist + 2neighborhood + 2seqAliL + strAliId . Cluster logos are presented in Fig 4 , while the residues that most distinguish each cluster are listed in S2 Table . Because this is a protein family of unknown function , result comparison is complicated . However , there is substantial correspondence between the two clusterings: Despite being completely automatic , our framework was still able to obtain clusters very similar to those manually produced in [1] , even better concentrating the enzymes active for 4-hydroxybenzoylacetate . The only problem in this case study was the GP system considered it to be more relevant , in terms of active site composition , to break group G7 into two relatively uniform clusters , while the authors in [1] opted not to divide this heterogeneous group since the corresponding enzymes are inactive for all tested substrates . Thus , we have successfully demonstrated the proposed framework’s ability and utility for detecting isofunctional subfamilies in families of unknown function . Protein kinases are enzymes that modify the functions of other proteins by adding phosphate groups usually removed from ATP , covalently binding them to the side chains of Ser , Thr , or Tyr residues [25] . They are one of the largest and most functionally diverse protein families , responsible for controlling the majority of biochemical pathways , performing key roles in regulating metabolic processes , cell differentiation , and proliferation of diverse cell types [60] . The main division in protein kinases is between Ser/Thr and Tyr kinases: Ser and Thr are similar in size and shape , while the reaction chemistry and substrate size are substantially different for Tyr [25] . The majority of kinases act upon Ser or Thr , while others are specific to Tyr , and some act upon all three . It is known that some positions confer specificity , such as subdomain VI , in which consensus sequence RDLKPEN is usually found in Ser/Thr kinases , while RDLAARN is typical of Tyr kinases [25] . After removing from the protein set used in [31] 314 proteins that became obsolete in UniProt , 3 , 087 remained in this family , of which 2 , 044 are labeled as Ser/Thr kinases and 1 , 043 as Tyr kinases , according to the labels employed in [31] , in which a subgroup of 235 Tyr kinases labeled as Epidermal Growth Factor Receptors ( EGFRs ) was also reported . Using the same parameter values applied in [31] for the original protein set , ASMC produced , for the updated set , a hierarchical clustering with three and seven clusters in its first two levels . Given two main subfamilies exist , our framework was applied to divide the family into two , three , and seven clusters . Table 8 presents the data combinations produced by the GP system which yielded the best results for protein kinases . Yet again , the presence of active site-related data is noticeable as expected due to the quality measure employed . Other outstanding data types were structural alignment identities ( strAliId ) and GO term similarities ( go ) . Proteases are a large enzyme family involved in peptide bond hydrolysis . Almost a third of all proteases are serine proteases , whose name derives from the Ser residue at the active site [61] . Serine proteases are involved in a huge number of biological processes , such as digestion , homeostasis , apoptosis , signal transduction , reproduction , immune response , and blood coagulation [61 , 62] . They present a catalytic triad composed of a Ser , an Asp , and a His [31] , whose 3D arrangement allows for moving protons in and out of the active site . All serine proteases act through a similar catalytic mechanism , but have different cleavage preferences due to active site changes [25 , 31] . For chymotrypsins , the active site is lined with hydrophobic residues , so proteins containing hydrophobic residues such as Leu or Ile form strong bonds in the correct orientation for the triad to act . The cavity in trypsins contains a negatively charged Asp , so their substrates must have a specifically positioned positively charged residue such as Lys or Arg . In turn , elastases have smaller cavities , so only proteins containing short-chained residues such as Gly or Ala can be acted upon [62] . After removing 140 sequences that became obsolete in UniProt since being used in [31] , 1 , 533 proteins remained , of which 43 are labeled as elastases , 26 as chymotrypsins , and 1 , 464 as trypsins , according to the subfamily labels employed in [31] , in which a subgroup of 13 trypsins were found to be kallikreins . When ASMC is run on the updated protein set with the same parameter values used for the original set , the family is not divided . Hence , the main parameter ( -C 0 . 25 ) was reduced in 0 . 05 decrements until a value which divided the family was found: 0 . 15 , which yielded a hierarchical clustering with four clusters in its first level , and eleven in its second . For comparison purposes , the proposed framework was run to divide the family into four and eleven clusters . Table 10 shows the data combinations that yielded the best results for the serine proteases in five runs for each considered number of clusters . Again , the active site-based similarity matrices showed strong utility , as expected . The other data type that stood out was the global sequence alignment score ( seqAliG ) . The crotonase superfamily enzymes catalyze a wide range of metabolic reactions . Some have been shown to display dehalogenase , hydratase , and isomerase activities , while others have been implicated in carbon-carbon bond formation and cleavage , as well as the hydrolysis of thioesters [63] . After applying the filtering process described in the “Protein family definition” subsection to the 7 , 908 crotonase superfamily proteins with known families in the SFLD [57] , 2 , 694 proteins remained , distributed among twelve families , all of which are in the crotonase like subgroup . The superfamily distribution is presented in Table 11 . The active site compositions were extracted from structurally aligning the models against reference structure 1MJ3’s active site . Given twelve families exist , the proposed framework was applied to divide the family into twelve clusters . The best result is obtained with equation ASid + seqAliG + strAliId , which yielded a clustering with MI = 52 . 18 . Cluster logos and compositions in terms of SFLD family labels are presented in the S11 Text . The distribution of crotonase families among clusters is presented in Table 12 . In comparison with SFLD’s family classification , the clustering produced by the GP system for dividing the crotonase superfamily into twelve clusters presents a Rand index of 0 . 80 , a Jaccard coefficient of 0 . 44 , 93 . 84% precision , 45 . 06% recall , an F1 score of 0 . 61 , a variation of information of 0 . 80 , and an edit distance of 26 . These values indicate the clustering produced by the GP system is in high agreement with the SFLD family classification , yet this agreement is more related to precision ( i . e . , pairs that are in the same cluster and actually have the same classification ) than to recall ( i . e . , pairs with the same classification that are actually in the same cluster ) . Given the enoyl-CoA hydratase , which accounts for 55 . 94% of the crotonase superfamily , had elements assigned to four different clusters , this greatly impacted the recall . Hence , the results suggest more clusters are required to properly separate the families , due to the existing variation among enoyl-CoA hydratases . When the SCI-PHY classification method was applied to the crotonase superfamily in [64] , the authors achieved a variation of information of 1 . 05 , and an edit distance of 32 . Although different protein sets were considered for each technique , this suggests the proposed framework outperforms SCI-PHY . Unfortunately , we were unable to properly compare the techniques on a same data set because the studied protein set is not presented in [64] . Enolase superfamily enzymes catalyze the abstraction of the α-proton of a carboxylic acid to form an enolic intermediate . This is mediated by conserved active site residues . Reactions catalyzed by these enzymes include racemization , β-elimination of water and of ammonia , and cycloisomerization . These enzymes have two structural domains: a N-terminal capping domain and a C-terminal TIM beta/alpha-barrel domain , both of which are required for function [65] . After applying the filtering process described in the “Protein family definition” subsection to the 31 , 182 enolase superfamily proteins with known families in the SFLD [57] , 4 , 791 proteins remained , distributed among six subgroups and twelve families . The superfamily distribution is presented in Table 13 . The active site compositions were extracted from structurally aligning the sequence models against reference structure 1MDR’s active site , chosen according to the process described in the Methods section . Given twelve families exist , the proposed framework was applied to divide the family into twelve clusters . The best result is obtained with equation 3APid + go + 2seqAliG , which yielded a clustering with MI = 98 . 18 . Cluster logos and compositions in terms of SFLD family labels are presented in the S12 Text . The distribution of enolase families among clusters is presented in Table 14 . Although nine of twelve clusters are pure ( i . e . , contain a single family ) , the mixture in clusters IX , XI and XII shows that more clusters are required in order to properly separate the families . As was the case with the serine proteases and crotonases , this is likely caused by family imbalance: the enolase family accounts for 52% of the protein set , and variation among the enolases may dominate the smaller families . Interestingly , despite the mixture in Cluster IX , all three families are in the muconate cycloisomerase subgroup of the enolase superfamily . In comparison with SFLD’s family classification , the clustering presents a Rand Index of 0 . 87 , a Jaccard coefficient of 0 . 62 , 87 . 30% precision , 67 . 83% recall , an F1 measure of 0 . 76 , a variation of information of 0 . 84 , and an edit distance of 34 . These values reflect the great agreement of the clustering produced by the GP system with the SFLD family classification . The somewhat low recall , however , further suggests more clusters are required to properly separate the families , likely due to the existing variation among the dominating enolase family . When applying the SCI-PHY classification method to the enolase superfamily in [64] , the authors achieved a variation of information of 1 . 37 , and an edit distance of 70 . Although the protein sets are different , this suggests the proposed framework outperforms SCI-PHY . However , experiments on a same dataset would be required to properly compare the techniques . Unfortunately , we were unable to perform such comparison since the studied protein set is not presented in [64] .
A deeper investigation is required into the semantics of the data combinations produced by the Genetic Programming ( GP ) system . A network of dependent or synonym variables should aid in better comprehending the redundancy among the data types employed as functional similarity evidence . Eliminating redundant data types from the GP system might ease the semantic analysis of the obtained data combinations , as well as improve the quality of the generated clusters . This requires experiments with different subsets of the studied data types . Additionally , we need to investigate the use of phylogenetic information as functional similarity evidence , given that proteins from a same species should subdivide a cluster into different , yet related , functions . This is extremely hampered , however , by data scarcity and the existing imbalance of known proteins among the species of origin . Considering the difficulty in dividing the serine protease family , as well as the crotonase and enolase superfamilies , due to ( sub ) family imbalance , it would be interesting to apply sampling methods to families in which this occurs , in order to evaluate the technique’s performance in more well-balanced databases . Furthermore , due to the added complexity for promiscuous protein families , we need to investigate the possibility of adapting the proposed framework to using fuzzy clustering algorithms . Unlike partitional clustering , fuzzy clustering would output , for each protein , a level of membership in each cluster . Thus , a protein which performs multiple functions could belong , at the same time , to different clusters . Lastly , further efforts are required in the pursuit of a clustering quality measure appropriate for this application scenario that will enable comparing clusterings with different numbers of clusters in order to determine the ideal number of clusters in a protein family . | The knowledge of protein functions is central for understanding life at a molecular level and has huge biochemical and pharmaceutical implications . However , despite best research efforts , a substantial and ever-increasing number of proteins predicted by genome sequencing projects still lack functional annotations . Computational methods are required to determine protein functions quickly and reliably since experimental investigation is difficult and costly . Considering literature shows combining various types of information is crucial for functionally annotating proteins , such methods must be able to integrate data from different sources which may be scattered , non-standardized , incomplete , and noisy . Many protein families are composed of proteins with different folds and functions . In such cases , the division into subtypes which share specific functions uncommon to the family as a whole may lead to important information about the function and structure of a related protein of unknown function , as well as about the functional diversification acquired by the family during evolution . This work’s purpose is to automatically detect isofunctional subfamilies in a protein family of unknown function , as well as identify residues responsible for differentiation . We integrate data and then provide it to a clustering algorithm , which creates clusters of similar proteins we found correspond to same-specificity subfamilies . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"sequencing",
"techniques",
"dehydration",
"(medicine)",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"enzymes",
"enzymology",
"serine",
"proteases",
"signs",
"and",
"symptoms",
"protein",
"structure",
"lyases",
"molecular",
"biology",
"techniques",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"sequence",
"alignment",
"proteins",
"protein",
"kinases",
"molecular",
"biology",
"protein",
"structure",
"comparison",
"biochemistry",
"diagnostic",
"medicine",
"adenylyl",
"cyclase",
"protein",
"domains",
"biology",
"and",
"life",
"sciences",
"proteases",
"macromolecular",
"structure",
"analysis"
] | 2016 | Isofunctional Protein Subfamily Detection Using Data Integration and Spectral Clustering |
One of the striking features of evolution is the appearance of novel structures in organisms . Recently , Kirschner and Gerhart have integrated discoveries in evolution , genetics , and developmental biology to form a theory of facilitated variation ( FV ) . The key observation is that organisms are designed such that random genetic changes are channeled in phenotypic directions that are potentially useful . An open question is how FV spontaneously emerges during evolution . Here , we address this by means of computer simulations of two well-studied model systems , logic circuits and RNA secondary structure . We find that evolution of FV is enhanced in environments that change from time to time in a systematic way: the varying environments are made of the same set of subgoals but in different combinations . We find that organisms that evolve under such varying goals not only remember their history but also generalize to future environments , exhibiting high adaptability to novel goals . Rapid adaptation is seen to goals composed of the same subgoals in novel combinations , and to goals where one of the subgoals was never seen in the history of the organism . The mechanisms for such enhanced generation of novelty ( generalization ) are analyzed , as is the way that organisms store information in their genomes about their past environments . Elements of facilitated variation theory , such as weak regulatory linkage , modularity , and reduced pleiotropy of mutations , evolve spontaneously under these conditions . Thus , environments that change in a systematic , modular fashion seem to promote facilitated variation and allow evolution to generalize to novel conditions .
The origin of the ability to generate novelty is one of the main mysteries in evolution . Pioneers of evolutionary theory , including Baldwin [1] , Simpson [2] , and Waddington [3] , [4] , suggested how useful novelty might be enhanced by physiological adaptations and by the robustness of the developmental process . These early theories were limited by a lack of knowledge of the molecular mechanisms of development . Recent decades saw breakthroughs in the depth of understanding of molecular and developmental biology . Many of these findings were unified in the theory of facilitated variation [5] , presented by Kirschner and Gerhart , that addresses the following question: how can small , random genetic changes be converted into complex useful innovations ? In order to understand novelty in evolution , Kirschner and Gerhart integrated observations on molecular mechanisms to show how the current design of an organism helps to determine the nature and the degree of future variation . The key observation is that the organism , by its intrinsic construction , biases both the type and the amount of its phenotypic variation in response to random genetic mutation [3] , [4] , [6]–[10] . In other words , the organism seems to be built in such a way that small genetic mutations have a high chance of yielding a large phenotypic payoff . To understand FV , it is important to compare it to the related concept of evolvability . A biological system is evolvable if it can readily acquire novel functions through genetic changes that help the organism survive and reproduce in future environments [11] . Evolvability is composed of two aspects: 1 ) variability: the capacity to generate new phenotypes 2 ) fitness: the fitness of the new phenotypes in future environments . Most studies of evolvability focused on the first aspect , variability . Such studies measured the range and diversity of the phenotypic variation that can be generated by a given mutation , usually without discerning between potentially useful phenotypes and non-useful ones [12]–[16] ( for an interesting exception see Ciliberti et al [17] ) . FV theory adds to previous considerations by focusing on the nature of the generated variation , and specifically on the organism's ability to generate novel phenotypes which are potentially useful . Facilitated variation ( FV ) is made possible by certain features of biological design . One of these is the existence of ‘weak regulatory linkage’ [5] , [10] , [18] , where general and non-instructive signals can trigger large pre-prepared responses . For example , changes in growth hormone concentration at a localized position ( limb bud in an embryo ) can trigger large useful changes in the shape of the limb , driven by the conserved mechanisms for growth of bones , muscles , blood vessels , and nerves [19] . A good example is the ease of changing beak shapes with any of many possible mutations that affect the concentration of a single morphogenic factor [20] ( Figure 1A ) . In weak regulatory linkage , the information about the output is pre-built into the regulated system without instruction from the regulator , which only selects between states . Such regulatory organization reduces the constraints for evolving new regulations and for generating complex potentially useful phenotypes . An additional feature that is important for FV is modular design [21]–[24] , seen for example , in the highly conserved body-plan of the embryo [25] , [26] and in the compartmental organization of gene regulation and signaling networks [27] . Modularity helps to relieve the concern that a mutation might interfere with many different parts of the organism . With properly designed modularity , variation within each module can be generated without harming other modules [28]–[31] . Facilitated variation can be in principle studied experimentally , for example by generating mutants and scanning the types of phenotypes generated . For example , a study on mutants of the lac regulatory region indicated that the shape of the gene input function is channeled in directions of AND-like and OR-like functions , rather than other possibilities [32] . An open question is how does FV spontaneously evolve ? It is not clear how selection in a present environment can lead to designs that increase the probability of useful changes in future environments . How does evolutionary theory account for the emergence of special designs that make it easy to generate novel and useful variation ? The key point in our study is the observation that environments in nature do not vary randomly , but rather seem to have common rules or regularities [33]–[35] . Specifically , environmental goals faced by organisms or molecules may be thought of as composed of a combination of subgoals [33] . When environments change , the organisms encounter a new goal that is still made of the same or similar subgoals . For example , on the level of the organism , the same subgoals , such as digesting food , avoiding predation , and reproducing , must be fulfilled in each new environment but with different nuances and combinations . On the level of cells , the same subgoals such as adhesion and signaling must be fulfilled in each tissue type but with different input and output signals . On the level of proteins , the same subgoals , such as enzymatic activity , binding to other proteins , regulatory input domains , etc . , are shared by many proteins but with different combinations in each case . One may thus propose that in many cases , the different possible environments share a language of modularity , in the sense that they are all made of certain combinations of a set of subgoals . We thus test the possibility that under such patterned varying environments , the organism can learn over many generations the language common to the environments encountered in its past . We ask whether FV arises in such systematically varying environments , by measuring the ability of simple model systems to adapt to new , previously unseen goals , which are in the same language as past goals . We employ two well-studied model systems: combinatorial logic circuits [33] , [34] and RNA secondary structure [12] . We find that the standard experiment of setting a goal which remains constant over time leads to highly optimized systems that show little FV . In contrast , FV is readily generated under modularly varying goals ( MVG ) , in which goals change over time but share the same subgoals [33] . We find that MVG evolution enhances the ability to generate novel phenotypes as long as novelty is modular: phenotypes with novel modules or novel combinations of modules . We show that organisms under MVG store information about past goals in their genomes , and evolve weak linkage that allows small genetic changes to unleash large phenotypic responses that do not ruin the modular structure of the organism . Our study thus suggests that environments that change in a systematic fashion promote the evolution of facilitated variation , and leave an imprint on the evolvability properties of the organisms , allowing them to generalize to new conditions that are in the same language as past conditions .
In the following , we mainly focus on two representative problems , logic circuits evolved towards combinations of XOR and EQ goals , and RNA molecules evolved towards cloverleaf-like RNA structure . Similar conclusions were found for all six Boolean goals studied and five other RNA structures tested , as detailed in Text S1 sections 1 . 1 and 1 . 2 . Under MVG evolution , the evolving circuits or RNA molecules were exposed to a series of goals that are related to each other by their shared set of subgoals . We find that within a few thousand generations , genomes evolve that are able to adapt rapidly , often within a single generation , to each new goal ( Figure 1B and 1C ) . Despite the fact that the phenotypic adaptation is large ( e . g . an entire hairpin changes to an unstructured open loop , or a change in about half the bits in the truth table of a circuit goal , see Methods ) , the adaptation is associated with a very small genetic change , usually only 1–2 mutations . In contrast , adaptation of organisms evolved under FG is slow when the goal is suddenly switched , even if the switch is to a goal with the same subgoals as the previous goal . FG-organisms take a dozen times more generations to satisfy the new goal ( Figure 3A ) , and require about five times more mutations on average , than organisms evolved under MVG . The same is true for the other goals tested in Text S1 . Thus , the response to changing goals is significantly slower than the response of MVG-evolved organisms to previously seen goals ( Figure 3A ) . We next asked what is special about the design of MVG-evolved organisms that facilitates their response to changing goals ? For this purpose , we considered the phenotypic neighborhood [37]–[39] , defined as the set of phenotypes that are accessible from a given genotype by a single point mutation . We find that the phenotypic neighborhood of MVG-evolved genomes includes phenotypes that have high fitness to the past goals seen in their history ( Figure 3B ) . This indicates that the evolved organism effectively remembers its past goals by storing information about it in its genome . In contrast , in genomes evolved under constant conditions ( FG ) , the fitness of the neighborhood for new goals is significantly lower . FG populations are known to evolve toward the center of the neutral network , defined as the set of all genotypes with the same phenotype that are connected by neutral mutations [40]–[42] . Thus the FG organisms are more robust to genetic mutations and their phenotypic neighborhood exhibits a lower degree of variation than the MVG organisms . These features are also found in the present study ( Text S1 , section 4 . 1 ) . In contrast , MVG organisms seem to be located at the edge of the neutral network that is closest to the neutral networks of the previously seen goals . This implies that temporally varying environments push populations towards special regions of the neutral network . In addition to genetic mutations , one can also study thermal fluctuations that give rise to alternative structures encoded by a single genotype [12] . Thus , in the RNA model , we considered in addition to the genetic neighborhood also the thermodynamic neighborhood: the set of structures for a given genome that have a free energy that is within 5kT of the minimal free energy ( MFE state ) and are therefore accessible with a non-negligible probability by thermal fluctuations [43] . We find that the thermodynamic neighborhoods of MVG-evolved genomes include structures that have high fitness for previously seen goals . The FG-evolved genomes we have tested have a thermal neighborhood whose fitness for new goals is significantly lower . In this respect , the thermodynamic neighborhood is similar to the genetic neighborhood ( Figure 3C , and Text S1 section 5 . 2 ) , a phenomenon called ‘plastogenetic congruence’ [12] ( Text S1 section 4 . 1 ) . We find that the rapid adaptation to previously seen goals in MVG organisms is facilitated by key positions in the genome that can stabilize a desired sub-structure or module among other potential outcomes . We term these positions ‘genetic triggers’ , since they can trigger a large and prepared phenotypic response . To detect genetic triggers one must search for genomic positions that vary in a way that is highly correlated to the change in the goals . This means that triggers carry high information content about the current goal . The genetic triggers can thus be detected by evaluating the mutual information between the environment ( goal ) and the genomic content at each position ( see Methods ) . Since mutual information measures how much the knowledge of one variable reduces the uncertainty regarding the other , the trigger positions are characterized by high mutual information with the environment ( Figure 4A ) . Trigger positions were readily detected for all MVG cases tested . In the RNA model , we find that mutual information is spread amongst more genomic positions than in the logic circuit model . Triggers can still be clearly detected at sites with much higher mutual information than the background . We find that these trigger nucleotides are positioned within the module that they affect , usually in the stem of a hairpin ( Figure 4C and 4D ) . In this respect , the hairpins evolved in MVG differ from hairpins evolved in FG in that a single change in the trigger can cause a flip between an open loop and a closed hairpin . Over time , under MVG conditions , it is evident that the mutual information between genomes and goals ( i . e . environments ) gradually becomes focused to a few trigger positions , allowing rapid adaptation when environment changes ( Figure 4B ) . Since trigger positions are small variations that lead to a sizable switch between pre-designed states , they may be considered as a simple example of weak regulatory linkage . So far , we analyzed the adaptation to previously seen goals introduced along MVG evolution history , which highlighted the ability of MVG organism to remember its past . We now turn to novel , previously unseen goals , where we test the ability to generalize based on the past . The main problem is to define what kind of novel goals might be encountered in future environments that are in the same context as the previous environments . Indeed , adaptation of MVG-organisms toward a randomly picked goal results in evolution that is as slow , or even slower , than FG-organisms ( Text S1 section 6 . 5 ) . But a randomly picked goal has no correlation with the past . To address this , MVG evolution offers the possibility of presenting a previously unseen goal which is in the same ‘language’ as previous history . This language , in the present case of logic circuits , is defined as the set of all goals that can be decomposed in the following way u ( x , y , w , z ) = f ( g ( x , y ) , h ( w , z ) ) , Figure 5A . In other words , the goals in the language are made of a hierarchy of three functions f , g and h , such that g responds to x and y , and h responds to the other two inputs w and z , and f responds to g and h . In the case of the RNA model , the language can be defined as the set of all secondary structures with independent structural modules ( e . g . , hairpin loops , open loops etc . ) that correspond in their genomic positions to the modules of the MVG goals ( see Methods and Figure 5B ) . Within this language , we defined two classes of possible future goals which are novel: ( a ) New-comb is a goal that presents previously seen subgoals but in a new combination ( Figure 6A ) ( b ) Novel-module refers to goals where one of the subgoals is a previously unseen one , while the other subgoals are kept unchanged ( Figure 6C ) . This represents a novelty that is restricted to one of the modules of the goal . We tested evolution under these two classes of novelty . We find that for both logic circuit and RNA models , MVG populations adapted faster than FG populations when introduced to new-comb goals ( Figure 6B ) . We also performed competition experiments in which initial populations were composed of 50% FG-evolved and 50% MVG-evolved genomes . When new-comb goals were presented , the descendants of MVG-evolved genomes took over the population in about 68% of the RNA model runs ( Figure 6B inset ) . Logic circuits showed similar behavior , where MVG-genomes took over the population in about 75% of the runs ( Text S1 section 6 . 3 ) . We also tested novel-module goals . Here , the RNA model did not show a significant difference between FG and MVG genomes . However , in the logic circuit model , MVG-populations adapted significantly faster also to novel-module goals ( Figure 6D ) . We tested 20 different novel-module goals . For example , a novel goal is generated by replacing a XOR module by a previously unseen 2-input Boolean function , such as AND or NOR defined by its truth table ( Figure 6C ) . We find that MVG's outperformance occurred only toward goals within the modularity language . MVG adaptation toward non-modular goals was not significantly different from FG's ( Figure 6E ) . In competition experiments [44] between FG and MVG genomes toward novel-module goals , populations were taken over by MVG-genomes in about 70% of the runs ( Figure 6D inset ) . In experiments toward randomly chosen goals , populations had equal chance to be taken over by either FG or MVG genomes ( Figure 6E inset ) . We further find that the harder the novel-module goal ( the more generations needed to solve it ‘from scratch’ ) , the more MVG organisms out-perform FG organisms ( see Text S1 section 6 . 4 ) . These results imply that temporally patterned environments not only lead to a memory of the past goals , but also to generalization: the population learned a language of its history of environments ( conditions ) that share the same common rules . To examine the mechanisms for enhanced evolution of novelty within the MVG language , we tested three suggested mechanisms proposed in the theory of FV [5] ( a ) mutations have large effect on their own module . This reduces the number of steps to novelty; ( b ) mutations have small effect on other modules , a property also called reduced pleiotropy [45] , [46]; and ( c ) mutations have reduced lethality , increasing viable genetic variance in the population and allowing access to higher diversity of potential phenotypes . We quantified the effects of mutations according to these suggestions . The results demonstrate that MVG organisms in the present study follow the first two mechanisms , but not the third . We begin with the first two mechanisms , and treat the third in the next section . To quantify the effect of mutations on their own module and on other modules , we mutated each of the genome positions that correspond to a given module in the phenotype , and tested its phenotypic effect on its own module and on the other modules . The effect of the mutation was quantified as phenotypic distance: Hamming distance between the structures of subsequences in the case of RNA , and between the series of outputs of the gates ( over all input combinations ) within each module in the case of logic circuits ( see Methods ) . The results are summarized in Table 1 . Significantly enhanced intra-module change and reduced pleiotropy were found in most cases . The two models differed in the extent of these mechanisms: logic circuits showed more reduced pleiotropy , and RNA structures primarily showed more enhanced intra-module change . We now turn to the third mechanism for novel adaptation proposed by FV theory , associated with an increase in the genetic variance of the populations . We evaluated the genetic variance in a population given its current goal by measuring the conditional genomic entropy ( Methods , Text S1 section 11 , [47] ) . In contrast to the suggested mechanism of FV theory , we find that MVG populations display lower genetic variance than FG populations ( Figure 7A ) . The reduction in genetic variance indicates that the rapid adaptation of MVG populations in this study is not due to population diversity but rather in useful potential variation within each individual . Why do MVG-populations show a lower genetic variance ? One possibility is that they evolve to store information about past environments in their genome , placing constraints on the sequence ( strong stabilizing selection ) . To test this , we studied the effect of increasing the number of goals introduced over time in MVG . We find that the more goals ( or more precisely the higher the information content in the environment ) , the lower the genetic variance in the population ( Figure 7B ) . Organisms evolved in constant environment seem to store less information and have higher genomic entropy ( Figure 7A ) . An additional way to understand the low variance in MVG genomes compared to FG genomes is to consider that the latter are more robust to genomic mutations ( see Text S1 section 4 . 1 ) . Hence , they display more positions in the genome that can be varied without affecting the phenotype . Robustness to mutations thus allows higher genetic variance in the population [48] , and conversely , strong constraints on the genome lead to lower genetic variance and sensitivity to mutations in MVG organisms . As an example for storage of information in the genome and its effect on genetic variance , consider the example of Figure 7C and 7D . Populations of RNA molecules that evolve toward a fixed secondary structure G1 that contains an open loop are found to show high variance in the genomic positions that form the open loop . This is because forming a loop is relativity easy as there are few constraints for base-pairing . On the contrary , populations evolved under MVG environments in which the goal repeatedly switched between G1 and G2 ( Figure 7C ) , show lower variance in the corresponding “loop region” . The evolved MVG loop carries information about its past , and is ready to become a stem by a single ‘trigger’ mutation . The information acquired by the loop is reflected in the pronounced decrease in the variance of that genomic region in MVG populations ( Figure 7D ) . We note that increase in variance might be expected in more complex models , especially when spatial heterogeneity can allow several metapopulations to exist by using recombination as an efficient adaptation mechanism . High variance may also occur if the genomes can not store the required information ( see Text S1 section 10 . 2 for an example ) . We find an additional property of MVG-evolved genomes which helps to overcome barriers to novelty and further reflects the ability to generalize , in the case of logic circuits . In a preceding section , we showed that the MVG phenotypic neighborhood is enriched with phenotypes that are close to previously seen goals . We now turn to possible future goals . We scanned the phenotypic neighborhoods for goals within the same modularity language as previous goals . We find , in the case of logic circuits , that the phenotypic neighborhood of a MVG-circuit is enriched with modular circuits that compute decomposable ( modular ) functions that are of the form u ( x , y , w , z ) = f ( g ( x , y ) , h ( w , z ) ) , ( Figure 8 ) . In contrast , the neighborhood of a FG-circuit includes more functions that are not decomposable and thus are not within this “modularity language” ( see Text S1 section 7 ) . This property was not found for the RNA model . Finally , we aimed to define a quantitative measure for facilitated variation . A desired measure should capture the two main components of biased variation: ( a ) the quantity component [41] , namely enriching of the phenotypic neighborhood with potentially useful phenotypes which are novel . ( b ) The quality component: accessing as many as possible different potentially useful novel phenotypes , which are as far as possible in phenotypic distance from the wild-type [49] . We chose a simple FV measure , among other possible choices , which is the product of these two components ( see Text S1 section 8 . 1 ) . The ‘quantity’ component is the probability of forming a potentially useful phenotype which is novel by a single point mutation; the ‘quality’ component is the average phenotypic distance between the wild-type and the potentially useful phenotypes within its phenotypic neighborhood . This measure is then normalized for its corresponding value with respect to non-useful neighboring phenotypes . Here , useful phenotypes correspond to phenotypes with the same modular structure as the goals in which the organism has previously evolved . Nuseful is the number of neighbors which have a modular phenotype ( useful ) and are different from the wildtype phenotype ( novel ) , and <d ( Puseful , P0 ) > is the mean distance between novel and useful phenotypes and the wildtype . Similar definitions apply for the denominator , where non-useful means phenotypes that do not have the modular structure of previous goals ( in the logic circuit model this includes either trivial functions such as an output of all ones or all zeros , or non-decomposable Boolean functions ) . According to the formula , an organism with high FV has a high likelihood of forming potentially useful variation and a relatively low probability of varying towards non-useful phenotypic directions ( see Methods and Figure 9A ) . We find that the FV measure increases with generations under both FG and MVG evolution ( Figure 9B and 9C , Text S1 section 8 . 2 ) . However , it increases significantly more under MVG . The increase under FG evolution seems to result from the increase in robustness ( increased probability of generating wild-type phenotype or close to wild-type phenotypes ) . Finally , we performed experiments in which the initial population consisted of genomes with high FV that were evolved by MVG . We then placed this population under a fixed goal , corresponding to their last seen goal , but presented constantly over time . We find that FV decreased rapidly within a few tens of generations provided there is even a slight selection pressure for small circuit size ( Figure 9D , following [33] ) . This result demonstrates the role of modularly varying goals in preserving facilitated phenotypic variation in the face of more optimal , low FV circuits .
This study quantitatively examined facilitated variation in model systems and demonstrated that it is enhanced in modularly varying environments as compared to constant environments . When the environment varies in a modular fashion ( or , more generally , in a systematic manner ) , it is possible to define feasible future environments that belong to the same ‘language’ as past environment . Hence , one can define a context specific evolvability: the extent to which organisms can generalize and generate novelty that is useful in the context of feasible future environments . The present results suggest that adaptation to new goals in MVG relies on the evolvability properties of each individual [50] . The evolved organisms are intrinsically designed for a certain class of changes . Organisms that evolve under MVG develop weak linkage implemented by ‘trigger’ genomic positions that elicit a large phenotypic payoff upon minimal genetic investment . The triggers elicit substantial changes in one module and have low effect on other modules ( low pleiotropy ) . The genomes are such that their genomic neighborhood is enriched with a wide range of potentially useful phenotypes – useful in the context of the previous goals ‘learned’ by the organism . Thus , the evolved genomes carry information about past goals . This information effectively prepares the organism for the future , provided that future goals are related to past goals . The evolution of facilitated variation is time-scale dependant: if goals switched very rarely , it would be equivalent to a succession of FG's . On the other hand , if goals switched too fast , the required information would not have the sufficient time to be assimilated . We find that in the case of the present models , the rate of environmental switching that gives rise to evolvable organisms spans several orders of magnitude [33] , [34] . This study employed two different models to study facilitated variation , logic circuits and RNA structures . Importantly , these two models differ in the type of modularity in their goals . RNA goals contained explicit structural modules ( e . g . hairpin loops ) . Every RNA structure that satisfies such goals is modular by definition . In contrast , the modularity in logic circuits goals is implicit . Circuits that satisfy a modular goal can have either a modular circuit structure or a non-modular one . Modular circuit structures are in fact much more rare , and tend to evolve only under MVG , where switching between goals with shared modules constrain the circuits to evolve structural modules [33] . This difference between RNA and logic circuit models may underlie the fact that logic circuits showed a very strong enhancement of facilitated variation in MVG compared to fixed goals , whereas RNA model had a more modest enhancement . These two models are approximations to different aspects of biological design: Cell signaling and regulation networks that compute responses to signals are more analogous to the circuit model , whereas molecular structures are akin to the RNA model . What happens if goals vary over time but in a non-modular fashion ? We find that an environment that varies between randomly chosen goals typically causes confusion , where no good solution is found that can rapidly adapt to both goals . It is possible , however , to find pairs of goals which are not modular and yet which have solutions that are only a few mutations away from each other . In other words , goals whose neutral networks happen to come very close at a certain point . Here , genomes evolve that show rapid adaptation each time that the goal switches , but do not have modular phenotypes . However , it is hard to define facilitated variation towards novel goals in this case , since one can not define the future goals that are in the same ‘language’ . Adaptation to novel goals is generally very poor ( see Text S1 sections 2 . 1 and 6 . 5 ) . In summary , evolution under non-modular varying environments might lead in certain cases to memory but not to generalization . Modularly varying goals seem to enhance facilitated variation because of two main effects ( i ) they greatly improve the chances for the existence of solutions for the different goals that are close in genetic space ( because the same modules need only be rewired by a few mutations ) ( ii ) they offer the possibility of learning not only past goals , but also generalize to future goals as long as they are made of the same subgoals or with the same division into modules as previous goals . Finally , we note that facilitated variation comes with a cost: organisms are less optimal to the current goal than they might have been . For example , logic circuits that have high FV are usually composed of more logic gates than the optimal circuits that evolve if this goal is kept constant for very long times . Modularity , genetic triggers , and storage of information about the past in the genome , seem to demand more genes than is absolutely required to solve the problem . Extreme optimality to present environments is sacrificed to provide readiness to future ones . Organisms or molecules that are under constant conditions [51] , [52] are predicted by the present theory to lose their FV design , and become less evolvable . One may test this prediction by comparing organisms that evolved in varying and relatively constant environments [51] , [52] . A further prediction is that any fluctuation in the system ( such as molecular noise [53] or thermal fluctuation ) would result in an output that is channeled in potentially useful directions . In summary , the present study aimed at studying facilitated variation in simple model systems . Populations evolved under systematically varying conditions were found to exhibit not only a memory of past goals but were also able to generalize to new conditions that are in the same language as previous conditions . Adaptation to useful novel goals was enhanced by organisms that have learned the shared subgoals that existed in past environments and are therefore likely to be encountered in future environments . Several elements of facilitated variation theory , such as genetic triggers , modularity , and reduced pleiotropy of mutations seem to evolve spontaneously under these conditions . It would be interesting to study the evolution of additional FV mechanisms such as exploratory behavior and body-plan compartmentalization using more elaborate models with hierarchical designs and developmental programs .
We used a standard genetic algorithm [54] , [55] to evolve combinatorial logic circuits and a structural model of RNA . The settings of the algorithm were as follows: a population of Npop individuals was initialized to random binary genomes of length B bits ( random nucleotide sequences of length B bases in the case of RNA , in the main examples B = 76 for the RNA and B = 104 for logic circuits ) . In each generation , Npop individuals were selected with repeats from the previous generation according to a probability that exponentially scales with their fitness ( selection strategy , see Text S1 section 1 ) . Pairs of genomes from the selected individuals were recombined , using crossover probability Pc ( Pc = 0 . 5 for the circuits model; Pc = 0 for the RNA model ) and then each genome was randomly mutated ( mutation probability Pm = 0 . 7/B per locus per genome ) . The present conclusions for the logic circuit model are generally valid also in the absence of recombination ( Pc = 0 ) . The present results were based on simulations of a population of size Npop = 5000 evolved for L = 105 generations for the circuit model , and a population of Npop = 500 evolved for L = 105 generations for the RNA model . These population sizes were empirically found to serve as minimal values for many of the presented effects , which seem to apply also for larger population sizes . For statistical analyses we considered only simulations that ended with maximal fitness of 1 within the predefined generation limit L . Similar conclusions were found when analyzing all runs . Circuits were composed of up to twelve 2-input NAND gates . The binary genome coded for the circuit wiring as described in [33] , [34] , [54] , [55] . Self loops and feedback loops were allowed . Goals were 4-input 1-output Boolean functions composed of XOR , EQ , AND , and OR operations . The goals were of the form u ( x , y , w , z ) = f ( g ( x , y ) , h ( w , z ) ) , where g and h were 2-input XOR or EQ functions , and f was an AND or an OR function [34] . Each Boolean function can be represented as a truth table , where each row represents a different combination of inputs values ( 0 or 1 ) , and the relevant output value ( again 0 or 1 ) . Thus each goal can be uniquely defined by the output column vector . The fitness of each circuit was defined as the fraction of correct outputs over all possible inputs . In the MVG simulations the goals were modularly related by changing the functions f , g or h . The goal changed over time in a probabilistic manner every E = 20 generations . We followed the work of Schuster [42] and Ancel and Fontana [12] and used standard tools for structure prediction available at http://www . tbi . univie . ac . at/RNA/ , and the “tree edit” structural distance [56] . The goals were secondary structures of length 60–90 nucleotides such as the Saccharomyces cerevisiae phenylalanine tRNA and synthetic secondary structures composed of three hairpins ( for the full list of structures see Text S1 section 1 . 2 ) . In MVG , the modular changes were applied by modifications of single hairpin at a time ( such as changing the shape of the hairpin to an open loop ) . Goals changed every E = 20 generations ( unless otherwise noted ) . Normalized fitness in Figures 3A , 6B , 6D , and 6E is defined as , where F is the maximal fitness in the population and Fr is the average maximal fitness of a population of Npop random genomes . Normalized fitness Fn = 1 means a perfect solution to the goal , and Fn = 0 means a solution that is as good as expected in a random population of the same size . For the purposes of computing the best fitness X of a genetic neighborhood of a given system with phenotype P , as in Figure 3B and 3C , we used a normalization in which Fr is the value of X averaged over Npop samples taken from genomes with the same phenotype P . In the case of logic circuits , genomes with the phenotype P were obtained by simulated annealing optimization algorithm which produced genomes that satisfy the desired goal . In the case of RNA structures , genomes with phenotype P were generated using a standard inverse fold algorithm [36] . The normalized fitness of the genetic neighborhood is . Following Adami et al . [48] , genetic variance was measured using entropy H computed as follows . In a RNA genome of length B , each position can hold one of the 4 possible nucleotides with the probabilities: Pi , j where i = 1 . . B and j = {C , G , A , U} . The entropy of position i is Hi = −ΣPi , jlog ( Pi , j ) . The maximal entropy per position ( using logarithm of base 4 ) is 1 , which occurs when the nucleotides distribution at that site is uniform . Perfectly conserved positions have zero entropy meaning that they contain maximal information ( see Text S1 section 11 . 1 ) . The nucleotide probabilities for each genomic position were computed from the population genomes . The genetic entropy is the sum of the entropies of all positions . We note that this is only an approximation of the full genomic entropy since we ignore the epistatic relations between positions . It is also important to note that this measure is not the marginal genomic entropy but the conditional entropy of the genome given its current environment ( for FG , the two measures coincide ) . For an example , see Text S1 section 10 . 1 . In order to detect the genetic triggers in a genome , we computed the mutual information I between target goal T and specific genomic site i , Xi , as I ( Xi , T ) = H ( Xi ) −H ( Xi |T ) where H is the entropy per site as described above [see Text S1 sections 10 and 11] . Triggers are defined by the positions with the highest mutual information ( I ) between goal and genomic contents . To define the effects of mutations on phenotype modules , we first computed the modules in each phenotype . For RNA this was based on the modular partition of the structure ( into hairpin loops etc . ) , and in logic-circuits , modules were defined using the Newman-Girvan algorithm [57] . We then measured the effects of each possible genomic mutation on the phenotype of its own module , and on the phenotype of all other modules . In the RNA model , the effect of mutations on the phenotype of each module was evaluated by the distance d between the wild-type and the mutant structure in each module ( Hamming distance between the string representations of the secondary structure [58] ) . In the case of logic circuits , the output series of each gate was evaluated , and the Hamming distance d between the mutant and wild-type was evaluated for each gate . Intra-module effects of mutations were the mean of all changes in the same module as the mutated gate , and inter-module effects of a mutation was the mean effect on the output of all gates in all other modules . The physical ranges of those effects were estimated by analyzing samples from the solution space ( obtained by optimization algorithms ) . To quantify the modularity of a network we used the normalized Qm measure of Kashtan et . al . [33] , [51] . | One of the striking features of evolution is the appearance of novel structures in organisms . The origin of the ability to generate novelty is one of the main mysteries in evolutionary theory . The molecular mechanisms that enhance the evolution of novelty were recently integrated by Kirschner and Gerhart in their theory of facilitated variation . This theory suggests that organisms have a design that makes it more likely that random genetic changes will result in organisms with novel shapes that can survive . Here we demonstrate how facilitated variation can arise in computer simulations of evolution . We propose a quantitative approach for studying facilitated variation in computational model systems . We find that the evolution of facilitated variation is enhanced in environments that change from time to time in a systematic way: the varying environments are made of the same set of subgoals , but in different combinations . Under such varying conditions , the simulated organisms store information about past environments in their genome , and develop a special modular design that can readily generate novel modules . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"computational",
"biology/evolutionary",
"modeling",
"evolutionary",
"biology/bioinformatics"
] | 2008 | Facilitated Variation: How Evolution Learns from Past Environments To Generalize to New Environments |
The budding yeast Saccharomyces cerevisiae alters its gene expression profile in response to a change in nutrient availability . The PHO system is a well-studied case in the transcriptional regulation responding to nutritional changes in which a set of genes ( PHO genes ) is expressed to activate inorganic phosphate ( Pi ) metabolism for adaptation to Pi starvation . Pi starvation triggers an inhibition of Pho85 kinase , leading to migration of unphosphorylated Pho4 transcriptional activator into the nucleus and enabling expression of PHO genes . When Pi is sufficient , the Pho85 kinase phosphorylates Pho4 , thereby excluding it from the nucleus and resulting in repression ( i . e . , lack of transcription ) of PHO genes . The Pho85 kinase has a role in various cellular functions other than regulation of the PHO system in that Pho85 monitors whether environmental conditions are adequate for cell growth and represses inadequate ( untimely ) responses in these cellular processes . In contrast , Pho4 appears to activate some genes involved in stress response and is required for G1 arrest caused by DNA damage . These facts suggest the antagonistic function of these two players on a more general scale when yeast cells must cope with stress conditions . To explore general involvement of Pho4 in stress response , we tried to identify Pho4-dependent genes by a genome-wide mapping of Pho4 and Rpo21 binding ( Rpo21 being the largest subunit of RNA polymerase II ) using a yeast tiling array . In the course of this study , we found Pi- and Pho4-regulated intragenic and antisense RNAs that could modulate the Pi signal transduction pathway . Low-Pi signal is transmitted via certain inositol polyphosphate ( IP ) species ( IP7 ) that are synthesized by Vip1 IP6 kinase . We have shown that Pho4 activates the transcription of antisense and intragenic RNAs in the KCS1 locus to down-regulate the Kcs1 activity , another IP6 kinase , by producing truncated Kcs1 protein via hybrid formation with the KCS1 mRNA and translation of the intragenic RNA , thereby enabling Vip1 to utilize more IP6 to synthesize IP7 functioning in low-Pi signaling . Because Kcs1 also can phosphorylate these IP7 species to synthesize IP8 , reduction in Kcs1 activity can ensure accumulation of the IP7 species , leading to further stimulation of low-Pi signaling ( i . e . , forming a positive feedback loop ) . We also report that genes apparently not involved in the PHO system are regulated by Pho4 either dependent upon or independent of the Pi conditions , and many of the latter genes are involved in stress response . In S . cerevisiae , a large-scale cDNA analysis and mapping of RNA polymerase II binding using a high-resolution tiling array have identified a large number of antisense RNA species whose functions are yet to be clarified . Here we have shown that nutrient-regulated antisense and intragenic RNAs as well as direct regulation of structural gene transcription function in the response to nutrient availability . Our findings also imply that Pho4 is present in the nucleus even under high-Pi conditions to activate or repress transcription , which challenges our current understanding of Pho4 regulation .
When environmental conditions change , the budding yeast Saccharomyces cerevisiae , like other microorganisms , makes a decision about growth , cell division , and which responses to elicit in a coordinated fashion . Starvation for nutrients , alterations in temperature or salt concentration , and the presence of toxic agents are critical stresses for yeast cells and elicit signals that evoke cellular responses favoring survival under the new conditions . Nutrient status is probably the most important condition that must be accurately and rapidly sensed and responded to in order to ensure cell survival . In this process , nutrient-sensing kinases including cyclic-adenosine-monophosphate-dependent kinase , Snf1 , Tor , and Pho85 kinases play important roles in regulation at levels ranging from transcription to the activity of individual enzymes [1 , 2] . Transcriptional regulation is the most fundamental process in the nutritional response , and DNA-binding transcription factors and genes under their control are extensively characterized by conventional genetic and biochemical analyses and , more recently , expression profiling via DNA microarray and chromatin immunoprecipitation ( ChIP ) -on-chip analysis . The PHO system is a well-studied case in which a set of genes ( PHO genes ) is expressed to activate inorganic phosphate ( Pi ) metabolism for adaptation to Pi starvation [3] . The Pho4 transcription factor that activates PHO genes is regulated by phosphorylation to alter its cellular localization: under high-Pi conditions , the Pho85 kinase phosphorylates Pho4 , thereby excluding it from the nucleus and resulting in repression ( i . e . , lack of transcription ) of PHO genes . Pi starvation triggers an inhibition of Pho85 kinase , leading to the migration of unphosphorylated Pho4 transcriptional activator into the nucleus and enabling expression of PHO genes [4–6] . Transcriptional regulation responding to nutrient change is also extensively studied in glucose repression and in amino acid starvation , cases in which a complex interplay between activators and repressors acting on the structural genes involved in the respective process is well documented [7 , 8] . Recent studies on transcriptional regulation have revealed the participation of novel regulators in addition to protein factors , specifically , an involvement of RNA in the regulation of protein expression responding to external signals including nutrient changes [9 , 10] . Prokaryotic mRNAs that change their conformation upon binding of specific metabolites can alter transcription elongation or translation initiation and are called riboswitches [11] . Noncoding ( nc ) RNAs including small inhibitory ( si ) , micro ( mi ) , and small nucleolar ( sno ) RNAs modify RNA species to regulate gene expression: siRNA and miRNA target mRNA to cause mRNA cleavage and inhibition of translation , respectively , whereas snoRNA targets rRNA . Numerous ncRNAs , however , have been found that do not show these known functions , including antisense ( AS ) RNAs and transcribed pseudogenes [10] . In S . cerevisiae , several ncRNAs involved in transcriptional regulation are reported: SRG1 intergenic RNA functions in repression of SER3 [12 , 13] , and an AS RNA in the PHO5 locus appears to facilitate PHO5 transcription upon activation [14] , whereas those in PHO84 and IME4 function in gene silencing in aging cells [15] and inhibition of transcription [16] , respectively ( Accession numbers for genes described in this article are listed in Table S1 ) . Recent large-scale cDNA analysis [17 , 18] and mapping of RNA polymerase II binding using a high-resolution tiling array [19] have identified a large number of intergenic , intragenic , and AS RNA species whose functions are yet to be clarified . The Pho85 kinase has a role in various cellular functions other than regulation of the PHO system via Pho4; these functions include nutrient sensing , cell cycle progression , stress response , and control of cell morphology [20–23] . Pho85 monitors whether environmental conditions are adequate for cell growth and represses inadequate ( untimely ) responses in these cellular processes [1] . When nutrient is sufficient , the kinase phosphorylates Gsy2 , a glycogen synthase [24] , and represses the expression of UGP1 , which encodes an enzyme that catalyzes the production of UDP-glucose for glycogen synthesis [25] . Both of these events lead to the down-regulation of glycogen synthesis . Pho85 also facilitates the degradation of Gcn4 , a transcription factor that activates genes involved in amino acid metabolism when amino acids are depleted [26] . In contrast , the known cellular function of Pho4 seems rather limited to the PHO system [3] . Recent microarray analysis , however , has demonstrated that some genes involved in stress response and various other metabolic functions are activated under Pi-limiting conditions [27] , implying that Pho4 may activate these genes . Indeed , Pho4 is required for G1 arrest caused by DNA damage [28] . These observations suggest that Pho4 facilitates stress response by activating genes involved in the process . The fact that overproduction of Pho4 causes growth arrest of yeast cells in the absence of Pho85 [29] supports the antagonistic function of these two players on a more general scale when yeast cells must cope with stress conditions . To explore the general involvement of Pho4 in stress response , we tried to identify Pho4-dependent genes by a genome-wide mapping of Pho4 and Rpo21 binding ( Rpo21 being the largest subunit of RNA polymerase II ) using a yeast tiling array . In the course of this study , we found that Pho4- and Pi-dependent AS and intragenic RNAs modulate Pi signaling , leading to stimulation of expression of PHO genes , which demonstrates that nutrient-regulated RNA species other than mRNA are functioning in nutrient-responsive pathways in yeast cells . We also found that Pho4 was involved in transcriptional regulation of stress-responsive genes , either positively or negatively , and in some cases independently of environmental Pi conditions , which challenges the current model of Pho4 regulation [4 , 5] .
To analyze the Pho4 binding sites in the yeast genome , we used two kinds of oligo-DNA arrays , an Affymetrix high-density oligo-DNA array harboring 25-mer oligonucleotides with 4-nucleotide spatial resolution ( high-resolution [HR] chip ) and an Agilent yeast whole genome 44K array that had 60-mer oligos with ca . 270-nucleotide spatial resolution ( low resolution [LR] chip ) . The complete datasets of the HR chip analysis are found in the NCBI GEO database ( http://www . ncbi . nlm . nih . gov/geo/ ) under accession number GSE13350 . Analysis with the HR and LR arrays revealed that Pho4 bound to 51 and 57 genes , respectively , under low- but not high-Pi conditions ( Table S2 ) . Thirty-five genes were common in the two analyses , and all but two of these ( KCS1 and SHE9 ) had a prospective Pho4 binding sequence ( CACGTG/T or CTGCAC ) in their upstream regions . For the two exceptions , a binding sequence was present within the ORF ( Table S2 ) . Among the 35 genes , 16 had already been reported as PHO genes , that is , genes regulated by Pho4 in a Pi-dependent manner , by genetic , biochemical , and microarray analyses [30 , 31] . Our results demonstrated that Pho4 actually binds to the upstream regions of these genes depending on Pi conditions . In addition to the known PHO genes , our analyses identified 19 genes as possible novel PHO-type genes . Among them , eight genes , PST1 , MNN1 , HOM3 , HOR7 , PTK2 , CBF1 , SUR1 , and GLN1 , showed the expected pattern of Rpo21-binding ( Table S2 ) . Representative results for PHO8 and PST1 are shown in Figure 1A: Pho4 binding to the upstream regions of these two genes depending upon Pi condition is demonstrated by the two different ChIP-on-chip analyses ( Figure 1A ) , Pi- and Pho4-dependent transcription by Rpo21 binding ( Figure 1A ) and by northern analysis ( Figure 1B ) , and a Pi-dependent in vivo binding of Pho4 to their promoter regions by gene-specific PCR of a chromatin-immunoprecipitated ( ChIPed ) fragment ( Figure 1C ) . Pi- and Pho4-dependent transcription of MNN1 , PTK2 , CBF1 , and PST1 was demonstrated by northern analysis ( Figure 1B ) . In vivo binding of Pho4 to the upstream regions of these genes depending on Pi conditions was demonstrated by PCR using ChIPed DNA as a template and primers specific to the respective gene , which was in good accordance with the results of ChIP-on-chip analysis ( Figure 1C and Table S2 ) . AIR2 serves as a negative control for Pho4 binding , because it has no prospective Pho4 binding sites in its ORF and its expression is not affected by Pi conditions or Pho4 . ARO9 shares a divergent promoter with SPL2 , a known PHO-gene . Expression of ARO9 appeared dependent on both Pi and Pho4 ( Figure 1B ) , and Pho4 bound to its promoter in a Pi-dependent fashion ( Figure 1C ) . Binding of Pho4 to the CYC3-CDC19 region appeared to be Pi-dependent ( Table S2 ) , but CDC19 expression appeared independent of Pi conditions or Pho4 ( Figure 1B and 1C ) , indicating that Pho4 binding to the upstream region of CDC19 did not play a major role in transcriptional regulation of CDC19 . ChIP-on-chip analyses with the HR and LR arrays revealed that 140 and 30 genes , respectively , showed Pi-independent binding of Pho4 , and among them , nine genes ( URA3 , MUC1 , CIS3 , ILV3 , PDC1 , YPS3 , YLR137W , HPF1 , and ASN1 ) were commonly detected . Each of these nine genes has the prospective Pho4 binding site in its promoter or ORF ( Table S3 ) . We focused on ILV3 , ASN1 , CIS3 , and YPS3 for further analysis . Analyses of the binding profiles of Pho4 to these genes by ChIP-on-chip analysis using the two different platforms are shown in Figure 2A . Although the binding profiles of Rpo21 in the wild-type ( wt ) or Δpho4 mutant exhibited some inconsistency with those expected from Pho4 binding profiles , northern analysis demonstrated Pi-independent but Pho4-dependent expression , either a decrease ( ILV3 and ASN1 ) or an increase ( CIS3 and YPS3 ) in the absence of Pho4 ( Figure 2B ) . Gene-specific PCR demonstrated that Pho4 bound to the upstream region of these genes in vivo irrespective of Pi conditions ( Figure 2C ) . These results raise possibilities that Pho4 can bind to a promoter under high-Pi conditions and that Pho4 binding can lead to transcriptional activation ( ASN1 and ILV3 ) or repression ( CIS3 and YPS3 ) . The former hypothesis challenges the current model of Pho4 regulation in which , under high-Pi conditions , Pho4 is excluded from the nucleus through phosphorylation by Pho85-Pho80 [5] . Therefore , we further analyzed whether the activity of the ASN1 promoter was dependent on Pho4 under high-Pi conditions by measuring reporter activity in cells grown in a high-Pi medium . The wt promoter was active under high-Pi conditions ( Figure 2D ) , and its activity level decreased to almost 50% when a prospective Pho4 binding site ( at −451 with A of ATG as +1 ) in the promoter was mutated ( ASN1mut ) . In the absence of Pho85 where Pho4 became active , the activity of the wt promoter was stimulated by 1 . 3-fold compared to that in the wt strain , but the ASN1mut promoter showed a level of the activity similar to that observed in the wt strain . In the absence of Pho4 ( Δpho4 ) , the activity of the wt ASN1 promoter decreased to 50% whereas that of the mutant promoter showed a similar level of the activity in the wt cells . Gcn4 activates ASN1 under amino-acid-starvation conditions [32] , and in the absence of Gcn4 ( Δgcn4 ) , the promoter activity was decreased drastically ( Figure 2D ) . These results further demonstrated that Pho4 can activate ASN1 regardless of Pi conditions , which requires the Pho4 binding sequences in its promoter . Our ChIP-on-chip analyses using two different platforms demonstrated the binding of Pho4 within the KCS1 and SHE9 ORFs depending on Pi conditions ( Figure 3A and Table S2 ) . Because , to our knowledge , gene transcription mediated by the binding of a yeast transcription factor within an ORF is very rare with only three precedents [33–35] , we further analyzed the regulation of KCS1 by Pho4 . Whereas gene-specific PCR using a primer set specific to the KCS1 promoter ( −908 to +70 ) failed to detect an enrichment of the Pho4-bound fragment ( Figure 3B , top panel ) , the ORF-specific primer ( +36 to +1102 ) could detect an enrichment of the Pho4-bound fragment prepared only from cells grown in low-Pi medium ( Figure 3B , second panel from the top ) . In the absence of Pho85 where Pho4 became active , Pho4 binding to the ORF was detected under both high- and low-Pi conditions ( Figure 3B , bottom panel ) whereas that to the upstream region was not ( Figure 3B , second panel from the bottom ) . These results demonstrated that Pho4 binding within the KCS1 ORF depends on Pi conditions . We also analyzed Rpo21 binding by ChIP-on-chip and found that Rpo21 was localized to the KCS1 locus in Pi- and Pho4-dependent manners ( Figure 3A , bottom two panels ) . This result indicated the presence of Pi- and Pho4-dependent transcription in the KCS1 locus . To examine whether binding of Pho4 within the ORF could direct KCS1 mRNA transcription , initiation of transcription within the ORF ( intragenic RNA ) , or synthesis of AS RNA , we carried out northern analysis using RNA probes specific to the sense or antisense strands of KCS1 ( Figure 3C ) . With an RNA probe hybridizing with the KCS1 mRNA ( Figure 3E ) , a transcript of ca . 3 , 200 nucleotides ( nt ) , approximately the size of the KCS1 ORF ( 3 , 150 nt; Figure 3E ) , was detected , which was not dependent on either Pho4 or Pi conditions ( Figure 3C , top panel , lanes 1 to 4 , designated by upper arrow ) . Judging from its size , this RNA species is highly likely to be the KCS1 mRNA , because this band was not detected in Δkcs1 cells ( unpublished data ) . In the absence of Pho85 , a transcript with a smaller size ( ca . 2 , 600 nt ) was detected ( Figure 3C , top panel , lanes 5 and 6 , designated by lower arrow ) , together with the KCS1 mRNA and short transcripts ranging from 2 , 300 to 1 , 800 nt in length ( lanes 2 , 5 , and 6 , designated by a vertical bar ) . These short sense transcripts were also detected in the wt cells under low-Pi conditions , though weakly ( lane 2 ) , but not observed in a Δpho4 ( lanes 3 and 4 ) or Δpho4 Δpho85 double mutant ( lane 7 ) . These results indicated that Pho4 binding within the KCS1 ORF can activate the transcription of RNAs shorter than the KCS1 mRNA in both Pi- and Pho4-dependent manners . It should be noted that the KCS1 mRNA level appeared reduced when these short RNA species were abundant in Δpho85 cells ( lanes 5 and 6 ) . On the other hand , an AS RNA probe covering from −715 to +295 of the KCS1 gene ( Figure 3E ) could detect a short transcript of ca . 500 nt , which was dependent on both Pho4 and Pi conditions ( Figure 3C , middle panel , lanes 1 to 7 ) , whereas that covering from −715 to −228 failed to do so ( unpublished data ) . Therefore , the AS RNA was highly likely to be encoded between +295 and −228 . The presence of AS RNA in PHO genes is reported in PHO5 [14] and PHO84 [15] , and in both cases , RRP6 affects the stability of the AS RNA . We tested the effect of a Δrrp6 mutation and found that the mutation also stabilized the AS RNA in the KCS1 locus while not significantly affecting the amount of sense RNAs ( Figure 3D ) . To examine whether the 2 , 600 nt transcript is a processing product and to confirm the presence of the AS RNA , we determined transcriptional start points of the sense and AS RNAs in the Δpho85 strain by the 5′ rapid amplification of cDNA ends ( RACE ) method and found that the sense RNAs started mainly at −14 and +537 and that the AS transcription started at +235 ( Figure 3E ) . The sense RNA starting at −14 is highly likely to be the KCS1 mRNA , because the transcription start points of the sense RNA in the wt were also mapped mainly to this point ( unpublished data ) . The one starting at +537 can be ca . 2 , 600 nt in length when transcribed through the ORF , the size of which coincides well with the estimated size of the short transcript detected by northern analysis ( Figure 3C , top panel , designated by lower arrow ) . This result supported the conclusion that the 2 , 600 nt RNA is not a processing product but is transcribed from within the KCS1 coding sequence . Three prospective Pho4 binding sites ( at +406 , +1127 , and +1193 ) are present in the 5′ half of the KCS1 ORF ( Figure 3E ) , and the one closest to the 5′ end is sandwiched by the transcription start points of the intragenic ( +537 ) and AS RNAs ( +235 ) , which suggests a possibility that binding of Pho4 to the +406 site activates both of the 2 , 600 nt sense and AS RNAs . We constructed the KCS1 mutant ( KCS1mut ) in which three prospective Pho4 binding sites were mutated while keeping the amino acid sequence intact and analyzed the wt or KCS1mut in a low copy ( YCp ) plasmid in Δkcs1 cells . We found that the KCS1 mRNA was normally produced ( Figure 4A , top panel , lanes 1 to 4 ) , whereas the AS and intragenic RNAs were produced from wt KCS1 but not from KCS1mut under low-Pi conditions ( second panel , lanes 2 and 4 ) . Both the AS and truncated sense RNAs were synthesized from wt KCS1 in Δkcs1 Δpho85 cells ( lanes 5 and 6 ) whereas KCS1mut produced RNA of the wt size but not the intragenic or AS RNAs ( lanes 7 and 8 ) . These results indicated that the generation of the 2 , 600 nt intragenic as well as AS RNAs are activated by Pho4 , which requires the presence of at least one of the prospective Pho4 binding sites . Short sense transcripts around 2 , 000 nt ( Figure 4A , lanes 2 , 5 , and 6 , designated by a vertical bar ) decreased to below detectable levels with KCS1mut ( lanes 7 and 8 ) , which suggested that these transcripts are also dependent on Pho4 . Antisense RNA functions in cis to inhibit sense RNA transcription by transcriptional collision as reported in the IME4 case [16] and in trans to form a hybrid with sense RNA to inhibit its function [9 , 10] . To reveal the role of the KCS1 AS RNA that is regulated by Pho4 and Pi conditions , we first asked whether the AS RNA could affect the synthesis of the Kcs1 protein by immunoblotting . Kcs1 protein whose C-terminus was tagged with the c-myc epitope was produced from the wt KCS1 or KCS1mut gene in a YCp plasmid ( Figure 4A , second panel from the bottom ) . Kcs1 protein of the wt size ( 1 , 143 amino acids including the myc tag ) was detected in all cases ( lanes 1 to 8 ) , and Δkcs1 Δpho85 cells harboring the wt KCS1 plasmid produced a truncated protein together with the normal one ( Figure 4A , second panel from the bottom , lanes 5 and 6 ) . The band observed between 83 and 62 kDa markers was nonspecific staining with anti-myc antibody because it was also detected in the absence of Kcs1-myc protein in the extract ( unpublished data ) . Because both the AS and truncated sense RNAs were produced in Δkcs1 Δpho85 cells , the AS RNA hybridizing to the 5′ region of the KCS1 mRNA might inhibit normal translation initiation or the truncated sense RNA might be translated , either of which could use the in-frame initiation codon at +676 ( Figure 3E ) to produce truncated Kcs1 protein composed of 825 amino acids plus 93 amino acids from the c-myc tag . To examine whether the AS RNA could act in trans , for example , by hybridizing to the 5′ region of the KCS1 mRNA , we constructed a plasmid in which the AS RNA was produced from the GAL1 promoter by placing a KCS1 fragment covering +295 to −950 downstream of the promoter and introduced it into a Δkcs1 strain harboring the KCS1mut gene , so that the AS RNA was provided only in trans and the short sense transcripts including the 2 , 600 nt intragenic RNA were not produced ( Figure 4A , top panel , lane 7 ) . The truncated Kcs1 protein was observed only when transcription of the AS RNA was induced from the plasmid in galactose medium ( Figure 4B , second panel from the bottom , lanes 9 and 10 ) . As expected , the short sense transcripts were not detected when the AS RNA was overproduced in trans ( Figure 4B , top panel ) . These results indicated that the AS RNA can act in trans ( i . e . , inhibition of the normal translation initiation , possibly by hybrid formation with the KCS1 mRNA ) . To test hybrid formation , we tried to detect the presence of the double-stranded ( ds ) RNA in the total RNA sample by RNase protection analysis using single-stranded RNA-specific RNase , followed by reverse transcription ( RT ) and PCR amplification of the protected fragments . Reverse transcription was carried out using either sense- or antisense-strand-specific primers , hybridizing to from +243 to +223 and from −16 to +5 ( sense and AS in Figure 4C , respectively ) . As shown in Figure 4C , dsRNA of ca . 250 bp in length protected from RNase digestion was detected in Δpho85 cells in which both sense and AS RNAs were present ( Figure 4C , second panel from the top , lanes 13 , 14 , 17 , and 18 ) . When the RNA sample was not digested with RNase , the sense strand was successfully amplified , whereas the AS strand was not ( lanes 16 and 20 ) , indicating that the KCS1 mRNA was present but the AS RNA was not in the RNA sample tested . In Δpho85 cells in which both sense and AS RNAs were present ( Figure 4C , lane 14 ) , dsRNA of ca . 260 bp in length protected from RNase digestion was detected ( Figure 4C , second panel from the top , lanes 13 , 14 , 17 , and 18 ) . The Δkcs1 mutant cells do not produce the KCS1 sense or AS RNAs , and accordingly the sense or AS primer failed to synthesize cDNA ( Figure 4C , bottom panel ) . When KCS1mut was expressed in Δkcs1 cells under low-Pi conditions , the sense RNA was produced but not the AS ( Figure 4A , lane 4; Figure 4C , the middle panel , lanes 16 and 20 ) , and therefore dsRNA was not detected ( lanes 13 , 14 , 17 , and 18 ) . Reciprocally , when the AS RNA was expressed in Δkcs1 cells , only the AS was detected ( Figure 4C , second panel from the bottom , lanes 16 and 20 ) , and the protected dsRNA fragments were , if any , below the detectable level ( lanes 13 , 14 , 17 , and 18 ) . To confirm that the protected dsRNA was specifically amplified , we carried out the RT reaction using primers hybridizing upstream of the transcription start points of the sense or AS RNAs ( at −14 and +235 , respectively , in Figure 3E , and those designated by asterisks in Figure 4C ) . The RNA samples prepared from the strains producing the KCS1 mRNA ( wt , Δpho85 , and Δkcs1 + KCS1mut ) could generate cDNA of ca . 430 bp in length when amplified with the sense*/AS primer pair only in the absence of RNase digestion ( Figure 4C , lanes 21 and 23 ) . Similarly , the AS*/sense primer pair could synthesize cDNA when the RNA samples containing the AS RNA were not digested by RNase ( lanes 24 and 26 ) . These results indicated that the protected dsRNA fragment was specifically transcribed and amplified by the RT-PCR reaction and therefore strongly suggested that the AS RNA can form a hybrid with the KCS1 mRNA in vivo . Such a hybrid may inhibit normal translation initiation , leading to the generation of the truncated Kcs1 protein . Although we demonstrated that the truncated Kcs1 protein could be generated independently of the short sense RNAs , it is still possible that the truncated protein is translated from the 2 , 600 nt intragenic RNA . To test this possibility , we constructed plasmids harboring N-terminally truncated KCS1 or KCS1mut ORF fragments ( +105 to +3150 ) and introduced them into Δkcs1 or Δkcs1 Δpho4 mutants . Because these KCS1 fragments lack the KCS1 promoter , the plasmids are unable to produce the full-length KCS1 mRNA . With a strand-specific RNA probe ( Figure 3E ) , we could detect Pi- and Pho4-dependent transcripts of ca . 2 , 600 nt in length ( Figure 4D , lanes 27 , 28 , 31 , and 32 ) . This transcript was not observed in the KCS1mut ( lanes 29 , 30 , 33 , and 34 ) , indicating that the presence of the Pho4 binding sites is required for the production of this RNA species . Thus Pho4 could activate transcription from downstream of its binding site in the KCS1 ORF . Although this downstream transcription was plasmid-borne , the observed similarity in the size of transcript and its regulation strongly indicate that the 2 , 600 nt transcript is actually transcribed in the chromosomal KCS1 locus in a Pho4-dependent fashion and not a processing ( or limited degradation ) product . This plasmid-derived transcript could produce protein that had a size similar to that of the truncated Kcs1 protein ( Figure 4E , lanes 36 and 39 ) , indicating that the 2 , 600 nt intragenic RNA can encode the truncated Kcs1 protein . Taken together , these results indicated that Pho4 binding within the KCS1 ORF provokes transcription of both of the AS and the 2 , 600 nt intragenic RNAs , which may lead to production of the truncated Kcs1 protein by alteration of translation initiation through the formation of a hybrid with the KCS1 mRNA and by translation of the 2 , 600 nt intragenic transcript . What is the biological relevance of Pi- and Pho4-dependent production of the AS and intragenic RNAs and consequently of the truncated Kcs1 protein ? KCS1 codes for inositol hexakisphosphate ( IP6 ) kinase synthesizing 5-diphospho myoinositol pentaphosphate ( 5-PP-IP5 ) [36] . The same substrate is used by another yeast IP6 kinase , Vip1 , to synthesize IP7 isomers , 4- or 6-PP-IP5 that function in Pi signaling in the PHO system [37 , 38] . Therefore , it is conceivable that a decrease in the Kcs1 activity can supply more substrate for Vip1 , thereby enhancing Pi signaling . The fact that overproduction of Kcs1 reduces the extent of PHO5 derepression whereas a deletion of KCS1 derepresses PHO5 under high-Pi conditions [39] supports this model . Although the levels of normal Kcs1 protein did not appear to be altered significantly in the presence of the AS RNA and 2 , 600 nt intragenic transcript ( Figure 4A ) , the formation of a hybrid RNA could affect the normal level of the KCS1 mRNA , which may cause a slight difference in the level of the Kcs1 protein not detectable by western analysis . In addition , the presence of the truncated Kcs1 may perturb normal function of Kcs1 . To test these hypotheses , we analyzed the effects of the AS RNA and the intragenic sense transcript on PHO84 and PHO5 expression by northern analysis ( Figure 5 ) . PHO84 responds more quickly to a change in Pi conditions than PHO5 [40] . In the wt cells , PHO5 and PHO84 were expressed only under low-Pi conditions ( Figure 5 , lanes 1 and 2 ) but not in the absence of Pho4 ( lanes 5 and 6 ) . The two genes were expressed under high-Pi conditions in Δkcs1 cells ( Figure 5 , lanes 7 and 8 ) as reported [39] , which was suppressed by the wt KCS1 in a YCp plasmid ( lanes 9 and 10 ) . This result indicated that the plasmid-borne Kcs1 is functional . KCS1mut that produced neither the AS RNA nor the intragenic transcript ( Figure 4A ) showed a decreased expression level of the two genes under low-Pi conditions ( Figure 5 , lane 12 ) , suggesting that the low-Pi signal was not transmitted sufficiently to activate Pho4 . On the other hand , overexpression of the AS RNA in the wt cells resulted in a significant derepression of the two genes under high-Pi conditions ( lanes 1 and 13 ) , suggesting that the low-Pi signal was transmitted to activate Pho4 under high-Pi conditions in this case . This stimulatory function of the AS RNA was dependent on the presence of Vip1 IP6 kinase ( lane 19 ) that functions in low-Pi signal transmission [38] . This result further supported the conclusion that the AS RNA functions in the low-Pi signal transduction pathway . We also overproduced a truncated Kcs1 protein ( dKcs1 , +670 to +3150 ) in the wt cells , which caused derepression of PHO84 , albeit weakly , and barely detectable expression of PHO5 under high-Pi conditions ( lane 15 ) . The different expression levels of PHO84 and PHO5 can be attributed to different responsiveness of the two genes against the change in the environmental Pi level [40] . We also assayed the activity of acid phosphatase encoded by PHO5 in the strains with a combination of various plasmids as tested in northern analysis and found that the levels of the enzyme activities correlated with the mRNA level ( unpublished data ) . These results indicated that the presence or absence of the AS RNA and the 2 , 600 nt intragenic RNA cause altered regulation of PHO5 and PHO84 responding to Pi conditions , and therefore it is likely that the Kcs1 activity was modulated by the RNAs , the truncated Kcs1 protein , or both . The apparently weak effect on Pi signaling of dKcs1 compared to that of the AS RNA suggests that the truncated protein is not solely responsible for the stimulation of the low-Pi signaling . The AS RNA could play a certain role in this stimulation process , possibly through modulation of the KCS1 mRNA and protein levels . Thus , Pho4 appears to enhance low-Pi signaling by expressing the AS and the intragenic RNAs from within the KCS1 ORF , thereby constituting the positive feedback loop in the Pi signaling pathway .
Judging from the Gene Ontology terms of these 18 newly recognized PHO-type genes ( Saccharomyces Genome Database [SGD] , http://www . yeastgenome . org/ ) , they apparently do not have any functional relationship to either Pi metabolism or Pi signaling and are not categorized in a specific functional group ( Table S4 ) . Their expression profiles by global analysis , however , showed some similarities in that 12 of them are induced by either nitrogen depletion or amino acid starvation [32 , 45] and 9 of them are induced in the stationary phase [45] . This raises the possibility that Pho4 is involved in the regulation of a certain set of genes that responds to these nutrient-limiting or stress conditions . Pho4 is also reported to activate the transcription of genes involved in G1 arrest caused by DNA damage [28] . Thus , Pho4 appears to activate the transcription of genes responding to various stress conditions . This notion implies possible cross talk between Pi starvation and other stress conditions , the requirement of function of some , if not all , of these 18 genes in the adaptation of yeast cells to Pi starvation , or both . The results in this paper imply that Pho4 is present in the nucleus even under high-Pi conditions to activate or repress transcription ( Figure 2 ) , an implication that challenges our current understanding of Pho4 regulation . If the current model were correct , then Pho4 should somehow avoid phosphorylation by Pho85 , or if phosphorylated , then the modified Pho4 should have much less affinity to the Msn5 exportin to remain in the nucleus . Recently , Zappacosta et al reported Pi-dependent phosphorylation of Pho4 at Ser242 and Ser243 by a kinase other than Pho85 [46] . Phosphorylation of these two sites , however , appears less dependent on Pi than that at those sites modified by Pho85 ( i . e . , Ser at 100 , 114 , 128 , 152 , and 223 ) [46] . We could imagine that , under high-Pi conditions , prior phosphorylation of the Ser242 , Ser243 , or both by this unknown kinase could prevent phosphorylation of the other Ser residues by Pho85 , thereby decreasing the affinity of Pho4 for Msn5 while increasing the affinity to the target promoter , including ASN1 . Alternatively , Pho4 modified at Ser242 , Ser243 , or both might have more affinity to a yet unknown factor than to the exportin , and the resulting complex might be recruited to the target promoter regardless of phosphorylation by Pho85 . The Pho4 transcription factor appeared to repress CIS3 and YPS3 , both cell wall constituents . Expression of CIS3 is repressed by nitrogen starvation and in the stationary phase , implying that Pho4 can function as both an activator and a repressor under these stress conditions . The functioning of a yeast transcription factor as both an activator and a repressor has precedents ( e . g . , Rap1 and Abf1 ) [47 , 48] . Transcriptional repression by the two factors is often accompanied by silent chromatin structure . In a separate paper , we have reported that Pho4 negatively regulates the expression of SNZ1 , a stationary phase-specific gene , and that this regulation is accompanied by alterations in chromatin structure evoked by Pho4 binding [49] . The mechanism underlying transcriptional repression of CIS3 and YPS3 by Pho4 is yet to be clarified , but we suppose that a similar mechanism with the SNZ1 may apply in these cases . We demonstrated the presence of Pi-regulated AS RNA in the KCS1 locus . A large-scale cDNA sequencing by Miura et al . revealed the presence of many AS RNA species [17] , including an AS RNA in the KCS1 locus transcribed from +293 to −43 . Although its start point is different a little from our result by 5′-RACE analysis ( Figure 3E ) , we think it highly likely that this AS RNA coincides with the Pi-regulated AS RNA that we have reported here . The Pi-regulated AS RNA in KCS1 , however , did not appear to be coregulated with the KCS1 mRNA in the wt cells ( Figure 3C ) , and this contrasts with the observation in higher eukaryotes that sense and AS pairs are frequently coregulated [50] . Although Miura et al . did not describe the regulation of the KCS1 AS RNA , they claim coregulation between the sense and AS RNA in the GAL10 locus . However , the fact that the fold induction of the AS RNA is much less compared to that of the GAL10 mRNA when cells are grown in galactose medium and that Gal4 dependency of the GAL10 AS RNA was not analyzed points out that more work is necessary to establish coregulation of the sense and AS RNA at the GAL10 locus . With respect to biological function of ncRNA in yeast , a noncoding intergenic transcript ( SRG1 ) , originating from upstream of SER3 on the same strand and activated by Cha4 transcription factor in the presence of serine [13] , inhibits the binding of activators to the SER3 upstream activating sequence and of TATA-binding protein to its TATA box , leading to repression of SER3 [12] . Yeast AS RNA has been reported in the IME4 locus , which is expressed only in the haploid state to inhibit the IME4 mRNA transcription by transcriptional collision and thereby determines cell fate ( i . e . , the entry into meiosis ) [16] . The AS RNA at the PHO5 locus is constitutively expressed at a low level from ca . 1 , 400 bp downstream of the PHO5 TATA box through its promoter and is proposed to increase chromatin plasticity to enhance histone eviction upon a shift to low-Pi conditions [14] . Those in the PHO84 locus are suggested to recruit and/or stimulate Hda1 histone deacetylase for silencing of PHO84 in aging yeast cells [15] . Although these AS RNA species are found in PHO genes , they have not been reported to be regulated by environmental Pi conditions to facilitate the activation of PHO genes . The KCS1 case presents a different situation from them in that the AS and intragenic RNAs are activated by Pho4 in response to Pi starvation and may modulate the level of Kcs1 IP6 kinase to enhance Pi signaling , thereby stimulating the activation of PHO genes . We observed a decrease in the KCS1 mRNA level when Pho4 binds to the KCS1 ORF under low-Pi conditions or in a Δpho85 mutant ( Figure 3C , lanes 2 , 5 , and 6 ) . This observation suggests a possibility that transcriptional elongation of KCS1 mRNA is inhibited directly by Pho4 binding within the ORF . However , the scenario may not be so simple , because the KCS1 mRNA transcription itself can interfere with Pho4 binding , as reported in the SER3/SRG1 case [12] . Alternatively , the AS RNA could cause transcriptional collision with the mRNA , and hybrid formation of the AS RNA with the mRNA could lead to degradation of the mRNA . Both of these events could lead to a reduction in the KCS1 mRNA level . The stimulation of low-Pi signaling by Pho4-dependent intragenic and AS RNA represents an autoregulation ( induction ) or positive feedback loop responding to Pi limitation that can be envisioned as follows . Upon Pi limitation , the low-Pi signal is transmitted to Pho81 , leading to inhibition of Pho85-Pho80 and thereby stimulating Pho4 migration into the nucleus [6] . Pho4 then activates transcription of the AS and intragenic RNAs in the KCS1 locus: the AS RNA could reduce the KCS1 mRNA level by hybrid formation and possible transcriptional collision , which can lead to stabilization of Pho4 binding , resulting in the production of more AS and intragenic RNAs and consequently of more truncated Kcs1 protein using the downstream ATG codon at +676 . These events could lead to down-regulation of Kcs1 activity , enabling Vip1 IP6 kinase to utilize more IP6 to synthesize 4- or 6-PP-IP5 functioning in low-Pi signaling . Because Kcs1 also can phosphorylate these IP7 species to synthesize 4 , 5- or 5 , 6-PP2-IP5 ( IP8 ) [37] , reduction of the Kcs1 level can ensure accumulation of the IP7 species to further stimulate low-Pi signaling , leading to complete inhibition of Pho85-Pho80 . When Pi becomes sufficient , this loop runs in an opposite way for efficient inactivation of Pho4 and consequent repression of PHO genes . Though putative Pho4 binding sequences are present at −464 and −154 in its promoter , VIP1 expression was dependent on neither Pi condition nor Pho4 ( unpublished data ) , as in the case of the KCS1 mRNA . Inositol polyphosphate ( IP ) plays an important role in intracellular signal transduction as second messengers . The absence of Kcs1 and Vip1 causes abnormal vacuolar function and cell morphology , respectively , suggesting that they bear important cellular function [37 , 51] . Therefore , it is reasonable that the genes involved in IP synthesis are not regulated directly by individual nutrients ( in this case , Pi ) but indirectly by AS and intragenic RNAs responding to the nutrient , so that signal transduction and normal cellular function are not easily perturbed by fluctuation in the status of an individual nutrient . Positive feedback in the PHO system is also suggested to function in switching of Pi transporters [52] , in which Spl2 , activated by Pho4 , down-regulates low-affinity Pi transporters , Pho87 and Pho90 , whereas high-affinity Pi transporter Pho84 is activated by Pho4 . When the intracellular Pi level increases via the high-affinity transporter , Pho4 is inactivated to switch the transporters . Our finding expands the role of the Pho4 transcription factor beyond the regulation of the PHO system . The current consensus view is that it is the master regulator of the genes involved in the system , in that Pho4 activates transcription of the structural genes composing the PHO system to coordinate cellular response to Pi starvation [3] . Our findings indicate that Pho4 can modulate the activity levels of the products of apparently non-PHO genes by activating antisense and intragenic RNA expression to stimulate low-Pi signal transduction . A Δpho85 deletion causes pleiotropic mutant phenotypes [1] , some of which could be based on otherwise dormant transcriptional initiation , either intergenic or intragenic , or on the AS strand , caused by hyperactive Pho4 . In fact , we have also found Pi- and Pho4-regulated AS RNA in the GTO3 locus and intragenic sense transcript in the SHE9 locus ( unpublished data ) . GTO3 encodes an omega class glutathione-S-transferase having glutaredoxin activity , which is suggested to maintain an adequate redox state of specific target proteins , not in the general defense against oxidative stress [53] . SHE9 , also known as MDM33 , encodes a mitochondrial inner membrane protein functioning to maintain mitochondrial morphology [54] . Although , at present , we are unable to elucidate whether the GTO3 AS RNA or the SHE9 intragenic transcript can affect the annotated function of the respective gene product , this line of work will lead us to uncover yet unknown protein functions in the cellular response to Pi starvation . Regulation of gene expression and function by these nonconventional RNA species that are regulated by nutrient signals need not be restricted to the PHO system . Other inducible systems including GAL , glucose repression , and various stresses may well have these RNA species regulated by corresponding signals . High-throughput cDNA sequencing by Miura et al . and other works using microarrays [17–19] have revealed the presence of many intergenic , intragenic , and antisense RNAs in the yeast transcriptome . The finding of nutrient-regulated RNAs that are not coding annotated proteins adds more complexity to the intimately wired transcriptional regulation system by which yeast cells adapt to alterations in environmental conditions . High-resolution mapping of transcription factor and RNA polymerase II binding and very recent development of DNA–RNA hybridization techniques [55] will help to identify these regulatory RNAs . The yeast system , which can be manipulated by an array of genetic tools and for which there exists a substantial body of genetic information , will be the best resource to explore the complexity of the genetic network , including ncRNA species that function in responding to external signals .
Standard yeast genetics and media were used as described [56] . For phosphate-limited medium , Yeast Nitrogen Base ( YNB ) without phosphate ( Q-Biogene ) was used instead of normal YNB ( SD medium ) and was supplemented with 0 . 2 μM or 2 mM sodium phosphate to make low- and high-Pi media , respectively . The yeast strains used in this work are listed in Table S5 . Standard Escherichia coli and yeast protocols were employed [56 , 57] . Plasmids and primers used in this work are listed in Tables S5 and S6 , respectively . A Δpho85::URA3 fragment [25] was used to disrupt the PHO85 locus of the BY4741 ( MFY371 ) strain , and successful disruption was confirmed by PCR and constitutive expression of acid phosphatase ( unpublished data ) . To disrupt the PHO4 locus , Pho4Δ-F and -R primers were used to amplify the LEU2 marker having PHO4 sequences ( from +1 to +100 and from +830 to 929 with A of ATG as +1 ) at its termini , and the resulting fragment was introduced into MFY371 . Successful disruption was confirmed by PCR and failure to express PHO5 . For disruption of the VIP1 and RRP6 loci , the adaptamer-mediated PCR method was employed to prepare the DNA fragments for disruption [58] . The detailed methods are described in Text S1 . Disruption of GCN4 is described elsewhere [49] . To construct PHO4-tagged strains , MFY376 and MFY377 , a fragment containing PHO4 tagged with His × 6 and Flag × 3 was amplified using primers Pho4-Flag-F and -R and pUG6H3Flag plasmid as a template [59] , followed by transformation of MFY371 and MFY373 , respectively . Rpo21 fragments tagged with His × 6 and Flag × 3 ( Rpo21-Flag-F and -R ) were used to construct MFY378 and MFY379 . To mutagenize the prospective Pho4 binding site in the ASN1 promoter and in the KCS1 ORF , a QuickChange II site-directed mutagenesis kit ( Stratagene ) and appropriate primers were used . The detailed methods are described in Text S1 . Successful mutagenesis and the whole sequence of the mutant ASN1 promoter and of the mutant KCS1 ORF were confirmed by DNA sequencing . The promoter ( −920 to −1 ) and ORF ( −1 to +3143 ) of KCS1 were amplified by PCR using MN1132/1133 and MN1134/1135 pairs , respectively , so that EcoRI-NcoI and NcoI-XhoI fragments containing the respective sequences were generated . The two fragments were then ligated through the NcoI site and introduced into pRS313 to generate the pMF1530 plasmid . The wt NcoI-BamHI ( +1875 ) fragment had been replaced by the mutant fragment that lacked the three prospective Pho4 binding sites prior to incorporation of the EcoRI-XhoI KCS1 fragment into pRS313 to generate the pMF1531 plasmid . To construct plasmids pMF1527 and pMF1529 producing the wt and mutant Kcs1 protein tagged with six copies of the c-myc epitope at their C-termini , respectively , the EcoRI-XhoI fragment containing the wt or mutant KCS1 sequence was introduced into pRS316 containing a 6 × myc sequence . Plasmids pMF1540 and pMF1560 overexpressing the KCS1 AS RNA were constructed by placing the KpnI-EcoRI ( +291 to −920 ) fragment downstream of the TDH3 and GAL1 promoters in the pRS323 plasmid , respectively . Plasmid pMF1563 overproducing N-terminally truncated Kcs1 protein ( dKcs1 ) was constructed by placing the NcoI-XhoI fragment ( +676 to +3150 ) that had been cloned by PCR downstream of the TDH3 promoter in pRS326 . Yeast cells producing tagged protein were cultivated in high- or low-Pi medium as described above to a cell density of A600 = 1 . 0–1 . 2 , and chromatin immunoprecipitation was carried out as described [60] . ChIPed fragments were amplified by the T7 RNA polymerase-mediated method ( T7RPM ) followed by cDNA synthesis and the ligation-mediated PCR ( LM-PCR ) method for HR and LR analysis , respectively , essentially as described [41 , 61] . For T7RPM , ChIPed DNA ( about 100 ng ) was dephosphorylated in the reaction mixture ( 30 μl ) containing 2 units of CIAP and 0 . 2 units of BAP at 37°C , followed by incubation at 50°C for 15 min each . DNA was purified using a MinElute Reaction CleanUp kit ( Qiagen ) and eluted from the column with 10 μl of elution buffer ( 10 mM Tris . HCl , pH8 . 0 ) . This cleanup method was used throughout the following procedure except for the cleanup of reactions containing RNA . Dephosphorylated DNA was then subjected to poly ( dT ) tailing reaction in a reaction mixture ( 20 μl ) containing terminal transferase buffer ( Roche ) , 1 . 25 mM CoCl2 , 2 . 5 μM dNTP , and 20 units of terminal transferase by incubating at 37°C for 15 min . T7A18B primer ( GCATTAGCGGCCGCGAAATTAATACGACTCACTATAGGGAG[A]18B , where B refers to C , G , or T ) was then annealed to the dT-tailed DNA by incubating at 94°C for 2 min , at 35°C for 2 min , and at 25°C , followed by extension reaction in a 50 μl reaction mixture containing 1 ng/μl tailed DNA , 0 . 5 mM dNTP , 1 unit of Klenow enzyme , and 5 units of Sequenase at 37°C for 60 min . DNA was then subjected to in vitro transcription using a T7 Megascript kit ( Ambion ) in 20 μl of reaction mixture at 37°C for 4 h , followed by cleanup with an RNEasy Mini kit ( Qiagen ) . One microliter of 100 μM T7 degenerate primer ( GGATCCTAATACGACTCACTATAGGAACAGACCACCNNNNNNNNN ) was added to the RNA product , which was incubated at 70°C for 10 min , then on ice for 2 min , followed by cDNA synthesis . About 500 ng of cDNA was subjected to a second round of in vitro transcription and subsequent cDNA synthesis , followed by labeling using an in vitro transcription labeling kit ( Affymetrix ) . Labeled cRNA was hybridized to Affymetrix high-density oligonucleotide arrays of S . cerevisiae whole genome ( Watson strand ) or of chromosomes 3 , 4 , 5 , and 6 , which were processed and analyzed as described [62] . For LM-PCR , phosphorylation of 5′ termini of ChIPed DNA fragments by T4 polynucleotide kinase and ATP was performed prior to the blunt-end reaction , followed by ligation of annealed linkers ( MN974 and MN975 ) at 15°C for 16 h . The resulting fragments were amplified by PCR using MN974 primer , followed by PCR labeling with Cy3-dUTP and Cy5-dUTP for ChIPed and whole cell extract DNA , respectively . The data were analyzed with ChIP Analytics 3 . 0 software ( Agilent ) . The procedures for RNA isolation , northern and immunoblot analyses , and assay for β-galactosidase were as described [25 , 43] . DNA probes were prepared by PCR using digoxigenin ( DIG ) -PCR labeling mix ( Roche ) . RNA probes were prepared by transcribing DNA fragment cloned in pSP72 or pSP73 ( Promega ) with T7 RNA polymerase and a DIG-RNA labeling mix ( Roche ) . Gene-specific PCR was performed using primers listed in Table S6 and ChIPed DNA fragments or DNA in the whole cell extract ( WCE ) fraction as template under cycling condition as described [49] . A 5′-Full-RACE kit ( TaKaRa ) was used to determine the transcription start points in the KCS1 locus with total RNA from Δpho85 cells grown in high-Pi medium . Phosphorylated primers MN915 and MN1190 were used as a reverse transcription primer for the start points upstream of the initiating codon and within the ORF , respectively , and MN1134 for the start point of AS RNA . Amplified fragments were cloned using a TOPO TA cloning kit ( Invitrogen ) according to the manufacturer's protocol , and the transcription start points were determined by DNA sequencing . RNase protection assay and RT-PCR were carried out essentially as described [57] . Total RNA was digested with RNase ONE ( Promega ) at 30°C for 1 h and was recovered by precipitation in the presence of ethanol . First-strand cDNA was then synthesized using a primer specific to the sense or antisense strand , followed by PCR amplification after inactivation of reverse transcriptase and addition of appropriate reverse primer . The PCR products were separated by electrophoresis on a 5% polyacrylamide gel . | How does a microorganism adapt to changes in its environment ? Phosphate metabolism in the budding yeast Saccharomyces cerevisiae serves as a model for investigating mechanisms involved in physiological adaptation . The nutrient inorganic phosphate ( Pi ) is essential for building nucleic acids and phospholipids; when yeast cells are deprived of Pi , genes required for scavenging the nutrient are activated . This activation is mediated by the Pho4 transcription factor through its migration into or out of nucleus . The Pi-starvation ( low-Pi ) signal is transmitted by a class of inositol polyphosphate ( IP ) species , IP7 , which is synthesized by one of two IP6 kinases , Vip1 or Kcs1 . However , the IP7 made primarily by Vip1 is key in the signaling pathway . Here we report that under Pi starvation Pho4 binds within the coding sequence of KCS1 to activate transcription of both intragenic and antisense RNAs , resulting in the production of a truncated Kcs1 protein and the likely down-regulation of Kcs1 activity . Consequently Vip1 can produce more IP7 to enhance the low-Pi signaling and thus form a positive feedback loop . We have also demonstrated that Pho4 regulates , both positively and negatively , transcription of genes apparently uninvolved in cellular response to Pi starvation and that it sometimes does so independently of Pi conditions . These findings reveal mechanisms that go beyond the currently held model of Pho4 regulation . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"cell",
"biology",
"molecular",
"biology",
"genetics",
"and",
"genomics"
] | 2008 | Nutrient-Regulated Antisense and Intragenic RNAs Modulate a Signal Transduction Pathway in Yeast |
Schistosomiasis , a neglected global pandemic , may be curtailed by blocking transmission of the parasite via its intermediate hosts , aquatic snails . Elucidating the genetic basis of snail-schistosome interaction is a key to this strategy . Here we map a natural parasite-resistance polymorphism from a Caribbean population of the snail Biomphalaria glabrata . In independent experimental evolution lines , RAD genotyping shows that the same genomic region responds to selection for resistance to the parasite Schistosoma mansoni . A dominant allele in this region conveys an 8-fold decrease in the odds of infection . Fine-mapping and RNA-Seq characterization reveal a <1Mb region , the Guadeloupe Resistance Complex ( GRC ) , with 15 coding genes . Seven genes are single-pass transmembrane proteins with putative immunological roles , most of which show strikingly high nonsynonymous divergence ( 5-10% ) among alleles . High linkage disequilibrium among three intermediate-frequency ( >25% ) haplotypes across the GRC , a significantly non-neutral pattern , suggests that balancing selection maintains diversity at the GRC . Thus , the GRC resembles immune gene complexes seen in other taxa and is likely involved in parasite recognition . The GRC is a potential target for controlling transmission of schistosomiasis , including via genetic manipulation of snails .
Schistosomiasis is by far the most important helminth parasitic disease of humans . Schistosomes infect over 200 million people worldwide [1 , 2] , causing a chronic , debilitating disease that can lead to lifelong disability [3 , 4] . The disability-adjusted-life years lost to this disease are estimated at 13–56 million , a value rivaling that of malaria [3] . There are no effective vaccines against schistosomes , and effective treatment still relies on regular dosing with a single drug , praziquantel [5] . Praziquantel resistance in schistosomes can be easily selected for in the lab , suggesting that natural populations which infect humans could also evolve drug resistance [6] . In fact , there is now credible evidence of reduced drug susceptibility in some heavily-treated human populations [7] . This problem may increase substantially as mass treatment with praziquantel escalates under initiatives such as the Schistosomiasis Control Initiative [8] and the Gates Foundation’s SCORE project ( http://score . uga . edu ) . Alternate control strategies are therefore needed , including tactics for blocking transmission via the aquatic snails that serve as intermediate hosts . Understanding the molecular mechanisms by which snails and schistosomes interact will be the key for new approaches to interrupt transmission [9] . Most molecular research to date has focused on the schistosome parasite Schistosoma mansoni and its New World snail host Biomphalaria glabrata , recently aided by the newly sequenced genomes of both species ( [10]; https://www . vectorbase . org/organisms/biomphalaria-glabrata ) and transcriptomic studies [11 , 12] . Several lines of evidence demonstrate that resistance to infection is highly heritable in snails [13–17] . First , infection rates in inbred snail lines are consistent and typically either 0% or close to 100% [13 , 14] . Second , artificial selection experiments can produce snails with significantly increased or decreased susceptibility in a few generations [15 , 17] . Third , resistance to infection can be mapped to genetic markers in linkage crosses [16] . However , despite recent advances [18–20] , our knowledge of B . glabrata-S . mansoni interactions lags behind that of other host-parasite systems such as mosquito-Plasmodium [21 , 22] . Expression levels of some genes are known to influence resistance of B . glabrata to S . mansoni ( FREPs , [23 , 24]; Hsp 90 , [25] ) . However , to date there is only one genic locus known at which allelic variation associates with resistance ( sod1 , [26 , 27] ) , and the causality of this association still needs to be proven . Furthermore , there is substantial B . glabrata-strain by S . mansoni-strain ( G × G ) interaction in compatibility ( i . e . one strain of B . glabrata can be highly resistant to one strain of S . mansoni , but highly susceptible to another , and vice versa ) [28–32] . This pattern may reflect different per-strain combinations of highly diverse coevolving loci [33] which could conform to one of several models of host-parasite genotype matching [34] . In lab populations of B . glabrata and S . mansoni from natural populations in Guadeloupe , West Indies , only 40–50% of snails can be infected , no matter how many parasite miracidia are used to challenge them [33] . This scenario is consistent with a simple model of host-parasite genotype matching , in which the parasite population lacks the ability to be compatible with certain alleles in the host population [33] . Here we identify and characterize this resistance polymorphism in the snail population at the molecular level .
We started with a laboratory population of snails approximately ten generations removed from the wild , which had originated from hundreds of wild Guadeloupe snails and had been maintained as a randomly-mating population in the hundreds at the Université de Perpignan ( “Guadeloupe laboratory population” ) [33] . We selected two independent lines for resistance by challenging individual snails with either 10 ( line R10 ) or 30 ( line R30 ) miracidia and allowing only uninfected snails to found the next generation ( S1 Table ) . Susceptibility ( % of snails infected ) dropped from approximately 50% ( 53% +/- 4 ) to under 10% ( 3% +/- 2 for R10; 6% +/- 3 for R30 ) over five generations in the selection lines , but remained at approximately 50% ( 52% +/- 7 ) in the unselected control line GUA ( Fig . 1A ) . Linear regression of generation time on susceptibility ( logit transformed ) showed no significant effect for GUA ( p > 0 . 1 ) , but a significantly negative slope for both R10 ( odds of infection each generation decrease by 45% ( 95% CI = 36–54% ) ; p < 0 . 01 ) and R30 ( odds of infection each generation decrease by 42% ( 95% CI = 28–53% ) ; p < 0 . 01 ) . A combined regression with both selection lines showed no significant effect of population ( R10 vs . R30 ) on susceptibility ( p > 0 . 1 ) . In samples of 28 individuals each from GUA , R10 and R30 , we observed 6573 informative RAD markers that aligned to the reference genome , including 4142 with codominant ( two observable alleles ) variants ( SNPs or small indels ) and 2431 null markers . The same null marker , aligning to Scaffold1 , Site 1 . 814Mb , showed the highest allele frequency difference for both the R30-GUA comparison ( FST = 0 . 52 ) and the R10-GUA comparison ( FST = 0 . 30 ) , with the allele changing frequency in the same direction for both selection lines ( Fig . 2 ) . In 100 bootstrap replicates in which 28 samples of each population were chosen randomly with replacement , this marker was the top FST outlier for the R30-GUA comparison 84% of the time ( 95% CI of FST = 0 . 38–0 . 67 ) , and for the R10-GUA comparison 8% of the time ( 95% CI of FST = 0 . 15–0 . 50 ) . We selected this top outlier for further study . We next looked at the additional markers on Scaffold 1 and found two other null markers <250kb away showing relatively high FST ( Fig . 2 ) . We supplemented our analysis with Stacks [35] , a program that aligns reads to each other rather than to a reference genome , in order to find divergent markers on reads with low similarity to the reference genome . Because the genomic sequence flanking Stacks markers is unknown , they can be difficult to confirm with PCR , so we did not employ Stacks as a stand-alone analysis . However , Stacks revealed an additional RAD marker showing high allele frequency difference ( R30-GUA: FST = 0 . 46; R10-GUA: FST = 0 . 14 ) , and high linkage disequilibrium ( LD ) to Scaffold1 , Site 1 . 814Mb ( r = 0 . 94 , p < 10-15 ) , which had been too divergent to align unambiguously to the reference genome but which showed sequence similarity to Scaffold4 , Site 1 . 466Mb ( Fig . 2 ) . We determined the susceptibility/resistance phenotype of 289 additional snails from the Guadeloupe laboratory population by challenging each with 20 miracidia: 159 were resistant ( uninfected ) and 130 were susceptible ( infected ) . We genotyped them at one of the high-FST RAD markers on Scaffold1 ( “grc1” ) using PCR and Sanger sequencing . All Sanger genotypes could be unambiguously phased by eye and were found to comprise three alleles ( multi-SNP haplotypes ) , all of intermediate frequency . Allele R ( “resistance” ) with a frequency of 32 . 0% ( +/- 3 . 4% ) , showed a strong negative correlation with infection ( Fig . 1B ) and was dominant to the two S ( “susceptibility” ) alleles , S1 ( 42 . 6% +/- 3 . 2% frequency ) and S2 ( 25 . 4% +/- 3 . 6% frequency ) . Specifically , snails without an R allele ( n = 139 ) had a 70 ( ± 4 ) % chance of infection , while snails with at least one R allele ( n = 150 ) had a 22 ( ± 3 ) % chance of infection ( r2 = 0 . 23; odds ratio = 8 . 2; Fisher’s exact test , p < 10-15 ) . There was no significant difference in infection between RR homozygotes and RS heterozygotes ( Fisher’s exact test , p > 0 . 1 ) . Similarly , Sanger sequencing and genotype phasing by eye revealed three alleles ( multi-SNP haplotypes ) in a PCR-amplified region of Scaffold4 ( “grc2” ) , one of which was in near-perfect LD with allele R: only one individual out of 289 showed recombination between Scaffolds 1 and 4 ( r = 0 . 996 , p < 10-15 ) . The remaining two grc2 alleles were in modest LD with S1 and S2 ( r = 0 . 572 among the 139 SS snails , p < 10-12 ) . We genotyped a subset of 94 snails at additional sites on Scaffold1 and Scaffold4 ( Table 1; Fig . 3 ) , and found alleles in perfect LD with allele R extending from sites 1 . 541–1 . 940Mb on Scaffold1 and sites 1 . 187–1 . 468Mb on Scaffold4 , with the exception of a single recombinant at Scaffold4_1 . 187Mb . LD declined rapidly further upstream on both scaffolds , beginning at site 1 . 501Mb on Scaffold1 and at site 1 . 081Mb on Scaffold4 . The different FST values among the RAD markers on these scaffolds ( Fig . 2 ) is consistent with the imprecision of genotyping with null RAD tags ( e . g . null heterozygotes cannot be identified ) , rather than different allele frequencies in the sample . Thus , the ends of these two scaffolds are tightly linked and probably either adjacent or nearly so , forming a region of > 0 . 7Mb in near-perfect LD and showing a strong association with resistance . We refer to this region as the Guadeloupe Resistance Complex ( GRC ) . We performed RNA-Seq on outbred , parasite-unchallenged individuals of each genotype from an Oregon State University population derived from the original Guadeloupe laboratory population . The purpose of the RNA-Seq analysis was to facilitate gene annotation of the GRC , to obtain coding sequence for each allele , and to identify genes that might be differentially expressed between haplotypes . We sequenced 18 RR homozygotes , 9 S1S1 homozygotes , and 9 S2S2 homozygotes , all showing perfect LD among all three alleles at grc1 and grc2 . In order to exhaustively characterize differences in either sequence or expression among haplotypes , we identified 31bp sequences ( 31-mers; the largest size possible in our analysis pipeline ) showing significantly higher counts between RR and SS ( 8639 31-mers ) , between S1S1 and RR ( 5586 31-mers ) , and between S2S2 and RR ( 14976 31-mers ) . The large excess of 31-mers in S2S2 vs . RR , relative to the other comparisons , stems from the fact that LD between these two haplotypes extends ~0 . 7Mb farther upstream on Scaffold4 . We ignored these 31-mers that align upstream of Scaffold4_1 . 1Mb because recombination between R and S1 indicates that they do not contain the causal variant . We assembled transcripts from 27 genes ( defined as a transcribed sequence >500bp ) , including 15 coding and 12 putative noncoding ( no open reading frame >500bp; designated grcnc for “Guadeloupe resistance complex noncoding” ) genes that encompassed most of these 31-mers ( 95% for R , 90% for S1 , and 31% for S2; Table 2 ) . Nearly all transcripts aligned to the GRC region on Scaffold1 and Scaffold4 , with two exceptions: one transcript ( plmt ) aligned to Scaffold7477 , and another ( grctm2 ) aligned partially to Scaffold1 and partially to Scaffold7477 . However , SNPs in these transcripts were in perfect LD with the GRC , indicating that this section of Scaffold7477 ( at least 6kb in size ) occurs within the GRC . With one exception ( described below ) , all genes appeared in all three genotypes , although in some cases the orthologous transcript was under 500bp due to apparent truncation , and in some cases multiple isoforms of the same transcript were observed for some genotypes but not others ( S1 Fig ) . The coding genes represented a diversity of protein families , including hyaluronidase , myosin light chain kinase , hemolysin , paraspeckle component , zinc finger , ADP ribosylation factor GTPase-activating protein , and palmitoyltransferase ( Table 2 ) . Other coding genes showed low or no similarity to any characterized sequence . Seven genes , all on Scaffold1 between positions 1 . 6–2 . 0Mb ( middle of the GRC ) , show similar secondary structure , all encoding single-pass transmembrane proteins ( TM1s ) with >50 residues on either side of the transmembrane domain , and with a large ( >50% of the protein ) N-terminal ( extracellular ) domain in which β-strand residues are prevalent ( >25% of residues and >1 . 5 times as common as α-helix residues ) . TM1 proteins are a structural class seen in a wide variety of protein families with roles in processes like cell migration , adhesion , and growth , and typically acting as receptors for extracellular signals [36] . Notably , TM1s play an important role in the immune system , and include B- and T-cell receptors , Fc receptors ( binding the fragment crystallizable region of antibodies ) , some major histocompatibility complex ( MHC ) receptors , and Toll-like receptors , all of which are membrane-bound receptors recognizing foreign molecules [36] . Four of the TM1 genes in the GRC ( grctm2 , grctm3 , grctm4 , and grctm5 ) encode fibronectin III domains ( DELTA-BLAST , E < 10-3; Fig . 4 ) . Although the TM1 genes in the GRC do not all share primary sequence similarity , we designated them grctm1-grctm7 for “Guadeloupe resistance complex transmembrane” due to their similar secondary structure . In the parasite-unchallenged snails , although expression varied substantially among genes ( per-sample , per-site depth from 1x to >50x ) , expression differences among the three genotypes were low ( S1 Fig , S2 Fig ) . We performed 90 t-tests for expression depth among genotypes in the unchallenged snails , and thus employed a Bonferroni-corrected α value of 0 . 0006 . There was only a single instance of a gene not observed in all genotypes , and it was a short ( 605bp ) , noncoding sequence , grcnc8 , observed only in RR snails at relatively low coverage ( depth = 1 . 1x ) ( t-test between RR and SS , p<10-15 ) . Only one other gene , the non-coding grcnc9 , showed a >2-fold difference between RR ( 4 . 9x ) and both SS ( 0 . 8x , 0 . 9x ) genotypes ( t-test , p<10-9 ) . To be conservative with respect to identifying candidate genes , we also noted genes with expression differences that would be significant if uncorrected for multiple tests . There were two such genes showing small ( <2 fold ) differences between RR and both SS genotypes in the same direction: the non-coding grcnc5 ( RR = 9 . 7x , S1S1 = 6 . 6x , S2S2 = 5 . 0x; p<10-2 ) , and the TM1 gene grctm6 ( RR = 1 . 2x , S1S1 = 2 . 2x , S2S2 = 2 . 4x; p<10-2 ) . A few other genes showed >2-fold expression differences in some genotype comparisons , but not in both RR-S1S1 and RR-S2S2 comparisons . In order to test whether any expression differences are only apparent during parasite challenge , we also performed RNA-Seq on pooled DNA from six families ( 3 RR , 1 S1S1 , 2 S2S2 ) at 2 and 6 hour intervals after exposure to miracidia . We performed 62 t-tests for expression depth among genotypes in these challenged snails , and thus employed a Bonferroni-corrected α value of 0 . 0008 . Expression differences among haplotypes appeared slightly greater in challenged snails than in unchallenged snails , but still low overall and not significant . As with unchallenged snails , we conservatively noted all genes with expression differences that would be significant if uncorrected for multiple tests . There were 4 genes with greater expression in SS families: grcnc2 at both intervals ( 2 hr RR = 2 . 1x , 2 hr SS = 3 . 6x , 6 hr RR = 1 . 6x , 6 hr SS = 4 . 3x; p < 0 . 05 for both ) , grctm6 at 6 hr ( RR = 1 . 6x , SS = 4 . 5x; p < 0 . 05 ) ; adprf at 6 hr ( RR = 13 . 6x , SS = 17 . 2x; p < 0 . 05 ) and hely at both intervals ( 2 hr RR = 1 . 9x , 2 hr SS = 4 . 2x , 6 hr RR = 1 . 6x , 6 hr SS = 3 . 0x; p < 0 . 05 for both ) . Only grcnc8 , which was absent in both challenged and unchallenged SS snails , was higher in challenged RR families ( 2 hr RR = 0 . 8x , 6 hr RR = 0 . 9x ) . Additionally , there were 5 genes with >2-fold but nonsignificant ( even if uncorrected ) challenged expression differences among RR and SS snails: hyal at both intervals ( 2 hr RR = 2 . 8x , 2 hr SS = 9 . 9x , 6 hr RR = 3 . 3x , 6 hr SS = 8 . 5x ) , grcnc4 at 2 hr ( RR = 7 . 9x , SS = 3 . 7x ) , grcnc6 at 2 hr ( RR = 12 . 3x; SS = 3 . 5x ) , grcnc9 at both intervals ( 2 hr RR = 1 . 2x , 2 hr SS = 0 . 4x , 6 hr RR = 1 . 0x , 6 hr SS = 0 . 4x ) , and grctm6 at 2 hr ( RR = 1 . 8x , SS = 3 . 8x ) . Notably , there were no coding genes with increased expression in RR families . In contrast to expression differences , sequence divergence among alleles was often very high ( Fig . 3; Fig . 4 ) . Silent ( synonymous and noncoding ) divergence ( dS ) was over 5% in at least one pairwise comparison for 12 genes , and over 10% for 4 genes . All coding genes harbored nonsynonymous variants , but for non-TM1 coding genes , pairwise nonsynonymous divergence ( dN ) was 1% or less . In contrast , six of the TM1 genes had mean dN >5% , and all of these had at least one 75bp window with dN >15% in at least one comparison . The alleles at the remaining TM1 gene , grctm1 , showed dN <1% but still encoded very different proteins because the S2 allele had a premature stop codon approximately 30% of the way through its open reading frame . Because dS was also very high among TM1 alleles , dN did not significantly exceed dS . Only two genes showed high nonsynonymous substitution ( >3% ) specific to the R allele: grctm5 ( both R isoforms ) and grctm6 . Thus , the TM1 genes not only show structural similarity to known immune-relevant receptor genes , they also show the highest dN among alleles of all the genes in the GRC . In this snail population the GRC harbors three highly divergent haplotypes that have very even allele frequencies ( >25% ) . In order to test for the signature of balancing selection , we compared Tajima's D and nucleotide diversity ( π ) for the GRC to that in the rest of the genome . Tajima’s D at Sanger-sequenced marker locus grc1 ( 15 polymorphisms , 289 samples ) is 3 . 99 . Tajima’s D in the GRC cannot be directly estimated from the RNA-Seq data because samples were chosen non-randomly based on GRC genotype . However , if we adjust RNA-Seq SNP frequencies based on the population allele frequencies of the three haplotypes , Tajima’s D in the GRC is estimated as 3 . 85 . In order to compare this estimate to genome-wide patterns , we calculated Tajima’s D in two independent datasets: the RNA-Seq data from the 36 unchallenged snails , and the RAD data from the 28 GUA snails . We identified 779 genomic scaffolds for which we could confidently align at least 15 RNA-Seq SNPs with no missing genotypes ( 15–157 SNPs per scaffold; excludes SNPs in the GRC ) . Genome-wide , mean Tajima’s D is significantly less than in the GRC ( mean = 1 . 68; 95% CI = -0 . 98 to 3 . 62; empirical p for D ≥ 3 . 85 is 0 . 01 ) . In the RAD data , only 709 variants had no missing data across all GUA samples , and only a single scaffold contained at least 15 such variants , so we could not estimate a distribution of Tajima’s D as we did with the RNA-Seq data . However , overall Tajima’s D for these RAD variants is 1 . 76 , a very similar estimate to the RNA-Seq data , and again much lower than at the GRC . Both genome-wide estimates are still significantly higher than the neutral expectation of D = 0 ( coalescent simulation , p < 0 . 05 ) . Similarly , silent and noncoding diversity ( πS ) in the GRC ranged among genes from 0 . 16% to 7 . 40% , with a majority of genes ( 15 out of 26 universally expressed genes ) showing πS > 2% . Nonsynonymous diversity ( πN ) in the GRC ranged among genes from 0 . 07% to 5 . 69% , with six TM1 genes showing πN > 3% . In order to compare π in the GRC to genome-wide patterns , we identified 3061 genomic scaffolds for which we could confidently align at least 100 contiguous bp of RNA-Seq reads at high coverage ( length = 100–11 , 694bp; excludes reads in the GRC ) . Genome-wide π is significantly lower than for most GRC genes ( mean = 0 . 49%; 95% CI = 0 . 06–1 . 75%; empirical p for π ≥ 2% is 0 . 01; empirical p for π ≥ 3% is 0 . 001 ) . Scaffold1 includes the most interesting candidate genes as well as some obvious assembly errors ( e . g . Scaffold7477 must be inserted somewhere ) . Therefore , in order to verify the assembly of the Scaffold1 portion of the GRC , we sequenced eight BAC clones that putatively tiled across the GRC region , and aligned them to the B . glabrata reference genome ( BAC clones and the genome assembly were created using a Brazilian strain , BB02; [37] ) . All aligned to Scaffold1 as expected and largely confirmed the reference genome assembly between sites 1 . 471–1 . 971Mb , with a few exceptions . One large exception is the insertion ( relative to the genome assembly ) of a 69kb region aligning to Scaffold4587 , Scaffold7002 , and Scaffold7477 . This complete insertion is found in two BACs , which also overlap on Scaffold1 between sites 1 . 74–1 . 80 Mb . This result is consistent with our RNA-Seq transcripts , one of which ( plmt ) is on Scaffold7477 but in perfect LD with the GRC , and one of which ( grctm2 ) is partially on Scaffold7477 and partially on Scaffold1 , sites 1 . 79–1 . 81 Mb . This insertion is accompanied by a 14kb deletion between sites 1 . 762–1 . 776 Mb on Scaffold1 . Notably , this deletion contains a putative gene encoding an acetylglucosaminyltransferase . Thus this gene is not part of the GRC , consistent with RNA-Seq results . A second insertion of indefinite size ( at least 67kb ) occurs after Scaffold1 site 1 . 971Mb , as observed in two BACs . We did not observe sequence corresponding to the end of Scaffold1 from site 1 . 971Mb to 2 . 184Mb , which may occur after this final insertion or may have been incorrectly assembled onto Scaffold1 . A small but interesting insertion matching Scaffold866 , sites 57 . 7–58 . 1kb , is an exon of gene grctm6 , corresponding exactly to the region of the gene where allele S1 shows no sequence similarity to the other alleles ( Fig . 4 ) .
The GRC contains 15 coding genes , some of which are duplicated within some haplotypes . Pinpointing the specific causal gene ( s ) responsible for resistance in this system will require additional experimental evidence , but the most promising candidates are the TM1 genes , especially those with many nonsynonymous variants unique to the resistant haplotype . Gene regulation in cis is not a strong candidate for the functional mechanism because there were few significant and/or large ( >2-fold ) expression differences , especially in coding genes ( S1 Fig; S2 Fig ) . Although we cannot rule out a role for subtle expression differences , including in noncoding transcripts , the simplest mechanism of dominant resistance via gene expression would be overexpression of the resistance allele in a coding gene , and such a pattern is never observed . In contrast , the strikingly high amino acid sequence divergence is a much stronger candidate for the functional mechanism . All genes show some nonsynonymous differences , but only the TM1 genes show high dN ( Fig . 3 ) . The TM1 alleles also harbor many length polymorphisms , and some alleles with short open reading frames may be nonfunctional ( e . g . grctm1_S2 , grctm5_S1 , grctm7_S1 , Fig . 4 ) . We also note that the cluster of TM1 genes is in the center of the region of statistical association with resistance , the most likely position in which to find the causal locus or loci . That these TM1 genes also code for proteins that share structural similarity to known pathogen-recognition molecules [36] is particularly intriguing . Typical TM1s usually have their transmembrane domain very close to the N- or C-terminal ( <50 residues ) [38] , but in all seven GRC TM1 genes , the transmembrane domain is >50 residues from either end , a feature they share with Toll-like receptors and a minority of other TM1s . Most also contain fibronectin III domains which are often involved in molecular recognition [39] . Thus , although not all the TM1 genes show sequence similarity to each other , they may be functionally or even evolutionarily related . Specifically , we hypothesize that their extracellular ( N-terminal ) domains recognize foreign substances such as parasite PAMPs ( pathogen-associated molecular patterns ) , while their intracellular ( C-terminal ) domains transmit this signal to other cellular components , leading to a physiological response . For example , S . mansoni produces polymorphic mucins , SmPoMucs [40] , that interact with snail FREP immunity proteins [20 , 41 , 42] . It is possible that the TM1 proteins recognize SmPoMucs , SmPoMucs-FREP complexes , or unrelated PAMPs . The most compelling candidates for controlling resistance to S . mansoni in the TM1 gene cluster are grctm5 and grctm6 . Only these two genes show high nonsynonymous substitution specific to the R haplotype ( Fig . 4 ) . Gene grctm5 encodes a fibronectin III protein with sequence similarity to chitinase . Intriguingly , it is present in two copies in RR snails but only one copy in the susceptible genotypes . Furthermore , these two R isoforms are substantially divergent from each other ( dN = 7 . 9% , dS = 7 . 4% ) , as well as from S1 ( dN = 4 . 2–6 . 8% , dS = 7 . 1–7 . 8% ) and S2 ( dN = 7 . 4–7 . 5% , dS = 6 . 7–9 . 1% ) alleles . Both grctm5 R sequences show large , significant expression differences between RR ( depth = 2 . 8x , 3 . 0x ) and S1S1 ( depth = 1 . 0x; t-tests , p < 10-3 for both ) , but not between RR and S2S2 ( 2 . 3x; t-tests , p >0 . 1 for both ) ( S1 Fig ) . Gene grctm6 has no strong sequence similarity to any known proteins . It shows high divergence among all three haplotypes , and includes a ~300bp segment where the S1 haplotype shows no sequence similarity to R or S2 . The grctm6 gene also shows significantly lower expression in RR than in S1S1 or S2S2 . The GRC , especially the TM1 genes , shows three patterns of genetic variation consistent with balancing selection . First , with three intermediate frequency alleles , the site-frequency spectrum is highly skewed away from neutral expectation , as evidenced by the extremely high Tajima's D value . Second , high LD across several Mb is an unusual pattern suggestive of selection [43] . Although high LD can also be caused by low recombination rates , this explanation seems unlikely . In most genomes , regions of low recombination are gene-poor [44] , unlike the GRC . The main exceptions are chromosomal inversions which can suppress recombination across suites of genes; however , an inversion cannot maintain high divergence among more than two haplotypes , so the three divergent haplotypes of the GRC require another explanation . Both Tajima’s D and LD can be elevated by demographic bottlenecks . Genome-wide values of Tajima’s D in the Guadeloupe laboratory population are higher than the neutral equilibrium expectation , suggesting such a demographic effect . However , Tajima’s D at the GRC is significantly higher still , suggesting that selection has also played a role . Because the Guadeloupe laboratory population has had a census size in the hundreds at collection and ever since , we expect genetic drift in these snails to have been minimal . Thus , both allele frequencies and the extent of LD at the GRC should be similar to values in the wild , which we cannot directly estimate . Guadeloupe B . glabrata undergoes population fluctuations in the wild [45] , and these natural demographic dynamics likely contribute to the genome-wide population genetic patterns . In addition , our genome-wide estimates of Tajima’s D are based on samples that did experience minor laboratory bottlenecks ( N = 100 over five generations for GUA RAD; N = 30 for the founding of the Oregon State University population from which the RNA-Seq data is derived ) . Therefore , genome-wide Tajima’s D and LD may be lower in the wild , and the GRC may be an even more extreme outlier . The third non-neutral pattern , high genetic diversity , cannot be explained by demographics . The GRC is an outlier with respect to nucleotide diversity ( π ) at both synonymous and nonsynonymous sites , especially in six of the TM1 genes . The high nonsynonymous diversity ( πN = 3 . 2–5 . 7% ) observed at six of the TM1 genes is quite remarkable , as it greatly exceeds the presumably neutral values seen throughout most of the genome ( p < 0 . 001 ) . Of course , reads at highly diverse loci might be less likely to align to the reference genome , so our genome-wide estimate of π is likely biased downward . Still , such high nonsynonymous diversity is rarely seen in any species; for example , 99 . 9% of Drosophila simulans genes show lower πN values [46] . We did not observe a dN/dS ratio significantly greater than one at any gene . At grctm6 , dN is 1 . 65 times as high as dS ( synonymous plus noncoding ) divergence ( Fig . 3 ) , but only 0 . 83 times the magnitude of synonymous divergence alone ( excluding noncoding sites ) . Such high diversity at both synonymous and nonsynonymous sites is evidence that alleles have coexisted for an unusually long time at intermediate frequency , presumably because selection has maintained them in balance [47 , 48] . As with other immunity genes that show similar patterns [49] , it may be the case that heterozygotes have a fitness advantage because they can recognize and respond to a greater diversity of parasites , or it may be that parasite community composition changes so quickly that no one allele is advantageous for long enough to fix . Because S . mansoni arrived in the New World only in historical times [50] , the selection pressure maintaining these alleles must be due to other native parasites , perhaps including other trematodes which have a very close and ancient relationship with snails [17 , 51] . As with Tajima’s D and LD , π in the Guadeloupe laboratory population may be slightly different than in the wild , although we expect the effect of genetic drift in the lab to have been minimal . If anything , we would expect even greater diversity in natural populations , as there may be additional unsampled divergent alleles , especially if populations outside of Guadeloupe are considered ( indeed , the reference genome provides an example ) . Additional work is needed to study the relevance of the GRC to phenotypic variation in resistance across the range of B . glabrata , or in other snail species . Théron et al . [33] showed that the shape of dose-response curves in the Guadeloupe snail-schistosome community is consistent with a simple host-parasite phenotype compatibility model , in which allelic variation in the snails controls matching with the parasite , likely via a complimentary locus ( or loci ) in the Guadeloupe population of S . mansoni . Such phenotype compatibility could occur by several mechanisms . Under the matching alleles ( MA ) model , parasites avoid detection by matching host self determinants , whereas under the inverse matching allele ( IMA ) model , host molecules recognize parasite molecules leading to an immune response [34] . Importantly , heterozygous hosts are susceptible under MA and resistant under IMA [34] . Thus , the genetic dominance of resistance at the GRC favors IMA . Similar signatures of balancing selection are seen across a wide range of host taxa at immunity loci that interact directly with infectious disease agents [52–54] . In the Anopheles/Plasmodium pair , the best studied system of invertebrate host and eukaryotic parasite , several genomic loci have been associated with host resistance and show patterns paralleling those seen at the GRC [55–57] . For example , the Anopheles APL1 locus also represents a cluster of structurally similar genes that show extraordinarily high nonsynonymous diversity of the same magnitude as the GRC TM1 genes ( 3–6% ) [55 , 57] . Likewise , nonsynonymous divergence is high among alleles of the Anopheles TEP1 locus , although gene conversion rather than balancing selection appears to be responsible [56] . Nonsynonymous diversity exceeding neutral expectations is also consistently observed in the MHC of vertebrates [49] and plant R genes [58] . As with the GRC , genetic dominance of resistance is common across these systems , consistent with the IMA model . Furthermore , the close physical linkage of the TM1 genes resembles the gene clusters that form many of these resistance genomic regions in other taxa . Such clustering may be a neutral artifact of tandem gene duplication , or it may have adaptive significance through shared gene regulation or the maintenance of beneficial multi-gene haplotypes [59] . The GRC is an obvious target for applications in the control of schistosomiasis . Genetic manipulation of disease vectors is a promising approach that is already underway for mosquitoes and other pest species [9 , 60 , 61] , and driving resistance genes into B . glabrata populations is under discussion [27] . Gene knockdowns in B . glabrata have shown repeated success [23 , 24 , 62 , 63] , and targeted inhibition of GRC genes or their products may pinpoint the specific causal gene . Alternatively , transfection with a multi-gene haplotype could have practical utility even without perfect understanding of each gene’s functional contribution . Although genetic modification of snails is still in its infancy , the CRISPR nuclease system shows promise for fine-scale modification of even non-model species [9] . Importantly , transgenic vectors would not need to have higher fitness than wild-type organisms if transgenes were spread by gene drives [9] . Thus , if only the R allele recognizes S . mansoni while other alleles recognize other parasites of no importance to human health , it would be possible to drive the R allele to fixation even without an adaptive benefit to the snails , if it were coupled to a gene drive conferring preferential non-Mendelian inheritance of the R allele [9] . Conversely , if multiple alleles at this locus are required to recognize all strains of S . mansoni , it might be possible to engineer a single haplotype with duplicated genes that included all relevant sequences . Alternatively , if the GRC genes initiate an immune signal cascade upon recognition of the parasite , future research could seek to manipulate snails such that this signal cascade is constitutively upregulated regardless of genotype or infection status . The matching loci in the parasite , once they are identified , may also be targets for drugs or genetic manipulation .
Mice and hamsters were used to maintain the schistosome parasites and to produce miracidia for challenge experiments . Infection is through contact with inoculated water and involves minimal discomfort . Infected rodents are euthanized with CO2 prior to showing clinical signs of disease and are dissected to recover parasitic worms and their eggs . This research was approved by OSU IACUC and the French veterinary agency . B . glabrata and S . mansoni were collected on the island of Guadeloupe , West Indies , in 2005 as described previously [33] . Selection lines were created at the University of Perpignan beginning in 2008 . We selected for resistance by challenging snails with 10 ( line R10 ) or 30 ( line R30 ) S . mansoni miracidia and allowing only uninfected snails in each independent line to contribute to the next generation ( N = 50–100 snails challenged per line per generation , and so the numbers used to found each subsequent generation varied each generation and increased from as low as 38 to 90 as the susceptibility of the two lines decreased; S1 Table ) . This was done for five generations . We measured susceptibility in each line as the percentage of snails infected ( Fig . 1A ) . As a control , we maintained a line of uninfected snails ( GUA ) at a similar population size for the same time period ( N = 100 snails per generation ) . For GUA , 50–70 snails from each offspring generation were challenged in order to measure the susceptibility of GUA each generation ( percentage of snails infected , S1 Table ) , but these challenged snails did not contribute to subsequent generations . We tested for a correlation between generation and susceptibility by logit transforming susceptibility and using linear regression . We estimated the standard errors of proportions using √ ( ( p* ( 1-p ) ) /N ) . We used SbfI in RAD genotyping [64] of 28 individuals each from the control population ( GUA ) and the two selection lines ( R10 and R30 ) ( Illumina data at NCBI SRA , Bioproject Accession PRJNA268191 ) . We aligned reads to the B . glabrata reference genome with BWA [65] . All genomic analyses in this study used version 4 . 3 of the reference genome , and positions were subsequently converted to nearly identical reference genome version BglaB1 ( https://www . vectorbase . org/organisms/biomphalaria-glabrata ) . In order to maximize the density of markers , we considered both markers with observable variants as well as null markers for which a RAD tag aligned in only a subset of individuals . To estimate allele frequencies of SNPs , we counted two alleles per individual with 10x or greater depth , and two half-alleles per individual with 2–9x depth to account for possible nonobserved alleles , while genotypes with <2x depth were counted as missing . We only analyzed sites with a per-population allele count of at least 24 for both GUA and at least one other population , and which did not violate Hardy-Weinberg equilibrium ( χ2 > 20 ) in any population nor show overall heterozygote excess among populations . Allele frequencies for null markers were estimated using the method of Zhivotovsky [66] . For each informative marker we estimated the difference in allele frequency between GUA and each selected population ( FST ) as 1 – ( mean within-population expected heterozygosity ) / ( total expected heterozygosity ) . To assess RAD tags with low sequence similarity to the reference genome , we aligned RAD tags to each other with Stacks [35] and tested for any previously undetected high-FST markers . We assessed the robustness of FST results with 100 bootstrap replicates in which 28 samples from each population were chosen randomly with replacement . We chose the single most extreme outlier ( highest FST ) , along with markers showing evidence of close physical linkage to this marker ( <1Mb away on same scaffold and/or high LD ) , for further analysis . For these promising markers , we designed primer pairs to amplify and Sanger sequence the RAD tag in order to obtain the complete codominant genotype for all individuals . We phenotyped an independent sample of 289 snails by challenging them with 20 miracidia each and then designating them as either infected or uninfected following Theron et al . [33] . We examined markers showing exceptionally high FST by Sanger sequencing them in these 289 snails and testing for an association with phenotype . We calculated allele frequencies and standard errors of proportions . We calculated the proportion of phenotypic variation explained by the R allele as the square of the correlation coefficient ( r2 ) from a logistic regression of phenotype ( infected or not ) versus genotype ( R/- or not ) . We then developed primers to amplify nearby genomic regions in order to find and characterize the genomic region showing the strongest association with resistance . We tested these markers on a subset of 94 phenotyped snails . We extracted RNA from whole bodies of randomly-chosen , size-matched juvenile snails that had not been parasite-challenged . We prepared samples using the TruSeqTM RNA v2 kit ( Illumina RS-122–2001 ) following the low-throughput protocol found in the Sample Preparation v2 Guide . We performed RNA-Seq on the following 36 samples: 12 RR homozygotes ( single-end ) , 6 RR homozygotes ( paired-end ) , 6 S1S1 homozygotes ( single-end ) , 3 S1S1 homozygotes ( paired-end ) , 6 S2S2 homozygotes ( single-end ) , and 3 S2S2 homozygotes ( paired-end ) ( we first Sanger genotyped a large number of randomly-chosen snails from tentacle snips , and then we chose the above 36 to use for RNA-Seq ) . Sequencing was conducted on the Illumina HiSeq 2000 at Oregon State University ( Illumina data at NCBI SRA , Bioproject Accession PRJNA264063 ) . We converted FASTQ files to FASTA and used Jellyfish 1 . 0 . 2 [67] to count 31-mers in each sample , which is the largest kmer size that Jellyfish can count , but which is long enough to be typically unique in the genome and therefore represent specific transcripts . We then used a custom perlscript ( https://github . com/jacobtennessen/HOLDRS ) to identify 31-mers that differed significantly among genotypes , defined as showing at least a 2-fold difference in total count and a Welch’s t-statistic of at least 5 . These parameters were chosen to encompass both sequence and expression differences: expression differences under 2-fold may not be biologically meaningful , true sequence differences should result in a much larger ( all or nothing ) count ratio , and this t-statistic typically corresponds to a p-value under 10-4 , which will minimize the number of false positives among 31-mers from thousands of transcripts , while still detecting 31-mers that are truly overabundant in one sample ( N = 9 or 18 ) versus a control ( N = 18 ) . We identified reads containing these divergent 31-mers and used ABySS v . 1 . 3 . 4 [68] to assemble them into contigs . We used grep and manual alignment to extend these contigs and assemble additional contigs from divergent reads . We focused subsequent analysis on contigs >500bp , as well as shorter contigs with putative orthology to a contig >500bp in another genotype ( sequences in NCBI Transcriptome Shotgun Assembly Database , Bioproject Accession PRJNA264063 ) . In order to measure sequencing depth and identify polymorphisms , we aligned reads to these long contigs using BWA [65] and converted genotypes to vcf format . We chose six snail lineages homozygous at the GRC ( 3 RR , 1 S1S1 , 2 S2S2 ) . For each family we exposed 6 randomly-chosen , size-matched juvenile individuals to 20 miracidia each , and then we extracted RNA from whole bodies at 2 hr ( n = 3 ) and 6 hr ( n = 3 ) post infection . RNA from each set of 3 snails from the same family and time point was then pooled . We prepared and sequenced samples single-end as described above ( Illumina data at NCBI SRA , Bioproject Accession PRJNA264063 ) . We aligned reads to the assembled transcripts from the unchallenged dataset and compared depths as described above . We BLASTed the GRC portion of Scaffold1 ( B . glabrata reference genome ) against end sequences from the BAC library generated for B . glabrata strain BB02 ( ftp://ftp . ncbi . nih . gov/pub/TraceDB/biomphalaria_glabrata/; [37] ) . We chose BACs from among the hits , ordered them from the Arizona Genomics Institute , extracted DNA , and sequenced them in a Nano run of the Illumina MiSeq at Oregon State University ( Illumina data at NCBI SRA Bioproject Accession PRJNA268208 ) . We aligned the reads from eight BACs to the B . glabrata reference genome with BWA [65] . We calculated linkage disequilibrium between markers as the correlation coefficient , r , in number of R alleles per genotype ( 0 , 1 , or 2 ) We classified transcripts as “coding” if they contained an open reading frame of at least 500bp , otherwise they were designated “noncoding . ” We used DELTA-BLAST [69] to match candidate transcripts with known protein families . Genes were named following the guidelines in Bayne [70] . We characterized secondary structure with Jpred3 [71] and identified transmembrane domains using TMHMM v . 2 . 0 ( http://www . cbs . dtu . dk/services/TMHMM/ ) . We measured pairwise synonymous ( dS ) and nonsynonymous ( dN ) divergence , as well as allele-specific nonsynonymous substitution , with custom Perl scripts and DnaSP [72] . For each transcript , we used t-tests to identify significant differences in adjusted per-sample , per-site depth among genotypes in which all values were divided by the ratio of the total count of reads in that individual relative to the mean per-individual read count , and then log-transformed . Depth values of zero were arbitrarily recorded as 1/6 , which assumes they are present but too rare ( much less than 1x ) to have been sampled . For each transcript under each experimental scenario ( unchallenged , 2 hours post-challenge , and 6-hours post-challenge ) we compared RR vs . SS . For unchallenged snails we also compared RR vs . S1S1 and RR vs . S2S2 . If multiple sequences of the same gene were observed for one genotype , these were all tested separately . In order to estimate Tajima’s D [73] in the GRC , we calculated the frequency of all RNA-Seq SNPs on their haplotype ( R , S1 , or S2 ) , and then multiplied this frequency by the unbiased haplotype frequency estimate from the 289 snails Sanger sequenced at grc1 . In order to estimate genome-wide Tajima’s D and π , we aligned RNA-Seq reads to the B . glabrata reference genome with BWA [65] . We identified regions with at least 100 contiguous bp showing high ( ≥6x ) depth in all 36 samples . All of these high-depth regions were used to estimate π , while scaffolds with at least 15 SNPs in such high-depth regions were included in the estimate of Tajima’s D . Significance of Tajima’s D was estimated with standard neutral coalescent simulations in DnaSP [72] . | Schistosomes are water-borne blood-flukes that are transmitted by snail vectors . They infect over 200 million people in more than 70 countries and cause severe and chronic disability . Snails naturally vary in resistance to this parasite even within species , so bolstering snail resistance in the wild would block transmission . We artificially selected snails for resistance and observed a rapid evolutionary response , with the greatest change occurring in the same genomic region in two independent trials . We subsequently confirmed that the selected haplotype conveys resistance to infection by schistosomes . The extraordinarily high sequence divergence among haplotypes in this region appears to be elevated due to ongoing natural selection , likely via host-parasite co-evolution . We observed the highest variation in genes encoding putative parasite recognition proteins , suggesting that these control the resistance phenotype in a manner reminiscent of immune gene complexes in other taxa . Thus , this gene cluster presents a potential new target to interfere with parasite transmission at the vector stage . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | Hyperdiverse Gene Cluster in Snail Host Conveys Resistance to Human Schistosome Parasites |
Plasmacytoid and conventional dendritic cells ( pDCs and cDCs ) arise from monocyte and dendritic progenitors ( MDPs ) and common dendritic progenitors ( CDPs ) through gene expression changes that remain partially understood . Here we show that the Ikaros transcription factor is required for DC development at multiple stages . Ikaros cooperates with Notch pathway activation to maintain the homeostasis of MDPs and CDPs . Ikaros then antagonizes TGFβ function to promote pDC differentiation from CDPs . Strikingly , Ikaros-deficient CDPs and pDCs express a cDC-like transcriptional signature that is correlated with TGFβ activation , suggesting that Ikaros is an upstream negative regulator of the TGFβ pathway and a repressor of cDC-lineage genes in pDCs . Almost all of these phenotypes can be rescued by short-term in vitro treatment with γ-secretase inhibitors , which affects both TGFβ-dependent and -independent pathways , but is Notch-independent . We conclude that Ikaros is a crucial differentiation factor in early dendritic progenitors that is required for pDC identity .
Dendritic cells ( DCs ) are essential modulators of the immune response [1] . They can be broadly divided into conventional DCs ( cDC ) , which are required for antigen presentation , and plasmacytoid DCs ( pDC ) , which secrete high quantities of type-I interferon ( IFN-α , -β , -ω ) upon certain viral infections [2 , 3] . cDCs are further divided into cDC1 ( CD8+ ) and cDC2 ( CD11b+ ) subsets . Both DC lineages develop in the bone marrow . Monocyte and dendritic progenitors ( MDPs ) are the earliest known DC precursors , and they give rise to monocytes and common dendritic progenitors ( CDPs ) [4–6] . In turn , CDPs differentiate into pDCs and pre-cDCs , the latter of which migrate to the periphery to become cDCs . The molecular circuits regulating DC cell fate have been intensively studied , and some transcriptional regulators ( Ikaros , E2 . 2 , PU . 1 , IRF8 , GFI1 , NFIL3 , BATF3 , BCL11a ) and canonical signaling pathways ( TGFβ , Notch , Wnt ) have been identified [3 , 7–12] . However , the relationships and interactions between these players remain unclear , and this is important to understand if we wish to manipulate DC function . Deficiency of the Ikaros zinc finger DNA-binding protein and tumor suppressor , encoded by the Ikzf1 gene , is associated with profoundly impaired DC development . Mice homozygous for a dominant-negative ( dn ) Ikzf1 mutation lack all cDCs , while animals with a null mutation predominantly lack cDC2s [13] . In contrast , mice carrying the hypomorphic IkL/L mutation show a selective block in bone marrow ( BM ) pDC development , leading to an absence of peripheral pDCs , although cDCs appear normal [14] . These studies highlight the sensitivity of the DC lineages to Ikaros levels , where pDC development requires more Ikaros function than cDCs . In man , patients with germline IKZF1 mutations also exhibit reduced pDC , but not cDC numbers , indicating a conserved role for Ikaros in DC development [15] . Interestingly , IKZF1 deletions are associated with blastic plasmacytoid dendritic cell neoplasms ( BPDCN ) , a malignancy of pDC precursors with poor prognosis [16–18] . Thus Ikaros is required for DC development , but little is known about its molecular mechanisms . Here we show that Ikaros deficiency leads to multiple defects in pDC and cDC development . In particular , Ikaros is required for CDP development , where it antagonizes TGFβ function to promote pDC differentiation . We further show that Ikaros cooperates with Notch pathway activation to support the homeostasis of DC progenitors . Lastly , we show that a transient incubation of bone marrow cells with γ-secretase inhibitors rescues pDC development from WT and Ikaros-deficient BM progenitors , revealing a potentially novel way to enhance pDC function .
To determine how pDC differentiation is affected by Ikaros deficiency , we evaluated DC progenitor populations in IkL/L mice . IkL/L cells express functional Ikaros proteins at ~10% of WT levels , and although IkL/L mice die from Notch-dependent T cell leukemias at 4–6 months of age , the animals studied ( 6–8 weeks of age ) showed no signs of transformation ( normal CD4/CD8 profiles , T cell receptor chain usage , Notch pathway activation ) [19–21] . Successive stages of DC development were analyzed , which included BM Lin-Sca1-CD135+ cells , containing CD117hiCD115+ MDPs and CD117loCD115+ CDPs , as well as the more downstream BM CD11c+CD317+ pDCs and CD11c+CD135+MHCII-CD172a- pre-cDCs , and splenic cDCs ( Fig 1 ) [4–6 , 22] . In the BM , CDP numbers were significantly increased and pDC and pre-cDC numbers were significantly decreased in IkL/L mice , suggesting that IkL/L DC differentiation is blocked at the CDP stage ( Fig 1A–1C , 1E and 1F ) . In the spleen , IkL/L animals had no detectable pDCs , as previously reported [13 , 14] , fewer cDC2s , but similar numbers of cDC1s compared with WT ( Fig 1D and 1G ) . Thus Ikaros deficiency results in the specific accumulation of BM CDPs . We previously observed in a genome-wide study that genes associated with the Notch pathway ( eg . Hes1 , Ptcra , Uaca ) are upregulated in the BM pDCs of IkL/L mice [14] . Higher Hes1 and Ptcra mRNA levels were confirmed by RT-qPCR ( Fig 2A ) . To determine if Ikaros deficiency results in Notch activation throughout pDC development , we crossed IkL/L mice with animals carrying a Hes1-GFP knock-in ( KI ) reporter [23] . Total BM cells from Ik+/+ ( WT ) and IkL/L Hes1-GFP KI mice contained similar frequencies of GFP+ cells ( mostly CD19+ B cells ) ( Fig 2B and 2C ) . In contrast , GFP+ cells were nearly absent in WT BM pDCs , but they were present in a fraction of IkL/L pDCs ( 7–35% ) . IkL/L GFP+ pDCs were mostly SiglecH+CCR9lo , suggesting an immature phenotype ( Fig 2D ) [24 , 25] . CCR9lo pDCs from WT Hes1-GFP KI mice did not express GFP . These data indicated that the Hes1 locus , and perhaps the Notch pathway , are ectopically activated during pDC development in the mutant mice . To determine if ectopic Notch activation interferes with pDC differentiation in IkL/L mice , we first blocked Notch signaling in Flt3L-supplemented cultures of total BM cells , using a γ-secretase inhibitor ( GSI , Compound E ) [26 , 27] . As γ-secretase is required to cleave and activate ligand-bound Notch receptors , GSIs are potent inhibitors of Notch function . In the absence of GSI ( DMSO ) , WT cultures generated robust numbers of CD11c+CD137+CD11b- pDCs over an 8-day period , while IkL/L cultures did not ( Fig 3A ) . Strikingly , GSI treatment significantly enhanced WT pDC differentiation , and rescued pDC development in the IkL/L cultures to levels of WT cells . This occurred early , as GSI treatment at day 0 was both necessary and sufficient to rescue IkL/L pDC development ( Fig 3B and 3C ) . Similar results were obtained with other GSI compounds ( DAPT and MRK003 ) . In addition , early GSI treatment resulted in an increase in total cell numbers ( Fig 3D ) , which correlated with an expansion of immature CD11c- cells , particularly in the IkL/L cultures ( Fig 3B ) . The pDCs produced in the GSI-treated cultures were more immature , and expressed low levels of CCR9 and Ly49Q ( Fig 3E ) ; B220 levels , however , remained unchanged after GSI treatment . Importantly , the GSI-rescued WT and IkL/L pDCs expressed mRNA for Ifna following TLR9 stimulation in vitro with CpG ODN 1885 ( Fig 3F ) , suggesting functionality . Because GSI treatment at day 0 of culture was sufficient to induce differentiation , GSI was added only once at the onset of culture in subsequent experiments . To identify the DC progenitor cells sensitive to GSI , we co-cultured WT and IkL/L total BM cells , purified Lin-Sca1- cells , MDPs or CDPs ( all CD45 . 2+ ) , with CD45 . 1+ supporting WT BM cells , in the presence of GSI and Flt3L , for 8 days ( Fig 4A ) . The ability of the different CD45 . 2+ populations to give rise to pDCs was evaluated . GSI treatment consistently increased pDC differentiation from IkL/L CDPs ( Fig 4B ) . On the contrary , GSI did not affect WT MDPs ( 2 out of 3 experiments ) and CDPs , even though it enhanced pDC development from total WT BM cells . We also analyzed Lin-Sca1-CD117loCD135+CD115- cells ( "CD115- CDPs" ) in these assays , as they were reported to contain pDC-specific precursors [28] , even though they existed in similar numbers in WT and IkL/L BM ( S1A and S1B Fig ) ; GSI did not affect the pDC production from either WT or IkL/L CD115- CDPs ( S1C and S1D Fig ) , and these cells were not studied further . These results therefore suggested that Ikaros negatively regulates a γ-secretase-sensitive pathway mainly in ( CD115+ ) CDPs . To determine if transient GSI treatment rescues IkL/L pDC development in vivo , we adoptively transferred GSI-treated BM cells into recipient mice . WT and IkL/L BM cells ( CD45 . 2+ ) were cultured with Flt3L and GSI for 2 days , and then transplanted into irradiated hosts ( CD45 . 1+CD45 . 2+ ) along with CD45 . 1+ supporting WT BM cells . BM and spleen cells were analyzed 9 days later for CD45 . 2+ pDCs ( Fig 4C and 4D ) . In the BM , IkL/L cells generated few CD11c+CD137+CD11b- pDCs , regardless of GSI treatment ( S2 Fig ) . However , in the spleen , GSI-treated IkL/L cells generated CD11c+CD137+CD11b- pDCs while the DMSO-treated cells did not ( Fig 4C and 4D ) . WT cells generated slightly more pDCs after GSI treatment compared with DMSO . Importantly , the CD45 . 1+ supporting cells produced similar frequencies of pDCs in all conditions , indicating that GSI treatment enhanced IkL/L and WT pDC differentiation in a cell-intrinsic manner . Collectively , our results indicated that γ-secretase inhibitors rescue Ikaros-deficient pDC development in vitro and in vivo . Because γ-secretase inhibitors affect other pathways in addition to Notch , we tested the role of Notch activation in pDC development by genetic means . IkL/L mice were crossed with animals carrying a floxed null allele for Rbpj ( Rbpjf/f ) , the Notch transcriptional mediator , and the R26-CreERT2 transgene [29 , 30] . IkL/L Rbpj+/+ Cre+ ( IkL/L RBPJ WT ) and IkL/L Rbpjf/f Cre+ ( IkL/L RBPJ KO ) mice , along with control animals , were treated with tamoxifen for 5 days to delete Rbpj , and analyzed 5 days after the last injection . Deletion was confirmed by Western blot ( S3 Fig ) . BM cells from the tamoxifen-treated mice were cultured with Flt3L for 8 days , in the presence or absence of GSI , and cell expansion and pDC development were studied ( Fig 5A and 5B ) . In these experiments , we reasoned that if GSI rescues pDC development by inhibiting Notch signaling , then ( i ) Rbpj inactivation should mimic the effects of GSI , and ( ii ) GSI should not have additional effects when Rbpj is deleted . When cell numbers were evaluated , we observed that the samples treated with GSI contained significantly higher numbers of cells , regardless of RBPJ and/or Ikaros status ( Fig 5B ) . This suggested that GSI inhibits the function of pathways other than Notch . Likewise , when pDC development was evaluated ( Fig 5A and 5B ) , Rbpj deletion by itself did not enhance the differentiation of IkWT DMSO-treated cells , indicating that Notch activation is not required to limit pDC development when Ikaros is present . Further , in IkWT cells , GSI treatment enhanced pDC differentiation in both RBPJ WT and KO conditions , suggesting that GSI enhances pDC differentiation in the absence of Notch . Interestingly , when similar experiments were performed in IkL/L conditions , GSI treatment rescued pDC development in the RBPJ WT samples , as expected , but no rescue was observed when both RBPJ and Ikaros were mutated . GSI nevertheless still increased total cell numbers in the cultures from the RBPJ-Ikaros double mutant cells , indicating that its effects on pDC differentiation and cell expansion are separable . To determine why GSI treatment did not rescue pDC development in IkL/L RBPJ KO BM cultures , we analyzed the BM DC progenitor populations of tamoxifen-treated IkL/L RBPJ KO mice and littermate controls ( Fig 5C and 5D ) . Specifically , we evaluated the CDP population in the double mutant mice , as GSI rescues IkL/L CDP differentiation . These experiments revealed that MDPs and CDPs were barely detectable in most of the IkL/L RBPJ KO BM samples ( 4 out of 5 ) , while the BM from single mutant mice contained easily recognizable MDP and CDP cells . These results indicated that the GSI target population is absent in the IkL/L RBPJ KO BM , and suggested that Ikaros and Notch activation cooperate to generate or maintain MDP and CDP cells in the BM . Collectively , our results demonstrate that GSI treatment inhibits a Notch-independent pathway important for CDP development . To further investigate the molecular pathways targeted by Ikaros and γ-secretase in CDPs , we studied the transcriptome profiles of WT and IkL/L MDPs and CDPs , cultured in the presence or absence of GSI . We used a protocol similar to the one above , and co-cultured CD45 . 2+ WT or IkL/L MDPs , and CDPs , with supporting CD45 . 1+ WT BM cells . CD45 . 2+ cells were purified after 24h , and their transcriptomes were analyzed by microarray . In the vehicle-treated samples , 963 genes were differentially expressed >1 . 5-fold between IkL/L CDPs and all WT populations ( Fig 6A ) . Approximately 30% of these genes were deregulated in both IkL/L MDPs and CDPs ( clusters III and IV ) , and 70% were deregulated only in the IkL/L CDPs ( clusters I and II ) . To determine how these genes are expressed during WT DC development , we compared their levels of expression in IkL/L CDPs with those in WT progenitors and mature DC populations , as reported by the ImmGen Compendium ( GSE15907 ) , using unsupervised clustering [7] . Interestingly , this revealed that , among the genes up-regulated in IkL/L CDPs ( Fig 6B ) , the large majority ( >70% ) were related to mature cDC genes , and not pDCs . The remainder of the genes were DC progenitor-related . In contrast , among the genes down-regulated in IkL/L CDPs ( S4A Fig ) , most were related to DC progenitor ( CMP , MDP , CDP ) genes . Further , gene set enrichment analyses ( GSEA ) indicated that both the up- and down-regulated genes in the IkL/L CDPs correlated with those normally expressed in WT cDCs ( S4B Fig ) . Thus , Ikaros is required to repress the premature expression of cDC-associated genes in CDPs . We then asked if the cDC transcriptional hallmarks that characterize the IkL/L CDPs were also retained in the BM pDCs from IkL/L mice . Indeed , GSEA analysis showed that genes up- or down-regulated in IkL/L pDCs ( transcriptome data from [14] ) were also up- or down-regulated in mature cDCs ( S4C Fig ) , thereby confirming our hypothesis . Among the genes deregulated in IkL/L CDPs , only 70 were differentially expressed between GSI and DMSO treated samples ( Fig 6C , S1 Table ) . To identify the potential upstream pathways involved in the regulation of their expression in DC progenitors , we performed Ingenuity Pathway Analysis ( Fig 6D ) . This revealed Ikaros to be a significant probable regulator , which validated our approach , and showed the importance of Ikaros in CDPs . The top candidate , however , was TGFβ1 , which was interesting because TGFβ1 was previously reported to skew CDP differentiation towards the cDC lineage at the expense of pDCs [9] . We therefore asked if the deregulated genes found in IkL/L CDPs were enriched for TGFβ-associated genes , by GSEA . These results showed a strong and direct correlation between the genes down-regulated in IkL/L CDPs and those down-regulated by TGFβ1 signaling ( Fig 6E ) [9] . Conversely , the genes up-regulated in IkL/L CDPs were up-regulated by TGFβ1 activation ( Fig 6F ) . Thus , Ikaros expression is correlated with reduced TGFβ1 signaling in CDP cells . To determine if Ikaros directly regulates the TGFβ pathway , we investigated its capacity to bind TGFβ target genes . The low number of CDPs in WT mice did not allow us to directly investigate Ikaros binding in these cells . We therefore compared Ikaros binding to chromatin from 2 unrelated precursor cell types ( pre-B cells and DN3 thymocytes ) [31 , 32] , because conserved binding might indicate that Ikaros regulates similar elements across hematopoietic cell types . These analyses showed strong and conserved Ikaros binding to several TGFβ target genes implicated in DC differentiation ( eg . Axl , Irf1 , Irf4 , Nfkb2 , Nfkbie , Rel , Relb ) ( Fig 6G and S4 Fig ) , and suggested that Ikaros may directly regulate the expression of TGFβ target genes in CDPs . To determine if the TGFβ pathway is activated in IkL/L CDPs , we studied the mRNA expression of genes encoding upstream components of this pathway . Although the level of transcripts encoding the type I and type II TGFβ receptors , and the downstream SMAD proteins , were similar between WT and IkL/L MDPs and CDPs , we found that the mRNA levels of Eng , encoding the type III TGFβ receptor Endoglin , was higher in IkL/L MDPs ( 2x ) and CDPs ( 2 . 8x ) , regardless of GSI treatment ( Fig 6A ) . Endoglin ( CD105 ) , is an auxiliary receptor for the TGFβ receptor complex , which has been shown to positively modulate TGFβ signaling [33] . Higher CD105 expression was also detected on IkL/L MDPs , CDPs and pDCs ( Fig 7A and 7B ) . In contrast , CD105 levels were stable in other BM populations , including CD11c+CD317- cDCs ( Fig 7B ) , indicating that Endoglin expression is specifically increased in IkL/L pDCs and DC progenitors . In addition , we observed that Ikaros bound to the Eng locus in pre-B and DN3 cells , suggesting that it is an Ikaros target gene ( S5A Fig ) . Lastly , we analyzed the functional consequence of TGFβ inhibition on pDC development . WT and IkL/L BM cells were cultured with Flt3L for 8 days , in the presence or absence of a TGFβR1 inhibitor ( SB431542 ) , and/or GSI . SB431542 treatment alone did not affect total cell numbers ( S5B Fig ) , but increased pDC numbers in both WT and IkL/L cultures ( Fig 7C and 7D ) . In contrast , GSI treatment alone increased both total cell numbers and pDC numbers . The combination of SB431542 and GSI gave similar total and pDC numbers , compared with GSI alone . These results suggested that TGFβ inhibition promotes pDC differentiation in IkL/L CDPs .
Here we identify Ikaros as a promoter of early DC development . We show that Ikaros cooperates with Notch signaling to enhance the emergence and/or survival of MDPs and CDPs in the BM ( Fig 7E ) . We also show that Ikaros is required to promote CDP differentiation and cell fate specification towards the pDC and cDC lineages , in large part by correctly regulating the expression of DC-specific target genes , and secondly , by antagonizing TGFβ function . These results indicate that the general absence of mature DCs in Ikaros null mice [13 , 34] , as well as the selective absence of pDCs in Ikaros hypomorphic animals [14] , are due at least in part to CDP defects . Our results suggest that Ikaros antagonizes a TGFβ-dependent gene expression program in CDPs . TGFβ was previously reported to skew CDP differentiation towards the cDC lineage at the expense of pDCs , in part because it induces the expression of Id2 , which inhibits the master pDC regulator E2 . 2 [9 , 35–37] . We show that Ikaros-deficient CDPs display a premature cDC gene expression signature , indicating that Ikaros represses the expression of mature cDC-associated genes in DC progenitors . In addition , Ikaros-deficient BM pDCs also display a cDC signature , suggesting that the mutant CDPs that commit to the pDC lineage continue to express a promiscuous cDC gene expression program . Neither Id2 nor E2 . 2 are affected at the mRNA level in Ikaros-deficient CDPs and pDCs , suggesting that Ikaros promotes CDP differentiation independently of E2 . 2 . How Ikaros antagonizes TGFβ function remains only partially understood . Certain TGFβ target genes are enriched among the genes deregulated by GSI in CDPs . Furthermore , Ikaros-deficient CDPs ectopically express high levels of endoglin which can potentiate TGFβ signaling [38] . Because no other TGFβ receptors or downstream SMAD factors are deregulated in these cells , endoglin upregulation probably plays an important role in activating the TGFβ pathway in the mutant DC progenitors . Interestingly , betaglycan , a type III TGFβ receptor closely related to endoglin in structure and function , is a substrate of γ-secretase , and GSI inhibits TGFβ2-mediated reporter gene expression via betaglycan inactivation in HepG2 cells [39] . γ-secretase cleavage of type III TGFβ receptors may therefore inhibit TGFβ receptor signaling in Ikaros-deficient cells . If true , this suggests that Ikaros may be a novel upstream regulator of TGFβ signaling . In addition to its role in CDP differentiation , Ikaros is also required for MDP and CDP homeostasis . Observed only in compound mutants deficient for Ikaros and RBPJ where both populations are absent , our results demonstrate that Ikaros cooperates with Notch activation to maintain DC progenitor survival and/or expansion . Notch signaling by itself was previously found to promote DC development in vitro via up-regulation of the Frizzled family Wnt receptors [10] , but the basis for its cooperation with Ikaros remains to be elucidated . We have reported that Ikaros antagonizes Notch function in T cells , and interacts directly with the activated Notch1 protein to control a set of common target genes [40] . Whether Ikaros and Notch regulate common genes in DC progenitors remains to be investigated . Other studies have suggested that the Notch and TGFβ pathways interact to regulate common genes . Indeed , Hes1 is a common target of both pathways , because it is transcriptionally regulated by the Notch receptor intracellular domain or by Smad3 following TGFβ signaling [41] . In IkL/L cells , however , Hes1 up-regulation was observed in pDCs but not in the more immature dendritic progenitors , suggesting that Hes1 is differentially regulated by Notch and TGFβ activation in these populations . Finally , our results with γ-secretase inhibitors are unexpected and intriguing , and indicate that these compounds can be exploited to enhance and rescue WT and Ikaros-deficient DC development in vitro , though the effects are stronger in the mutant cells . We showed that transient GSI treatment promotes the generation of CD11c- cells , probably the upstream precursors of MDPs and CDPs , and pDC differentiation from CDPs . These actions suggest that GSI molecules might be considered as a potential treatment to enhance pDC function during certain viral infections , like chronic HIV or hepatitis C virus . Conversely , it will be important to test if GSI molecules might have a differentiating effect on BPDCN cancers , a rare and fatal leukemia with few options for treatment [42] .
All mouse procedures were approved by the IGBMC Ethical Committee ( Com'Eth ) ; APAFIS#8752–20 170 1261 0337966 v2 . The mouse lines used in this study were described previously: IkL/L , Hes1-EmGFPSAT , RBPJf/f and R26-CreER ( T2 ) [19 , 23 , 29 , 30] . Mice were used between 6–9 weeks of age . To delete Rbpj in vivo , RBPJf/f R26-CreET ( T2 ) + or IkL/L RBPJf/f R26-CreET ( T2 ) + mice were injected intraperitoneally daily for 5 days with 75 mg/kg of tamoxifen dissolved in sunflower oil , and analyzed 10 days after the first injection . pDC cultures were performed as described [26] . Briefly , BM cells were seeded at 2x106 cells/ml , and cultured in RPMI 1640 containing 10% fetal calf serum , 20 mM HEPES , 2 mM L-glutamine , 2 mM Sodium Pyruvate , 50 μM β-mercaptoethanol , 1x MEM non-essential amino acids , and antibiotics . Cultures were supplemented with conditioned medium from a Flt3L-producing cell line ( B16-Flt3L ) [43] , or rFlt3L at 100 ng/ml ( Peprotech ) . After 4d , half of the medium was replaced with fresh medium containing 2x Flt3L . GSI ( Compound E , Calbiochem ) and SB431542 ( Selleckchem ) were used at 5 μM . pDC cultures from DC progenitors were performed as above in 1 ml of Flt3L-supplemented medium using FACS-sorted Lin-Sca1-ckit+ , MDPs , CDPs or CD115-CDPs from BM cells ( CD45 . 2+ ) which were co-cultured with 106 CD45 . 1+ whole BM cells . For CpG oligo-deoxynucleotide ( ODN ) stimulations , pDCs ( CD11c+CD317+CD11b- ) were sorted after 8 days of culture and stimulated at 2x106 cells/ml in 96-well plates . CpG ODN 1585 or an ODN control ( InvivoGen ) were used at 2 . 5 μM . Cells were collected after 16h of stimulation . BM cells from donor mice ( CD45 . 2+ ) were cultured with Flt3L in the presence or absence of GSI for 48h . 2x105 cells from these cultures were then transplanted with 2x105 supporting WT BM cells ( CD45 . 1+ ) into lethally-irradiated ( 9 Gy ) CD45 . 1+CD45 . 2+ recipient mice . Mice were sacrificed and analyzed 9 days after the transfer . RNA was extracted with the RNeasy ( Qiagen ) or Nucleospin RNA ( Macherey-Nagel ) kits , and reverse transcribed using Superscript II ( Invitrogen ) . Hes1 , Ptcra and Hprt were amplified using the QuantiTect SYBR green system with the Mm_Hes1_1SG , Mm_PtcrA_1SG and Mm_Hprt_1SG primer sets ( Qiagen ) . Ifna mRNA was amplified using the SYBR green master mix ( Roche ) with 50 cycles of 10s 95°C , 30s 66°C , 15s 72°C . Primers used to amplify most of the Ifna subtypes were 5'-cctgctggctgtgaggaaata and 5'-gcacagggggctgtgtttct . Primers for Ubiquitin ( Ubb ) were 5'-tggctattaattattcggtctgcat and 5'-gcaagtggctagagtgcagagtaa . Hes1 and Ptcra levels were normalized to that of Hprt , while Ifna expression was normalized to that of Ubb . We used the following antibodies: anti-CD11b ( M1/70 ) eFluor450 or PE; anti-CD11c ( N418 ) AlexaFluor700; anti-human/mouse CD45R ( B220 ) eFluor650NC; anti- CD59 and Gr1 ( RB6-8C5 ) biotin; anti-CD199 ( CCR9 ) PE/Cy7; anti-CD317 ( ebio927 ) AlexaFluor488 or eFluor450; anti-MHCII ( M5/114 . 15 . 2 ) FITC or PE/Cy5; anti-Sca1 ( D7 ) biotin ( eBioscience ) ; anti-CD3 ( 145-2C11 ) biotin; anti-CD4 ( RM4-5 ) biotin; anti-CD8 ( 53–6 . 7 ) biotin; anti-CD11b ( M1/70 ) biotin; anti-CD45 . 1 ( A20 ) PE; anti-human/mouse CD45R ( B220 ) biotin; anti-CD115 ( c-fms ) APC; anti-CD135 ( A2F10 ) PE; anti-CD172a ( P84 ) APC; anti-NK1 . 1 ( PK136 ) biotin; anti-Ter119 biotin ( BD Biosciences ) ; anti-CD11c ( N418 ) biotin or APC; anti-CD19 ( 6D5 ) biotin; anti-CD45 . 1 ( A20 ) FITC; anti-CD45 . 2 ( 104 . 2 ) PE or AlexaFluor700; anti-CD105 ( Endoglin ) Alexa488; anti-CD117 ( c-kit ) APC/Cy7; anti-Ly49Q ( 2E6 ) PE ( MBL ) ; anti-SiglecH ( 551 . 3D3 ) PE ( BioLegend ) ; AlexaFluor™ 405 ( InvitroGen ) or AlexaFluor488 Streptavidin ( Jackson ImmunoResearch ) . Lineage staining was performed using a mixture of anti-CD3 , -CD4 , -CD8 , -CD19 , -CD11b , -CD11c , -Gr1 , -Ter119 , -NK1 . 1 and -B220 antibodies for Lin , and anti-CD3 , -CD19 , -Ter119 , -NK1 . 1 and -B220 for Lin* . Cells were analyzed on a LSRII analyzer ( BD Biosciences ) and sorted on a FACSAriaIISORP ( BD Biosciences ) . Sort purity was >98% . Total protein extracts from 106 BM cells were separated on SDS-PAGE gels . Immunoblots were analyzed with anti-RPBJ ( T6719; Institute of Immunology , Japan ) , and anti-β-actin ( A5441 , Sigma ) polyclonal antibodies . All secondary antibodies were horseradish conjugated ( Santa Cruz , Jackson ImmunoResearch ) . Transcriptome analyses were performed with Affymetrix Gene ST 1 . 0 arrays . Unsupervised hierarchical clustering and K-means clustering were performed using Cluster 3 . GSEA was performed using the GSEA 2 . 0 software [44 , 45] . Microarray data are available in the GEO databank ( GSE114108 ) . | Dendritic cells ( DCs ) are an important component of the immune system , and exist as two major subtypes: conventional DCs ( cDCs ) which present antigen via major histocompatibility class II molecules , and plasmacytoid DCs ( pDCs ) which act mainly as producers of type-I interferon in response to viral infections . Both types of DCs derive from a common dendritic progenitor ( CDP ) , but the genetic pathways that influence their development are not completely understood . A better understanding of these pathways is important , which may lead to protocols for generating specific DCs in culture , depending on the need . In this study , we have discovered important roles for the Ikaros transcription factor in DC development . We found that: ( i ) Ikaros cooperates with the Notch pathway to promote the development or homeostasis of CDPs; ( ii ) Ikaros controls pDC differentiation from CDPs through a γ-secretase sensitive pathway; and ( iii ) Ikaros antagonizes the TGFβ pathway to inhibit cDC differentiation . Our results thus identify Ikaros as a key player in the early steps of DC development . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"b",
"vitamins",
"chemical",
"compounds",
"biotin",
"gene",
"regulation",
"organic",
"compounds",
"notch",
"signaling",
"cell",
"differentiation",
"developmental",
"biology",
"regulator",
"genes",
"genome",
"analysis",
"gene",
"types",
"genomics",
"gene",
"expression",
"chemistry",
"vitamins",
"signal",
"transduction",
"cell",
"biology",
"organic",
"chemistry",
"tgf-beta",
"signaling",
"cascade",
"genetics",
"biology",
"and",
"life",
"sciences",
"transcriptome",
"analysis",
"physical",
"sciences",
"computational",
"biology",
"cell",
"signaling",
"signaling",
"cascades"
] | 2018 | Ikaros cooperates with Notch activation and antagonizes TGFβ signaling to promote pDC development |
From early dinosaurs with as many as nine wrist bones , modern birds evolved to develop only four ossifications . Their identity is uncertain , with different labels used in palaeontology and developmental biology . We examined embryos of several species and studied chicken embryos in detail through a new technique allowing whole-mount immunofluorescence of the embryonic cartilaginous skeleton . Beyond previous controversy , we establish that the proximal–anterior ossification develops from a composite radiale+intermedium cartilage , consistent with fusion of radiale and intermedium observed in some theropod dinosaurs . Despite previous claims that the development of the distal–anterior ossification does not support the dinosaur–bird link , we found its embryonic precursor shows two distinct regions of both collagen type II and collagen type IX expression , resembling the composite semilunate bone of bird-like dinosaurs ( distal carpal 1+distal carpal 2 ) . The distal–posterior ossification develops from a cartilage referred to as “element x , ” but its position corresponds to distal carpal 3 . The proximal–posterior ossification is perhaps most controversial: It is labelled as the ulnare in palaeontology , but we confirm the embryonic ulnare is lost during development . Re-examination of the fossil evidence reveals the ulnare was actually absent in bird-like dinosaurs . We confirm the proximal–posterior bone is a pisiform in terms of embryonic position and its development as a sesamoid associated to a tendon . However , the pisiform is absent in bird-like dinosaurs , which are known from several articulated specimens . The combined data provide compelling evidence of a remarkable evolutionary reversal: A large , ossified pisiform re-evolved in the lineage leading to birds , after a period in which it was either absent , nonossified , or very small , consistently escaping fossil preservation . The bird wrist provides a modern example of how developmental and paleontological data illuminate each other . Based on all available data , we introduce a new nomenclature for bird wrist ossifications .
The wing of birds is highly derived , having reduced the number of ossifications present in the wrist . Early dinosaurs had as many as nine ossifications ( Figure 1A ) [1] , whereas in birds , only four carpal ossifications remain , two distal and two proximal ( Figure 1B ) [2] . The two distal ossifications fuse to each other and to the metacarpi in the adult , forming part of the carpometacarpus . The two proximal ossifications do not fuse and are large , independent bones . Currently , the identity of all four ossifications is debatable . Importantly , two classic research fields , palaeontology and developmental biology , often label these bones differently . Figure 1C shows an identification of avian carpal ossifications commonly used in palaeontology , and Figure 1D shows another common for developmental biology , but different combinations of these labels may be found in any field , reflecting current confusion [3] . An important debate also exists over the identity of the digits of the bird wing: Traditionally , palaeontology labels them 1 , 2 , 3 [4] , [5] , whereas developmental biology labels them 2 , 3 , 4 [6]–[11] . In view of recent developmental evidence for 1 , 2 , 3 [12]–[14] , we will use 1 , 2 , 3 to refer to the digits and , especially so , their associated distal carpals ( here , dc1 , dc2 , and dc3 ) . However , it must be kept in mind that most developmental studies traditionally refer to the same distal carpals as dc2 , dc3 , and dc4 [3] , [8] , [15] . Developmental and paleontological data are routinely used for identifying homologies . They often illuminate and support each other , as shown by classic examples such as the bones of the mammalian middle ear [16] , [17] . Potential conflicts of data are thus important , demanding for an explanation and coherent integration . Here , we have studied the development of the embryonic wrist skeleton using classic clearing and staining techniques [18] for a broad taxonomic sample of species: wreath lizard , yacare caiman , Chilean tinamou , chicken , mallard duck , rock pigeon , Chilean lapwing , zebra finch , and budgerigar ( Phylogenetic relationships among these taxa [19]–[22] are presented in Figure 2 ) . We also used stacks of histological sections to assess tissue organization , such as the presence of an internal separation or “septum” within allegedly composite cartilages . Importantly , we used a new technique for whole-mount immunostaining of proteins expressed within embryonic cartilages . Traditional protocols only allowed antibodies to penetrate cartilage in thin histological sections . We observed the expression in chicken embryos of collagen type II ( Coll II ) , which marks cartilage formation [23]–[26] and collagen type IX ( Coll IX ) , which is indicative of endochondral cartilage maturation [24]–[27] . We also reviewed the paleontological evidence on the carpal bones present during the evolution of the bird line . This included direct observation of specimens in museum collections , especially “bird-like dinosaurs” ( the closest nonavian relatives of birds—that is , maniraptorans like Oviraptorosauria , Dromaeosauridae ) . The integration of our new developmental data with the information provided by the fossil record has important consequences for understanding the evolution of avian wrist bones , leading us to propose a new nomenclature .
Developmental studies are unclear on the identity of the proximal–anterior carpal bone ( anterior = medial ) . Some describe it as developing from a single radiale cartilage [6] , [8] , whereas others describe a composite of the radiale+intermedium cartilages [15] , [28] . In palaeontology , this bone is often labelled as the radiale in birds and bird-like dinosaurs , whereas ornithologists often use the term “scapholunare , ” a composite of the mammalian terms scaphoid ( radiale ) and lunare ( intermedium ) [29] . Whole-mount alcian blue staining in the chicken and budgerigar provides no evidence for two distinct elements , although diffuse staining is present in the entire anterior–central region ( Figure 3A ) , where both radiale and intermedium would be in other amniotes [28] . However , tissue organization in histological sections of chicken reveals two separate elements ( Figure 3B ) . Whole-mount immunofluorescence also reveals two distinct regions of Coll II expression at early stages ( Figure 3C ) . Traditional techniques for visualizing cartilage stain hyaluronic acid and glycosaminoglycans , which are highly concentrated in cartilage but are also present in other connective tissues [30] . Alcian blue often leads to diffuse staining and uncertainty on the number and limits of elements present . Coll II , in contrast , has only been reported in cartilage [23] , [26] , which may explain why we sometimes found more specific foci within larger domains of weak or diffuse alcian blue staining . Importantly , we found that in early embryos of duck , pigeon , tinamou , and zebra finch , whole-mount alcian blue staining is sufficient to observe a separate radiale and intermedium ( see pigeon and Chilean tinamou in Figure 3D , E ) . This condition was also previously reported in falcons , but no data were shown [28] . Thus , evolutionary variation is present , with greater coalescence of these cartilages in the chicken ( Galliformes ) and budgerigar ( Psittaciformes ) . These findings illustrate the advantages of observing several species: Some patterns of skeletogenesis are more easily detected in nonmodel taxa . At later stages , a single cartilage is apparent . However , the shape of this cartilage presents two distinct “lobes” ( Figure 3F , G ) that are also observable using Coll II expression ( Figure 3H ) . Histological sections in chicken indicate complete coalescence , with no septum or traces of separation , and a continuous cartilage matrix surrounded by a single perichondrium ( Figure 3I ) . However , two very distinct regions of late Coll IX expression are present within this cartilage . We confirmed these are largely separate domains using spinning-disc microscopy and 3D reconstruction , avoiding the effects of shape and superposition ( Figure 3J ) . Coll II is expressed upon cartilage formation , but Coll IX relates to cartilage differentiation: after cartilage formation , but before hypertrophy [24] , [31]–[33] . Accordingly , we have observed that the onset of Coll IX expression occurs after that of Coll II , and never outside boundaries of larger Coll II expression domains . Developmental studies have identified a single distal carpal 2 ( dc2 ) cartilage , at the proximal end of metacarpal 2 [6]–[8] , which gives rise to the distal–anterior ossification of birds . Palaeontologists label this ossification as the semilunate , a bone that in dinosaurs is a composite of dc1+dc2 . In embryos of the multiple bird species we observed , traditional whole-mount alcian blue staining shows a single region of continuous staining , providing no evidence for a composite of two elements ( Figure 4A ) . Histological sections at early stages of the chicken are ambiguous , revealing asymmetric tissue organization in this region , with weak alcian blue staining towards anterior and strongly stained , concentrically arranged cells towards posterior ( posterior = lateral ) ( Figure 5A–B ) . However , both Coll II and Coll IX in the chicken show two distinct regions of expression ( Figure 4B and 4C ) . At later stages , uniform Coll II expression indicates the cartilage matrix is continuous ( Figure 4D ) , and histological sections show a single , well-defined cartilage with no internal separation ( Figure 5C ) . However , Coll IX expression continues to show two very distinct , mostly separate regions ( Figure 4E ) , as confirmed by 3D spinning disc microscopy ( Figure 4F and Video S1 ) . In birds , the embryonic cartilage of the ulnare forms early as is typical for amniotes , being the first carpal element formed , at the distal end of the ulna ( Figure 6A ) . We confirm previous reports [8] , [34] that the ulnare thereafter ceases to grow and is lost in development ( Figure 6A–D ) . In the chicken , a late-forming cartilage has been described to “replace” the ulnare , which has been called “element x , ” suggesting it is a neomorphic element of birds [6] . This cartilage gives rise to the distal–posterior ossification in the bird wrist . These descriptions did not document whether “element x” is formed after the disappearance of the ulnare . We now present evidence that “element x” temporally coexists with the ulnare in the chicken , as observed using alcian blue whole mounts ( Figure 6A; “element x” is labelled as Dc3 ) , collagen expression ( Figure 6B–C and Video S2 ) , and histological sections ( Figure 6D ) . “Element x” and its coexistence with the ulnare was also observed in alcian blue whole mounts of tinamou , lapwing , pigeon , budgerigar , zebra finch , and duck ( Figure 7 ) . Although “element x” has been argued to “replace” the ulnare , we find this notion is misleading , as in all species observed it is distal to it and at the proximal end of metacarpal III , a position that corresponds to distal carpal 3 . This is especially evident in the Chilean lapwing ( Figure 7C , HH32 ) . Embryonic cartilages of distal carpals can form late , proximal to the preexisting metacarpi [35] , as observed for dc1 of the alligator [36] . Thus , we find no compelling reason to consider “element x” is neomorphic or a replacement of the ulnare . Rather , the term “distal carpal 3” is appropriate for this cartilage and the posterior–distal ossification that thereafter develops from it . Although the proximal–posterior carpal of birds is often identified as the ulnare in palaeontology , the embryonic ulnare is actually lost during avian development ( above section , Figures 6 and 7 ) . Most developmental studies identify the proximal–posterior bone as the pisiform [6] , [8] . Our observations confirm it originates from the embryonic cartilage that forms ventrally displaced and posterior to the contact between the ulnare and the ulna , a position that gives rise to the pisiform in other amniotes ( Figure 8A–D ) . The pisiform is a sesamoid that forms associated to a tendon at an articulation joint [37] , [38] , much like the patella in the knee . In monotremes , marsupials , placentals , turtles , lepidosaurs ( tuatara and “lizards” ) , and crocodylians , this tendon belongs to the flexor carpi ulnaris muscle , which begins from the epicondylus ventralis of the humerus , glides through the proximal end of the ulna , and attaches to the posterior side of the pisiform [39]–[47] . Immunosflourescence for tenascin confirms that the corresponding embryonic muscle of birds is attached posteriorly to the cartilaginous precursor of the proximal–posterior bone during its formation ( Figure 8E ) , indicating it is a sesamoid , as expected for a pisiform . The evolution of the wrist bones in the lineage leading to birds since early dinosaurs is summarized by the taxon sample shown in Figure 9 , including phylogenetic relationships [48]–[51] . Regarding the identity of the proximal–anterior bone , our data have confirmed it develops from an embryonic cartilage that is a composite Radiale+Intermedium . A separate ossification of the intermedium ( orange in Figure 9 ) has been described in some theropods such as Coelophysis rhodesiensis , Gorgosaurus libratus , and Guanlong wucaii [52]–[55] . Its presence has sometimes been overlooked , as in Acrocanthosaurus atokensis and Allosaurus fragilis , where it was mistakenly identified as the ulnare [56]–[59] . In all these taxa , the ossification of the intermedium is closely appressed or fused to the posterior aspect of the radiale ( purple in Figure 9 ) , providing evidence that is consistent with the evolution of a composite radiale+intermedium in birds ( purple–orange in Figure 9 ) . In the cartilaginous region that becomes the distal–anterior bone of the bird wrist , the presence of two domains of collagen expression is especially significant when paleontological data are integrated . This bone is comparable to the semilunate carpal of bird-like dinosaurs , which covered the proximal ends of both metacarpals I and II , and is considered a composite of dc1+dc2 . In early lineages like Allosaurus fragilis , dc1 and dc2 were separate ossifications ( yellow and green , respectively , in Figure 9 ) . In some coelurosaurs such as Harpymimus okladnikovi , Alxasaurus elesitaiensis , and Falcarius utahensis [60]–[64] , it presented a clear midline suture , indicating the presence of two roughly equal , fused ossifications of dc1 and dc2 . In taxa closer to birds , and in Mesozoic birds , a suture line is no longer observable , suggesting a single ossification [65] , although the suture may have been lost through bone remodelling during ontogeny [66] . The labelling of the ulnare reveals an apparent contradiction between palaeontology and developmental biology . Most paleontological papers identify the ulnare as present in the bird wrist . Previous embryological studies , however , described the embryonic ulnare was lost and “replaced” by a neomorphic “element x” or pseudoulnare . This complex process was not well documented , allowing for skepticism . According to our developmental data , “element x” is actually dc3 , which becomes the posterior–distal ossification: Whether it is a replacement of the ulnare is debatable ( see above sections , Figures 6 and 7 ) . However , we fully confirm that the embryonic ulnare is lost in avian development . This provides a strong reason to reexamine the evidence in a broad set of fossil taxa for labelling this bone as being present in birds . Indeed , except in the earliest lineages of theropod dinosaurs [67]–[69] and possibly the Ornithomimosauria [63] , [70] , there is no evidence of an ulnare ( brown in Figure 9 ) . Importantly , there is no ulnare in the most bird-like dinosaurs ( Oviraptorosauria , Dromaeosauridae , Troodontidae [71]–[75] ) , which are known from several well-preserved , articulated specimens ( Figure 9 ) . In many theropods , the ulnare was mistakenly considered present , having been confused with other elements , such as the intermedium [56] , distal carpal 2 [76]–[78] , and the posterior–distal dc3 , which in modern adult birds fuses to the carpometacarpus [79] , [80] . In early dinosaurs , some bird-like dinosaurs , and Mesozoic birds , dc3 is observable as a separate bone ( blue in Figure 9 ) that has been variably labelled as the ulnare , “element x” [81]–[84] , or dc3 [85] . The proximal–posterior bone of the bird wrist ( red in Figure 9 ) poses the greatest challenge to interdisciplinary integration . Paleontological data would seemingly exclude the hypothesis that it is a pisiform , because it provides evidence for its loss in the lineage leading to birds . Except for early theropods [52] , and possibly the Ornithomimosauria [63] , [86] , the pisiform is absent . The most bird-like dinosaurs show the presence only of the semilunate , the scapholunare ( often labelled “radiale” ) , and occasional preservation of dc3 [87] , but no pisiform ( Figure 9 ) . Thus , if present , the pisiform must have been at least very small or nonossified , consistently escaping preservation . Developmental data , in turn , provide compelling evidence that the large posterior–proximal ossification of modern birds ( often preserved in their fossil relatives ) is in fact a pisiform in terms of its embryological position , its sesamoid nature , and its muscular connectivity . An integrative explanation for both developmental and paleontological evidence is that a large , ossified pisiform was reacquired in the evolution of birds , after a period in which it was at least strongly reduced ( Figure 9 ) . Its evolutionary reappearance as a large , posteriorly displaced proximal carpal occurred early in the evolution of birds , consistently observed in Mesozoic taxa such as the cretaceous long-tailed bird Shenzhouraptor sinensis , the basal pygostylians Sapeornis chaoyangensis [83] , and Confuciusornis sanctus ( [88] , personal observation ) . A bone in appropriate position has been reported in the Eichstätt specimen of Archaeopteryx [89] , but not other specimens [90] . In other Mesozoic taxa closer to modern birds ( Ornithothoraces , [91]–[94] ) , this bone became v-shaped , like the pisiform of modern birds [95] , [96] .
The development of living species is expected to contain signs of their evolutionary lineage of origin . Because radically different data sources about evolution are available ( fossils vs . molecular and cell biology ) , transdisciplinary integration provides a great opportunity for independent confirmation . Several examples exist where molecular-developmental observations show great consistency with the information provided by the fossil record [105]–[107] . Sometimes , however , each area seemingly arrives to a different conclusion . It often occurs that one of the data sources needs revision or updating . However , when all facts are well documented , apparent contradictions may point to the need for a different interpretation . For instance , an explanation may be found in a previously unsuspected evolutionary transformation [12] , [108] . In the case of the bird wrist , a renewed look found support in both data sources for a composite radiale+intermedium , which had often been simply labelled as the radiale in both fields . The evidence for a composite semilunate cartilage shows how , despite claims to the contrary , avian development contains signals that are consistent with their origin from dinosaurs , which is a well-documented fact of palaeontology . Our detailed confirmation of the developmental loss of the ulnare led us to reexamine updated evidence from the fossil record . Paleontological evidence in fact strongly supports the loss of the ulnare in the bird line , ultimately revealing no inconsistency with developmental data . Perhaps the most interesting result of combining data sources is provided by the case of the pisiform . Sound fossil evidence indicates this ossification was absent in bird ancestors , but using developmental evidence alone would decidedly identify this bone as present in modern birds . The evolutionary reacquisition of a large ossified pisiform in birds can explain how both data sources could in fact be correct . The notion of important evolutionary reversals has historically met a lot of resistance in evolutionary thinking [109] . Although its empirical reality is now accepted [110] , it continues to be considered an oddity [111] . The reappearance of the pisiform in birds provides a compelling case documenting this intriguing evolutionary phenomenon . Integrating developmental and paleontological information can thus also be informative about what evolutionary processes are actually possible . These transformations would be hard to detect using only one source of information . Palaeontology and developmental biology often have radically different research objectives and methods . However , they intersect significantly . The avian wrist provides a striking new example of how they can illuminate each other in concrete ways . This is reflected in our updated proposal on the identity of bird wrist bones ( Figure 1E ) . Evolution , as documented by the fossil record , provides natural experiments that are outputs of the same developmental mechanisms that are conserved in living organisms . A complete separation of development and palaeontology misses opportunities for understanding evolution , much like a separation of astronomy and experimental physics would delay the advances of cosmology .
All procedures were formally approved by the Comité de Etica de la Facultad de Ciencias , Universidad de Chile , which certifies compliance with all aspects required for government funding ( http://www . conicyt . cl/fondecyt/2012/10/31/bioetica/ ) . Live animals were only kept to obtain fertilized eggs: None were euthanized or used for experiments . None of the wild species used is in a conservation category of concern ( http://www . iucnredlist . org ) , and eggs were taken with field permits of the Servicio Agrícola y Ganadero ( SAG , Government of Chile ) . Fertilized eggs of Gallus gallus ( Chicken , Galliformes ) , Anas platyrhynchos ( Mallard duck , Anseriformes ) , and Nothoprocta perdicaria ( Chilean tinamou , Tinamiformes ) were purchased from local farms: Chorombo S . A , Avícola Metrenco , and Tinamou Chile ( perdiz . cl ) . Fertilized eggs of Columba livia ( Rock pigeon , Columbiformes ) , Taeniopygia guttata ( Zebra finch , Passeriformes ) , and Melopsittacus undulatus ( Budgerigar , Psittaciformes ) were obtained from birds kept at facilities of the Faculty of Science , University of Chile . Eggs of Liolaemus lemniscatus ( Wreath lizard , Iguania ) were obtained from gravid females captured and then liberated with SAG permit ( AOV ) , and were incubated with the same protocol for Liolaemus tenuis [112] . Caiman yacare embryos ( Yacare caiman , Alligatoridae ) belong to Paula Bona ( Museo de La Plata , Argentina ) . Fertilized eggs of Vanellus chilensis ( Chilean lapwing , Charadriiformes ) were collected with permission from SAG ( D . S . -P . and D . N . -L . ) . The following specimens and casts of fossil taxa were observed directly: American Museum of Natural History , New York: Archaeopteryx lithographica ( Avialae ) cast FR 5120 ( Berlin specimen ) and cast FR 9495 ( Eichstatt specimen ) , Bambiraptor feinbergi ( Dromaeosauridae ) AMNH 30554 , Citipati osmolskae ( Oviraptorosauria ) TG002 IGM 100/978+3063 , Coelophysis bauri ( Coelophysoidea ) 30631 , and Khaan mckennai ( Oviraptorosauria ) MPC-D 100/1002 , MPC-D 100/1127; Natural History Museum of Los Angeles County: Confuciusornis sanctus ( Confuciusornithidae ) cast LACM 7852; Museo de Ciencias Naturales , Universidad Nacional de San Juan: Herrerasaurus ischigualastensis ( Herrerasauridae ) PVSJ 373; University of Texas: Coelophysis bauri ( Coelophysoidea ) MCZ 4329; University of California Museum of Paleontology , Berkeley: Dilophosaurus wetherelli ( Coelophysoidea ) UCMP 37302; and Peabody Museum of Natural History , Yale University: Deinonychus antirrhopus ( Dromaeosauridae ) YPM 5206 , Tanycolagreus topwilsoni ( Coelurosauria ) cast YPM 56523 . Embryos were fixed in 100% methanol for 2–3 d at room temperature ( RT ) . Methanol was replaced by 5∶1 ethanol/acetic acid solution with 0 . 03% 8G alcian blue for 2 d at RT in an orbital shaker . Then , embryos were cleared in a sequence of 1∶3 , 1∶1 , and 3∶1 glicerol/water , and photographed in a stereoscopic microscope . Two embryos per stage were used for Chilean lapwing . Five embryos per stage were used for Chilean tinamou , duck , and pigeon . Two or more embryos per stage were used for zebrafinch , budgerigar , and chicken . Three embryos per stage were used for Wreath lizard , and one for Caiman . Embryos were fixed in 4% paraformaldehyde ( PFA ) for 2 h at RT or overnight at 4°C . Then , forelimbs were dissected , washed in PBS 1% , and dehydrated in ethanol increasing concentrations ( 50% to absolute ethanol ) . Limbs were cleared in Neoclear and vacuum-embedded in Paraplast for 1 h . The paraplast-embedded material was cut into 10–12-µm-thick sections , rehydrated , and stained with alcian blue/nuclear red or safranin/hematoxylin . Six embryos for each stage were used for immunofluorescence of each primary antibody . Embryos were fixed in Dent's Fix ( 4∶1 methanol/DMSO ) for 2 h at RT , dehydrated in 100% methanol , and left at −80°C overnight . Before immunostaining , they were bleached in Dent's Bleaching ( 4∶1∶1 methanol/DMSO/H2O2 ) for 24 h at RT . For anticollagen immunostaining , embryos were fixed and bleached as above . Then , limbs were dissected and digested with 2 mg/ml of hyaluronidase ( Sigma ) in PBS for 2 h at 37°C . Embryos were rehydrated in PBS 1% triton ( PBST ) and incubated in primary antibodies for 2 d at 4°C in an orbital shaker . Primary antibodies were diluted in 2% horse serum , 5% DMSO in PBST at the following concentrations—1∶10 anti-tenascin ( M1-B4 , DSHB ) ; 1∶40 Colagen type II ( II-II6B3 , DSHB ) ; and 1∶40 collagen type IX ( 2C2-II , DSHB ) —and washed in PBST ( 3×10 min and 3×1 h in an orbital shaker ) . Secondary antibodies antimouse ( Alexa-488 and Alexa-Fluor 594 , Jackson ImmunoResearch , PA ) were diluted in 5% goat serum , 5% DMSO in PBST and incubated for 24 h at 4°C . After that , they were washed , cleared with Urea 4M , and photographed in a fluorescent stereoscopic microscope ( Nikon ) . For 3D reconstructions , 10 µm stacks were obtained in a spinning disk confocal microscope ( Olympus ) and projected in cellSens software ( Cellsens analysis Z stack was obtained for ·D reconstruction and 2D deconvolution/Nearest Neighbor analysis was obtained for removal of background fluorescence ) . | When birds diverged from nonavian dinosaurs , one of the key adaptations for flight involved a remodelling of the bones of the wrist . However , the correspondence between bird and dinosaur wrist bones is controversial . To identify the bones in the bird wrist , data can be drawn from two radically different sources: ( 1 ) embryology and ( 2 ) the fossil record of the dinosaur–bird transition . Currently , identifications are uncertain , but new developmental data can help resolve apparent conflicts . The modern bird wrist comprises four ossifications , arranged roughly in a square with its sides running proximal/distal and anterior/posterior . Our study integrates developmental and paleontological data and clarifies the relationship between each of these four ossifications and those found in nonavian dinosaurs . This integrative approach resolves previous disparities that have challenged the support for the dinosaur–bird link and reveals previously undetected processes , including loss , fusion , and in one case , re-evolution of a transiently lost bone . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"developmental",
"biology",
"paleontology",
"biology",
"and",
"life",
"sciences",
"evolutionary",
"biology",
"zoology"
] | 2014 | New Developmental Evidence Clarifies the Evolution of Wrist Bones in the Dinosaur–Bird Transition |
Gene regulatory networks are a crucial aspect of systems biology in describing molecular mechanisms of the cell . Various computational models rely on random gene selection to infer such networks from microarray data . While incorporation of prior knowledge into data analysis has been deemed important , in practice , it has generally been limited to referencing genes in probe sets and using curated knowledge bases . We investigate the impact of augmenting microarray data with semantic relations automatically extracted from the literature , with the view that relations encoding gene/protein interactions eliminate the need for random selection of components in non-exhaustive approaches , producing a more accurate model of cellular behavior . A genetic algorithm is then used to optimize the strength of interactions using microarray data and an artificial neural network fitness function . The result is a directed and weighted network providing the individual contribution of each gene to its target . For testing , we used invasive ductile carcinoma of the breast to query the literature and a microarray set containing gene expression changes in these cells over several time points . Our model demonstrates significantly better fitness than the state-of-the-art model , which relies on an initial random selection of genes . Comparison to the component pathways of the KEGG Pathways in Cancer map reveals that the resulting networks contain both known and novel relationships . The p53 pathway results were manually validated in the literature . 60% of non-KEGG relationships were supported ( 74% for highly weighted interactions ) . The method was then applied to yeast data and our model again outperformed the comparison model . Our results demonstrate the advantage of combining gene interactions extracted from the literature in the form of semantic relations with microarray analysis in generating contribution-weighted gene regulatory networks . This methodology can make a significant contribution to understanding the complex interactions involved in cellular behavior and molecular physiology .
Previous computational models used to reconstruct gene regulatory networks from microarray data have employed techniques such as Hidden Markov , Bayesian network , and stochastic differential equation models [11]–[15] . In a Hidden Markov Model , the real states of genes are treated as hidden variables , and the gene expression values are observed . This permits the states of genes at a given time t to be considered as depending only on the previous time point t-1 . A Bayesian network is a representation of a joint probability distribution . When applying Bayesian networks to genetic regulatory systems , nodes are identified with genes and their expression levels , edges indicate interactions between genes , and conditional distributions describe these interactions . In the stochastic differential equation , the change of expression of a given target gene over two adjacent time points is equal to the accumulation of the weighted results of the sigmoid function of its regulator genes plus a random error . These computational models all reflect one or more aspects of the nature of genes and gene regulatory networks , but computational complexity limits the dimensionality of the modeled networks , and sensitivity to noise in gene expression measurement largely reduces their accuracy . Genetic algorithms ( GA ) [16] have also been widely used for the inference of gene regulatory networks [17]–[20] . Genetic algorithms are inspired by Darwin's theory of evolution . Within this methodology , a population of candidate solutions to a problem is created and then evolves over a specified number of generations using phenomena such as cross-over ( swapping components from other candidates ) and mutation ( internal , random changes ) . The best candidates are propagated through with incremental changes in the overall structure and the final generation becomes the solution . At each generation , fitness functions are used to determine the fittest candidates . A genetic algorithm is a stochastic algorithm and therefore it is highly likely to find global optima and can easily escape local maxima . Since the genetic algorithm execution technique is not dependent on the error surface , it is capable of solving multi-dimensional , non-differential , non-continuous , and even non-parametrical problems , which is the nature of gene expression data . Keedwell et al . used small random gene subsets evaluated by an artificial neural network ( ANN ) that is optimized by gradient descent to form the population of the genetic algorithm [21]; Liu and Wu used a differential equation to model the GRNs and genetic programming for optimization [22]; Sîrbu et al . compared various evolutionary algorithms for GRN inferencing and found that Keedwell et al . 's method performs the best overall [10] , prompting us to use their method for the basis of our model . They also note that using literature-derived knowledge offers the potential for significant improvement over techniques that probe component genes randomly , and our use of such knowledge forms the basis of our deviation from the Keedwell et al . 's model . With the rapid rate of growth in the biology literature , text mining is increasingly seen as indispensable in managing and discovering new biological knowledge [23] . An active area of research in biological text mining has been extraction of interactions between biomolecular entities ( genes , proteins , etc . ) from the research literature . Many systems , adopting various representational means ( binary interactions , events , etc . ) and using a variety of rule-based and machine learning-based techniques , have been proposed for this task . Early systems that focused only on co-occurrence of entities were soon replaced by systems that relied on shallow parsing and hand-crafted syntactic rules to extract binary interactions [24]–[26 , among others] . These methods generally provided high precision at the expense of lower recall , in contrast to co-occurrence based methods . More recently , dependency parsing has become the predominant syntactic tool in extracting biological relations , as evidenced by the BioNLP Shared Task competitions [27] . These competitions have also signaled the increasing focus on events as the representational means for biological relations . Most commonly , dependency relations have formed the basis for syntactic features ( shortest paths , dependency n-grams , etc . ) for machine learning methods , along with lexical ( tokens , n-grams , part-of-speech tags , etc . ) and semantic ( entity types , hypernyms , etc . ) features . Best machine learning approaches have included pipeline models based on support vector machines [28]–[29] as well as model combination techniques [30] and joint inference [31] . Some rule-based systems have reported competitive results in this task , as well [32] . Recently , coreference resolution has also been beneficially integrated into several event extraction systems [33]–[34] . As these text mining methods mature , they are increasingly applied to practical needs of biologists , in tasks such as database curation and pathway generation . For example , two tasks in the recent BioNLP 2013 Shared Task competition [27] ( Pathway Curation [35] and Gene Regulation Network in Bacteria [36] ) investigated the feasibility of automatically constructing such networks from the literature alone . Given a set of relevant biomolecular entities , the former focused on extracting pathway-relevant events ( e . g . gene expression , regulation , binding , regulation ) , while the latter focused on constructing a gene regulation network for the model bacterium Bacillus subtilis involving these entities . For the latter task , participating groups could either directly construct a regulation network for the provided entities or extract the interactions from which such a network could be derived using a predefined algorithm that the organizers provided . While the results were encouraging , these tasks remain challenging as evidenced by the limited participation and the fact that both tasks presupposed that the entities involved were already known , thus , addressing only a fraction of the problem of network construction from the literature . From an opposing viewpoint , Miwa et al . [37] focused on linking interactions in biological pathways to supporting evidence from the literature; however , their work does not address the task of pathway construction . There have been several attempts at improving gene regulatory network modeling by incorporating existing knowledge . For example , Steele et al . used the correlation between different gene concept profiles to calculate the probability of edges in GRNs modeled by Bayesian networks [38] . These profiles are determined by the occurrence of terms in the literature , using the Unified Medical Language System ( UMLS ) for normalization . Gutierrez-Rios et al . used regulatory interactions described in RegulonDB , a database of the regulatory network of Escherichia coli K-12 , to establish the network of causal relationships to evaluate the congruence between the literature and whole-genome expression profiles [39] . Additionally , the literature contains examples of efforts to combine literature-derived networks and microarray analysis in contexts other than GRN inferencing . Duarte et al . reconstructed the human metabolomic network depending largely on a manual literature review combined with knowledge extracted from genomic databases and use gene expression analysis to fill in the gap for a subset of network components [40] . Ashley et al . used text mining to derive biological pathways from literature relevant to in-stent restenosis and analyzed which of these were most relevant to the expression profiles identified by microarray analysis of tissue samples from patients with this condition [41] . All of these techniques either use human review to identify asserted interactions or automated approaches to infer them based on co-occurrence of terms . Our approach combines automation , allowing for an exponentially greater survey of the literature , with the identification of assertions in the text by SemRep , moving beyond mere term co-occurrence and increasing the validity of the extracted relations . Several books , book chapters , and journal reviews are available detailing the pros and cons of various modeling approaches for GRN inferencing [42]–[47] . In addition to a description of various approaches , Karlebach and Shamir [45] also provide a comparative summary of the relative advantages of different types of models for various features . They align GRN models along a spectrum from logical models ( e . g . Boolean networks ) to continuous models ( e . g . linear differential equations ) to single-molecule level models ( e . g . stochastic simulation models ) . They identify the logical end of the spectrum as having a decreased detail , less faithfulness to biological reality , lower data quantity needs , and reduced ability to model dynamics , while having greater model size , computational speed , and inferencing ability . Models at the single-molecule level are positioned at the other end of the spectrum , with a higher level of detail , increased faithfulness to biological reality , greater data quantity needs , and increased ability to model dynamics and decreased model size , computational speed , and inferencing ability . Continuous models are positioned in the middle of the spectrum , with moderate levels of the assessed model's characteristics . Although not providing a complete picture of the comparative characteristics of GRN modeling techniques , they give an easily accessible summarization . The breast cancer microarray data used in our experiments comes from NCBI's Gene expression omnibus ( GEO: http://www . ncbi . nlm . nih . gov/geo/ ) [59] . GEO provides free access to raw and processed microarray and sequencing data submitted by researchers based on their published work . For the yeast study , we used the cdc15 time course data from the Yeast Cell Cycle Analysis Project website ( http://genome-www . stanford . edu/cellcycle/ ) originally used in [60] . For assessment of our methodology , we use the Kyoto Encyclopedia of Genes and Genomes ( KEGG: http://www . genome . jp/kegg/ ) [61] , which provides gold standard sets of molecular pathways . KEGG pathways are manually curated networks based on a review of protein-gene and protein-protein interactions described in the literature . It is worth noting that KEGG pathway maps are an abbreviated representation of known interactions , focusing on those considered to be best supported by evidence and most relevant . As a consequence , the interactions in a KEGG pathway form a subset of those in the literature , and it is necessary to assess separately the validity of interactions not in the KEGG pathway . A KEGG pathway can be accessed as a downloadable kgml file or an online map , which have slight differences regarding which genes are included and which interactions are specified . To mediate these differences we use both formats for our assessment and consider whether genes and interactions are included in either version .
In the first step , predication network generation , predications containing gene-gene interactions ( INTERACTS_WITH , STIMULATES , and INHIBITS predications ) are extracted from the MEDLINE citations that supported manual curation of each KEGG pathway; these citations are listed in the pathway entry on the KEGG website . These predications are augmented with those from additional relevant citations , which we identify by first extracting the Medical Subject Heading ( MeSH ) terms for each citation identified in KEGG and then manually refining that list to eliminate non-specific terms such as Humans , Male , or Biological Models . The resulting MeSH terms formed the basis of our PubMed searches . The terms that occurred in a significant distribution across the citations were grouped with an “OR” in the query when they were roughly equivalent or part of a set of different subtopics . As an example , the query for p53 citations was “ ( Tumor Suppressor Protein p53[mh] OR Genes , p53[mh] ) AND ( Apoptosis[mh] OR Signal Transduction[mh] OR ( Phosphorylation[mh] AND ( Neoplasms[mh] OR Neoplasm Proteins[mh] OR Tumor Markers[mh] ) ) ) AND physiology[sh] AND metabolism[sh] . ” This procedure was repeated to provide a citation list for each pathway . The number of citations for each pathway is given in Table 1 in ‘Citation’ column . The predications extracted from these citations were then retrieved from SemMedDB . The number of resulting predications for each pathway is shown under the column heading ‘Raw’ in Table 1 . To improve the reliability of our approach , we used three predication filtering mechanisms , explained below . These mechanisms result in a smaller set of predications extracted for each pathway , which forms subnetworks . These subnetworks , together , serve as the predication network . Figure 3 provides an example of a subnetwork created from predications related to p53 . This example has been pruned to provide a small network for simplicity . Argument-predicate distance is the number of intervening noun phrases between an argument and its predicate in the sentence from which they were extracted . It has been shown that smaller argument distance leads to higher precision in extracting interactions [62] . In the example below , both Bax and pro-caspase-3 are potential objects for p53 activity . Bax has an object distance of 1 since it is the first noun phrase subsequent to the predicate ‘promoted’ and pro-caspase-3 has a distance of 2 . Preferring an argument-predicate distance of 1 selects the predication p53 STIMULATES Bax , while eliminating the predication p53 STIMULATES pro-caspase-3 . In our experiments , we limit both the predicate-subject distance and the predicate-object distance to 1 , which decreased the number of predications to approximately 27% of the initial number of predications on average . The number of predications for each pathway after filtering using argument-predicate distance of 1 is shown in Table 1 ( under ‘Dist . ’ ) . After argument-distance filtering is applied , we normalize the set of predications by removing duplicate predications and mapping the subjects and objects to formal gene symbols based on the standard gene name dataset ( HUGO , http://www . genenames . org/ ) . If a predication argument cannot be mapped to a formal gene symbol , the predication is pruned . The numbers of predications after duplicate removal and normalization are given in Table 1 ( under ‘Uniq . ’ and ‘Norm . ’ , respectively ) . Document frequency for a predication is the number of citations in which the predication occurs . Document frequency filtering is based on the hypothesis that the confidence credential of a predication is in direct ratio to its occurrence in documents . In our experiments , we discard all predications that occur in fewer than two articles . The result is shown in Table 1 ( ‘Freq . ’ ) . Note that network filtering discards all relevant predications for the Focal Adhesion pathway , which was not considered in subsequent steps . In the final step we quantify gene-gene interaction strength in the network generated in previous steps , using a genetic algorithm , depicted in Figure 4 . In the initial population of chromosomes , each chromosome contains a candidate set of interactions between all genes in the pathway . The predication network determines the initial gene-gene interactions , while the strength of interaction is initially randomly generated for each of the 2000 chromosomes in the population . Then for each generation , the fittest candidates are replicated into the next generation and the balance of the population is filled through reproduction of the current generation , using random crossover and mutation to introduce novel diversity into the population . We compute the fitness of a chromosome , as compared to the time series microarray data , using an artificial neural network . These procedures are described in more detail in the following subsections . With the genetic algorithm used to infer the gene regulatory networks , we train the weights of the inbound interactions for every gene separately . A population of chromosomes is created where each chromosome is represented as a matrix of interaction weights between each possible pairing of genes identified from the predications . Each weight , valued between −1 ( greatest inhibition ) and +1 ( greatest stimulation ) , indicates the strength of interaction . A weight of zero indicates no interaction . The predications in the network define which genes have interactions and whether the interaction is inhibitory or stimulatory , but do not contribute to the determination of the weight . The absence of an interaction is represented in the matrix with a weight of zero but is not altered through subsequent generations . The direction of the interaction is maintained from the subject-object relation in the predication . We randomly generate 2000 chromosomes that contain different weights for the interactions in the predication network . The gene structure in each chromosome is the same , as well as the non-interacting/zero-weighted gene pairs , but the weights representing the strength of inhibition or stimulation of pairs found in the predications are varied . The population of chromosomes is then evaluated with the fitness function in Equation 1 , and the fittest 20 percent of the chromosomes are copied into the next generation directly . The rest ( 80 percent ) of the chromosomes for the next generation are generated by crossover at a specific ( randomly selected ) gene pairing between pairs of randomly selected chromosomes in the current generation with a mutation rate of 0 . 25 percent . This mutation rate approximates those commonly used to facilitate convergence of the chromosomes toward a fittest result and to avoid excessive intergenerational fluctuation . After evolving for 200 generations ( an empirically determined limit ) , the chromosome with the highest fitness value is selected as the final result . In our experiments , we use an artificial neural network ( ANN ) to model the interaction of the genes in each pathway . In this model , the gene expression level of a given gene at a given time point is a function of all other genes at the previous time point ( see Figure 5 ) . In a comparative study by Sîrbu et al . , this model outperformed 4 other recent GRN models ( GA + evolutionary strategy , differential evolution + Akaike's Theoretic criterion , genetic local search , and an iterative algorithm based on GA ) in a combined score of 6 performance measures ( data fit , parameter quality , noise , sensitivity , specificity , and scalability ) [10] . We follow the most common implementation of such a model , using a nonlinear weighted sum . Equation ( 1 ) shows the fitness test function based on the ANN model . ( 1 ) where ( 2 ) and T indicates the number of time points in the microarray dataset , R is the number of microarray replicates at each time point , r is the current replicate , n is the number of genes in the pathway , K is the activation function , is the weight of the interaction between gene j and gene i , and is the gene expression value in the microarray set for gene i at time point t . Since the difference of the expression value of a gene over each time course is considered to be a direct function of only the joint effect of other genes in the pathway and the gene expression values used have been normalized between −1 and 1 , we use a linear function as the activation function , effectively dropping the term . A visual representation of the ANN model is presented in Figure 6 . MEDLINE contains many more citations related to human genes than yeast genes . A simple PubMed search for “yeast and gene” returns a little more than 86 , 000 citations while the query “human and gene” returns over a million . When combining either species term with “cell cycle” ( a concept with a heavy amount of research done in yeast ) , there are over 96 , 000 citations related to human and just over 15 , 000 for yeast . In addition to this limitation , SemRep was designed to identify human genes and uses Entrez Gene database entries specific to humans for genes and proteins . Although this led to an initial study with human genes , we made modifications to SemRep to support a study with yeast . This required changing the data source to the Entrez Gene fungi dataset that contains Saccharomyces species and processing relevant citations ( resulting from the PubMed query “Saccharomyces[mh] AND cell cycle[major]” ) to extract predications . All subsequent procedures were consistent with the breast cancer study except that the predications were not filtered by argument distance or document frequency due to the lower initial number of predications . The number of citations and predications at each step are given in Table 2 .
As a measure of accuracy of the model , we determine how well the model fits the data , in line with previous research [21] . For this purpose , we use microarray data from a human breast cancer experimental set ( GEO: GSE29917 ) , which contains expression values for 7 time points and 6 replicates ( two microarrays are missing from the set for a total of 40 microarrays ) . We compare the accuracy of our model with that of the Repeated Genetic Algorithm with Neural Network model described by Keedwell et al . [21] , the best performing model in a recent comparison of evolutionary algorithms [10] and the basis for the interaction quantification component of our algorithm . Since we use the same algorithm for interaction quantification , the comparison helps isolate the effect of literature-derived knowledge on GRN inference . We downloaded their source code to be able to run their algorithm on our data , with only minimal modification for data format differences . Additionally we included a second comparison model , with a different type of algorithm ( time-delayed Spearman Rank-correlation or TDSRC ) to help identify any biases in our methodology [63] . This model was included by Gupta et al . as a part of a composite model and available for download and use within MatLab . For this comparison , we limit the microarray data to 78 genes included in the KEGG P53 pathway ( listed in Table 3 ) . We use a leave-one-out approach based on time points , sequentially using gene expression values for each time point as test data against models trained with values from all of the other time points . Fitness is defined as the standard deviation of predicted values from the average microarray expression of all 6 replicates in the test set . Because the microarrays for 2 different replicates at two different timepoints were missing , the total number of microarrays was 40 . The root mean square error ( RMSE ) combines fitness for all genes in the given test set . Figure 7 shows the root mean square errors over the 78 genes for each time point for each model . We used a paired t-test to assess the statistical significance of the differences between models . The p-values for each comparison are included in Figure 7 . Our model shows improvement at every test data set/time point over the Keedwell et al . model and is significantly different ( p<0 . 05 ) in 4 out of 6 time points and for all time points combined ( overall p = 8 . 26×10−8 ) . The RSP model performed significantly worse than both of the other models . This demonstrates that our model is significantly better in terms of fitting the microarray data . We assessed the value of gene regulatory networks generated by our approach by comparing them to KEGG pathways , both the kgml format downloaded from the KEGG website and manually using the search function in the online pathway map . Nodes , representing genes , and edges , representing interactions , were independently compared and the results are shown in Table 4 . The ‘Predication’ column represents the number of nodes and edges in the network generated by our approach after removing any genes that were not included in the microarray set , the ‘KEGG’ column provides the equivalent information for the kgml network , ‘Common’ indicates the intersection of our results and the corresponding kgml network , and ‘New’ provides interactions exclusive to our results . As the numbers indicate , new nodes and edges significantly outnumber those in common between the two networks . The p53 pathway had the highest ratio of common edges compared to the kgml network ( 19∶65 , 29 . 2% ) , so was chosen for further validation at the predication level . A comparison was made for each of the new interactions with the online KEGG p53 signaling pathway map . 49 of the 92 interactions contained a gene not included in the map , 7 had both genes present but no interaction , and 9 existed in the map but were not included in the kgml version . We also validated each new interaction in the resulting p53 network against the literature in three ways: a ) whether the literature asserts an interaction between the two at either a gene or protein level ( i . e . , the precision of our natural language processing techniques ) , b ) whether we capture correctly the direction of the effect , i . e . stimulatory or inhibitory , as compared to the weight generated by GRN analysis of the microarray data , and c ) combining the two above , whether we capture correctly both the presence of the interaction and the direction . In addition to assessing the accuracy of SemRep in capturing relevant interactions , this validation allows us to establish the specific contribution of using semantic predications over term co-occurrence . The comparison was limited to citations from which the predications were extracted , ranging from 1 to 87 citations for each interaction . Although there may exist another citation that would validate the resulting interaction , this approach was taken to limit the man-hours required to a reasonable amount while still allowing a reasonable possibility for validation . As seen in Table 5 , 78 . 3% of the new interactions in the resulting p53 network ( Figure 8 ) were asserted in at least one of the source citations . When comparing the sign of the interaction weight to the direction of the effect provided in the source literature , 76 . 4% were consistent . Those interactions that were both consistent with the literature and were weighted in the appropriate direction numbered 55 out of 92 ( 59 . 8% ) . We additionally focused on those interactions having a weight with absolute value >0 . 1 ( Table 5 , bottom ) , exploring the hypothesis that stronger interactions should be less affected by noise . Within these interactions a total of 32 interactions ( 84 . 2% ) were stated in the literature and 34 ( 87 . 5% ) were correct in the direction of effect , yielding 28 ( 73 . 7% ) correct on both counts . Finally , we investigated whether argument-predicate distance filtering had a detrimental effect on the results , since long distance syntactic dependencies are common in biomedical literature [64] . For this purpose , we focused on the Jak-STAT pathway and checked whether any of the 27 KEGG interactions , none of which appeared in our predication network , could be derived from the initial set of unfiltered , non-normalized predications . We found three such predications for the Jak-STAT pathway; one ( Jak1 INTERACTS_WITH PTPN11 ) was eliminated due to argument-distance filtering , while the other two ( Jak1 INTERACTS_WITH STAT1 and GRB2 INTERACTS_WITH PTPN11 ) passed the argument-distance filter but not the document frequency filter . Note that the former ( Jak1 INTERACTS_WITH PTPN11 ) occurred in a single document and , therefore , would also be eliminated in the subsequent document frequency filtering step , had it passed the predicate-argument distance filter . We compared the performance of our model against that of Keedwell et al . in the same manner as mentioned in our results section Comparison of model accuracy , but now on the yeast cell-cycle dataset , which contains 24 time points . Figure 9 shows the root mean-square errors for each time interval for the two models . We assessed statistical significance of the differences between models by calculating the p-values ( included in Figure 9 ) using a paired t-test . Using the yeast data , our model had increased fitness at every interval over the Keedwell et al . model , but this time every difference was significantly different ( p<0 . 05 ) .
Although some of the genes in the KEGG pathways did not appear in the results due to the strict filtering process , the presence of essential member genes in the resulting predication networks ( for example , TP53 , MDMD2 , BAX , CDKN1A , GADD45A , and CDK2 in the p53 network ) and known interactions ( p53 STIMULATES Gadd45a , E2F1 STIMULATES cdkn2a , and RCHY1 INHIBITS tp53 ) demonstrate the potential of this technique to replicate “known” pathways . Perhaps more importantly , we were able to identify new interactions that were not included in the KEGG maps , which is not surprising since these maps are curated and therefore provide only the most thoroughly established and important interactions as determined by the curators . Two of the highest-weighted new interactions included in the resulting p53 pathway but not in the kgml file are p53 STIMULATES BIK and MDM2 INHIBITS CDKN1A . The weight of the interaction between p53 and BIK was very strongly stimulatory at 0 . 977 . BIK ( BCL-2 interacting killer ) is not included in the KEGG P53 pathway map . It is a pro-apoptotic protein discovered in 1995 and interacts strongly with BCL-2 and BCL-xL [65] . P53 has been shown to induce expression of BIK under certain conditions , especially in breast cancer cells , as in our microarray dataset [66] . Our p53 pathway result also included an interaction between MDM2 an CDKN1A with a weight of −0 . 987 , i . e . , very strongly inhibitory . MDM2 is shown to inhibit p53 in the KEGG p53 map but there is only an indirect inhibitory action of MDM2 on CDKN1A through P53 ( by diminishing p53's activation of CDKN1A ) . However , a direct interaction as a negative regulator of CDKN1A is present in the literature [67] , supporting our result . A major limitation of this method is the accumulation of system errors . As shown in Table 5 , the accuracy of interactions identified using SemRep with the filtering used in our method is 78 . 3% overall and 84 . 2% for significantly weighted interactions ( >0 . 1 ) . The reproducibility of the microarray data between platforms and even with different algorithms on the same platform has been determined to be as low as 50–60% [68] . At a result , the accuracy of both the structure and the weights of the resulting GRNs is limited if viewed at the most precise level of granularity . The training of the weights is also limited by an insufficient time course in the microarray data , especially for large numbers of arguments ( i . e . genes in a pathway ) . Although it is beyond our means to improve microarray reproducibility or experimental design of publicly available datasets , improvements to SemRep with regard to gene and protein interactions can potentially improve results and the incorporation of multiple microarray datasets as training data may be able to overcome some of the inherent limitations . Additionally , although these limitations affect the use of these techniques for determining the precise quantitative nature of interactions , this is true generally of such studies , and does not prevent their use for hypothesis generation . Currently our network backbone is limited to genes included in the predications , but this set could be expanded by incorporating genes from the microarrays that are similar to the genes provided by the predications , using an ensemble of machine learning techniques such as support vector machines , random forests , and prediction analysis of microarrays . This approach of classifying genes from the microarray into each pathway would additionally reduce bias in the technique by not limiting interactions to what is already known . It would be valuable to maintain a distinction between the literature-based genes and the expansion set because a significant expansion in the final resulting networks would suggest that what is known in these pathways is only a fraction of what is waiting to be discovered . Another technique to extend the set of genes would be to adopt more sophisticated predication filtering mechanisms . The current mechanisms were effective in significantly reducing the computational complexity; however , our limited analysis of the effect of argument-distance filtering showed that some relevant predications were also eliminated due to filtering . A potentially useful approach would be to train a classifier that can identify ‘good’ predications , based on predication features , such as the predicate type , indicator types , etc . Predicate-argument distance could also serve as a feature for such a classifier . Additionally , although we used interactions from KEGG to evaluate our GRN , such interactions from this and similar databases could be incorporated into our interaction network to augment the proposed interactions from literature with established interactions , expanding the utility of prior knowledge . Although our current method of selecting source citations for predication extraction yielded usable results , there are many possible permutations in the method and a systematic comparison of source citations and their resulting networks should be explored . Our current search method facilitates the generation of hypothetical member genes for established pathways , but this approach can also be used to generate novel pathways by specifying a gene set and/or biological functions in the predication citation search , thereby providing an appropriate predication network . SemRep representation of biomolecular interactions ( subject-predicate-object triples ) is simple , intuitively accessible and lends itself easily to downstream applications . On the other hand , more complex representations , particularly event representation ( as discussed in Related work section ) , have been gaining in popularity in the BioNLP community , mostly due to available corpora [69] and shared task competitions [27] . An obvious question is whether such complex representations could provide an advantage over or could complement SemRep representation in the task of gene regulatory network inference . It seems very likely that generating a seed network from more complex representations would require some non-trivial post-processing along the lines of the algorithm described by Bossy et al . [36] . Since that algorithm essentially breaks down complex relations to simpler SemRep-style triples , it seems safe to assume that SemRep representation can adequately capture the complexity of biomolecular interactions . However , this needs further testing and validation . To assess the precision of interactions for the yeast study , we performed an automated comparison using Cytoscape ( http://www . cytoscape . org ) against a Biogrid ( http://thebiogrid . org ) yeast interaction dataset . We used the Biogrid Saccharomyces_cerevisiae-3 . 2 . 109 tab file , which contains 339 , 921 interactions and experimental information from multiple database sources including Saccharomyces Genome Database , MINT , IntAct , and Pathway Commons . As seen in Table 6 , although 346 out of 349 genes are found in the reference set , only 147 out of 520 interactions are included . That is a relatively low precision of 28% , which only increases to 31% for interactions weighted 0 . 1 or higher . This is not particularly meaningful without appropriate context . Since this is a relatively straightforward and easily undertaken evaluation approach , it would be worthwhile to conduct a study applying the same evaluation across various published models to see how precision compares among them . We present a methodology of gene regulatory network inference that combines literature knowledge and microarray data . Using SemRep , we extract gene and protein interactions from citations related to the pathways included in the KEGG Pathways of Cancer . These predications are linked together to form a network and a genetic algorithm is used on a breast cancer time sequence microarray dataset to determine the weight of contribution of each stimulating or inhibiting gene on its target , thereby providing a weighted gene regulatory network . Our model performs significantly better than comparable models in terms of fitness of predictive output to microarray results . The resulting networks contain both interactions included in the appropriate KEGG pathways and interactions not included but validated through a literature search . The accuracy of these interactions increases from 60% overall to 74% when minimally-weighted interactions are excluded . Our model also performed better in terms of fitness against a comparison model when modified for yeast predications and microarray data . This approach offers significant potential in elucidating new interactions in existing pathways as well as the possibility of identifying novel pathways . In a broader sense , it also provides a blueprint for incorporating automatically extracted knowledge from literature with large-scale biological analysis . Incorporating such knowledge underpins more accurate understanding of both normal and disturbed molecular physiology , leading to advances in diagnosis and treatment of disease . | We have developed a methodology that combines standard computational analysis of gene expression data with knowledge in the literature to identify pathways of gene and protein interactions . We extract the knowledge from PubMed citations using a tool ( SemRep ) that identifies specific relationships between genes or proteins . We string together networks of individual interactions that are found within citations that refer to the target pathways . Upon this skeleton of interactions , we calculate the weight of the interaction with the gene expression data captured over multiple time points using state-of-the-art analysis algorithms . Not surprisingly , this approach of combining prior knowledge into the analysis process significantly improves the performance of the analysis . This work is most significant as an example of how the wealth of textual data related to gene interactions can be incorporated into computational analysis , not solely to identify this type of pathway ( a gene regulatory network ) but for any type of similar biological problem . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"computer",
"and",
"information",
"sciences",
"natural",
"language",
"processing",
"network",
"analysis",
"regulatory",
"networks",
"biology",
"and",
"life",
"sciences",
"information",
"technology",
"computational",
"biology"
] | 2014 | Augmenting Microarray Data with Literature-Based Knowledge to Enhance Gene Regulatory Network Inference |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.